UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl) UvA-DARE (Digital Academic Repository) Organ motion in children for high-precision radiotherapy Why treat children like adults? Huijskens, S.C. Publication date 2019 Document Version Final published version License Other Link to publication Citation for published version (APA): Huijskens, S. C. (2019). Organ motion in children for high-precision radiotherapy: Why treat children like adults?. General rights It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons). Disclaimer/Complaints regulations If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You will be contacted as soon as possible. Download date:11 Jul 2022
Transcript
UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)
UvA-DARE (Digital Academic Repository)
Organ motion in children for high-precision radiotherapyWhy treat children like adults?Huijskens, S.C.
Publication date2019Document VersionFinal published versionLicenseOther
Link to publication
Citation for published version (APA):Huijskens, S. C. (2019). Organ motion in children for high-precision radiotherapy: Why treatchildren like adults?.
General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an opencontent license (like Creative Commons).
Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, pleaselet the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the materialinaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letterto: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. Youwill be contacted as soon as possible.
The research was supported by grants from KiKa and KWF Kankerbestrijding. Printing of this thesis was financially supported by Elekta B.V.
Organ motion in children for high-precision radiotherapy
Why treat children like adults?
ACADEMISCH PROEFSCHRIFT
ter verkrijging van de graad van doctor
aan de Universiteit van Amsterdam
op gezag van de Rector Magnificus
prof. dr. ir. K.I.J. Maex
ten overstaan van een door het College voor Promoties ingestelde commissie,
in het openbaar te verdedigen in de Aula der Universiteit
op vrijdag 17 mei 2019, te 13:00 uur
door
Sophie Cathelijne Huijskens
geboren te Leidschendam
Promotiecommissie Promotor:
Prof. dr. C.R.N. Rasch AMC-UvA
Copromotores:
Dr. A. Bel AMC-UvA
Dr. T. Alderliesten AMC-UvA
Overige leden:
Prof. dr. L.J.A. Stalpers AMC-UvA
Prof. dr. M.B. van Herk AMC-UvA
Prof. dr. B.W. Raaymakers Universiteit Utrecht
Dr. J.H. Maduro Rijksuniversiteit Groningen
Dr. J.H.M. Merks AMC-UvA
Faculteit der Geneeskunde
Contents
Chapter 1
General introduction
6
Chapter 2 Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: A multicenter study
22
Chapter 3 Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference?
36
Chapter 4 Abdominal organ position variation in children during image-guided radiotherapy
58
Chapter 5 Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy
78
Chapter 6 The effectiveness of 4DCT in children and adults: a pooled analysis
96
Chapter 7
Predictive value of pediatric respiratory-induced diaphragm motion quantified using pre-treatment 4DCT and CBCTs
116
Chapter 8
General discussion
132
Summary
156
Samenvatting
162
Addendum List of publications PhD portfolio Curriculum vitae Dankwoord
168
Chapter 1 General introduction
7
1.1 | Childhood cancer Undergoing treatment for childhood cancer has a big impact on the life of children and their families [1, 2]. Suddenly, they have to cope with many medical procedures, doctors, and long hospital stays. This can be an emotionally, frightening, and stressful period. As a result of improved treatment modalities, long-term childhood cancer survival has increased, however, accompanied by late treatment-related adverse events. Because of the young age of these cancer patients, and their potential longevity when surviving childhood cancer, extended surveillance and long-term follow-up procedures are needed to monitor the late consequences of therapy [3]. Incidence and survival rates Childhood cancer is a rare disease. In the Netherlands, each year around 500 children aged 0-18 years are diagnosed with cancer, which comprises 0.6% of all new cancer cases per year (Figure 1.1) [4]. Childhood cancer is the most common cause of disease-related mortality in children [5]. With continuous developments in diagnostic procedures and multimodal treatment strategies, survival rates have improved rapidly over the past decades, from approximately 25% in the 1960s up to 80% nowadays [6, 7]. With the improved survival rates, the number of childhood cancer survivors has grown impressively. As new treatment strategies and innovative techniques will continue to evolve and be integrated in childhood cancer care, the ambition is that this number continues to grow, as illustrated in Figure 1.2. Adverse events Increasing survival rates have also led to increasing incidence of (late) adverse events [8, 9]. These young cancer survivors have a long life expectancy and remain at risk for (late) adverse events due to the malignancy and/or treatment, including second malignancies and toxicities to important healthy organs. From the Emma Children’s Hospital / Academic Medical Center (EKZ/AMC) cohort [10], including 1362 5-year survivors of childhood cancer, almost 75% of survivors had one or more adverse events, and 25% had five or more adverse events. Furthermore, 40% of survivors had at least one severe or life-threatening or disabling adverse event. Specifically, a high or severe burden was observed in 55% of the survivors who received treatment with radiotherapy only. Therefore, solely focusing on cure of cancer is not enough. Although delivering adequate tumor dose to kill the tumor cells is the primary goal in radiotherapy, sparing the vital and long-term functions of adjacent organs is also paramount.
8
Figure 1.1 | Age-specific cancer incidence in the Netherlands in 2017. Adapted from [4].
Figure 1.2 | Survival rates for Dutch childhood cancer survivors. Adapted from [4].
60%
65%
70%
75%
80%
85%
90%
95%
100%
0 2 4 6 8 10 12
Prop
ortio
n su
rviv
ing
(%)
Survival from diagnosis (years)
1989-1994 1995-2000 2001-2005
2006-2010 2011-2015
9
Childhood cancer types Cancer starts when abnormal cells in the body grow uncontrollably, often forming a mass of tissue (i.e., a tumor) or affecting the blood or immune system. Childhood cancer is different from cancer in adults considering pathological characteristics, underlying biology, treatment approach, and survival opportunities. In adults, many cancer types are related to lifestyle or environmental risk factors, such as smoking, unhealthy diet, or consuming alcohol [11, 12]. Childhood cancer types are commonly not caused by lifestyle-related risk factors but develop for example by mutations in early fetal development or through the presence of genetic predisposition [13]. However, much uncertainty remains over what causes childhood cancer. Childhood cancer comprises a heterogeneous group of tumor types and can evolve in various body sites and organs. With classification depending on their cell of origin, childhood cancer types can be distinguished in hematological malignancies and solid cancers. They are mostly unique for children and are rarely presented in adults. Incidence rates for hematological malignancies are highest for leukemia and lymphoma [14]. Most frequently diagnosed solid tumors are central nervous system (CNS) tumors, neuroblastomas, sarcomas, Wilms’ tumors, and retinoblastomas. Neuroblastomas originate most frequently in the adrenal and retroperitoneal regions, but can also develop in the thoracic, pelvis or neck region [15]. Sarcomas can be divided in soft tissue sarcomas and bone sarcomas [16], and can be located at various sites. Wilms’ tumors and retinoblastomas are site-specific and originate in the kidney and eye, respectively. This distribution of cancers varies among age groups [4, 14]. In infants, the most predominant cancer is neuroblastoma, while leukemia is more common in 1-4-year-old children, and CNS tumors in 5-9-year-old children. For older children, bone tumors, lymphomas, and carcinomas are more common. Treatment options Since childhood cancer types differ from adult cancer types, the approach and response to treatment varies. Each type of cancer requires a different treatment approach. The three main components of treatment are surgery, chemotherapy and radiotherapy. Childhood cancers seem to respond better to high doses of chemotherapy. Secondly, pediatric tissues are still in development and their organs have lower tolerance to radiation than adults [17, 18]. Childhood cancer treatment is challenged by the small disease population. For most childhood cancer types, collaborative groups, such as the Children’s Oncology Group (COG) [19], the International Society of Paediatric Oncology (SIOP) [20], and Paediatric Radiation Oncology Society (PROS) [21], have developed child-specific diagnostic procedures and treatment protocols. Depending on histology, risk group, and/or stage of cancer, children are treated according to these proposed protocols. However, due to the rarity of the disease and patient-specific needs, a multidisciplinary team needs to work intensively together to create the optimal treatment plan for each individual child. The optimal treatment approach should aim to minimize adverse health-related outcomes and to maximize quality of life outcomes [8]. Historically, the use of radiotherapy has contributed to improved survival rates of several childhood cancers [22]. However, since the late treatment-related adverse events from radiotherapy in childhood cancer survivors are well known and documented [9, 10], effort has been made to reserve the use of
10
radiotherapy. Additionally, with effective chemotherapeutics and improved surgical techniques, the use of radiotherapy has varied over time [23]. Especially, (low risk) leukemia and lymphoma patients no longer receive radiotherapy as standard treatment option. For other cancer types, such as neuroblastomas, Wilms’ tumors, brain or bone cancers, the use of radiotherapy declines moderately or is stable [23, 24]. For those tumor types, radiotherapy, along with surgery and chemotherapy, plays an important role in the treatment. 1.2 | Pediatric radiotherapy Radiotherapy uses ionizing radiation to kill the tumor cells. The most commonly applied form of radiotherapy is photon therapy, using high energy X-rays (1-20 MeV) and is usually delivered with external beams generated by a linear accelerator (linac) (Figure 1.3). An optimal treatment plan in radiotherapy aims to deliver a sufficiently high dose to kill the tumor while at the same time minimizing doses to surrounding healthy tissues, or organs at risk (OARs). Developing the preferred treatment plan for each individual patient is a complex procedure and requires multidisciplinary collaboration of pediatric oncologists, radiation oncologists, physicists, and radiation therapists.
Figure 1.3 | A linear accelerator (Agility, Elekta AB, Stockholm, Sweden) with integrated kV-CBCT system used for image-guided radiotherapy.
11
Treatment planning At the start of the radiotherapy process, a pre-treatment computed tomography (CT) scan of the patient in radiation treatment position is made for treatment planning purposes. The target (i.e., tumor) and surrounding OARs are delineated on this planning CT scan, which is usually a three-dimensional (3D) scan, consisting of two-dimensional images that are taken along the cranial-caudal direction. The scan yields a short acquisition time over repeatedly changing anatomy during respiration, thereby not providing information on breathing motion that sometimes causes motion artefacts in the image. This is especially of concern in the abdominal and thoracic region, where target volumes and OARs are more prone to motion than in the cranial area or in extremities. For abdominal and thoracic target locations, the pre-treatment 3DCT scan is therefore more often replaced by a four-dimensional CT scan (4DCT) [25–27], where the complete respiratory cycle, from end-inhalation to end-exhalation, is divided into typically 10 phase scans. For treatment planning purposes, either the averaged 4DCT scan, or other strategies including all phases or one phase scan, are used. The radiation oncologist delineates the gross tumor volume (GTV), representing the macroscopic tumor visible on the planning CT [28]. However, in children, the tumor is often surgically removed before radiation treatment. For those cases, the target volume (i.e., tumor bed) includes the preoperative extent of disease or the residual macroscopic disease, and is delineated based on diagnostic information, preoperative imaging and surgical reports or clips (if present). The GTV is expanded with a margin to account for microscopic tumor growth, defining the clinical target volume (CTV). To account for geometrical uncertainties, such as a delineation error and anatomical variations, a margin is added to the CTV, resulting in the planning target volume (PTV) [28–30]. Similar margin definitions are also taken into consideration for OARs to define adequate planning risk volumes (PRVs) [31]. In the planning phase, a treatment plan is created including the delivery technique, describing the direction, intensity, and shape of the radiation beams. Treatment delivery
Radiotherapy courses are typically delivered in several fractions. Before delivery of each treatment fraction, patients are positioned on the treatment table identical to the position during the acquisition of the planning CT. This is done by aligning the in-room laser system with the pre-treatment applied skin marks. However, this positioning provides no information on actual internal anatomical variations relative to the skin marks. This led to the introduction of image-guided radiotherapy (IGRT), including an in-room kilovoltage (kV) or megavoltage (MV) imaging system integrated to the linac (Figure 1.3) [32]. The most used form nowadays is (kV) cone beam CT (CBCT) imaging, enabling pre-fraction 2D imaging of the internal anatomy, which is subsequently reconstructed to a 3D image and registered to the planning CT for set-up verification [33]. Usually, bony anatomy serves as a surrogate of the target for the rigid registration, because tumor visibility is difficult due to poor soft tissue contrast of the CBCT image (Figure 1.4). According to the resulting translations and rotations, the patients’ position is corrected ideally to the position identical to the one during the acquisition of the planning CT.
12
Next, the patient needs to remain alone in the treatment room during radiation dose delivery. For children, being separated from their parents, can cause additional anxiety, distress and inability of lying still [34–38]. To prepare children mentally and physically for this procedure, children are invited for a rehearsal session and are thereby guided by a specialized pediatric nurse who supports the child throughout the complete treatment course. During the rehearsal session, the child (and parents) experience a real treatment preparation. The steps of patient positioning and a dummy-session of dose delivery are undertaken, without the fear and pressure of actual treatment, to increase comfort and familiarity with the treatment room and procedure. Sometimes, in more complicated situations, children are immobilized in a vacuum matrass and some children, usually younger patients (< approximately 5 years [39, 40]) require general anesthesia (in the form of sedation). After proper patient-positioning, irradiation with high-energy X-rays starts. According to the planned beam directions, intensity and shape, the gantry rotates around the patient and delivers the radiation dose. In the past, static beams from typically two to four directions were used, resulting in high doses to the target volume but also dose to a large volume of surrounding healthy tissues. To create a high dose area more conformal to the PTV, a multi-leaf collimator (MLC) was introduced to enable beam shaping to anatomical structures in order to match the shape around the tumor whilst avoiding the surrounding OARs [41]. This is known as 3D-conformal radiotherapy (3D-CRT). The MLC can also control and modulate the intensity of the beams, achieving a specific dose profile, which is so-called intensity-modulated radiotherapy (IMRT) [42]. Dose delivery by continuously modulating the shape and intensity of the beam, while the gantry rotates around the patient, is called volumetric modulated arc therapy (VMAT) [43]. Although these newer delivery techniques potentially improve conformity and spare OARs, the introduction of these delivery techniques are moving very slowly into the field of pediatric radiotherapy. Image Guided Radiotherapy IGRT, by using an in-room imaging device for patient-positioning, has played a crucial role over the last decades and has contributed to higher accuracy in radiotherapy [44]. However, image acquisition for IGRT purposes adds to the dose from therapeutic treatment and could contribute to the risk of late adverse events [45]. Especially in the pediatric population, who have higher susceptibility to radiation compared to adults, the limitation of additional dose has to be considered. Moreover, the increased treatment time for IGRT may also increase the risk of motion. Therefore, use of IGRT in children merits specific attention, but it was slowly integrated in pediatric radiotherapy [45]. The first publication evaluating clinical practice of pediatric IGRT dates from 2014 [46]. Among international multi-institutes, IGRT was commonly used but consensus in applied procedures for any given tumor type was lacking [46]. Efforts to reduce imaging dose have resulted in low dose protocols, thereby avoiding unnecessary toxicity without compromising image quality [47]. Although reports on pediatric IGRT are steadily increasing, little is known on exact numbers of current practice of pediatric IGRT. In our institute, pediatric IGRT was introduced in 2010. The acquisition of CBCT imaging did not only allow for position verification, but also enabled to retrospectively investigate the anatomical variations that could occur during the treatment course (Figure 1.4). Tumors and OARs located extra-cranially are more prone to anatomical changes than when they are located intra-cranially. These anatomical variations could limit the accuracy of the treatment plan and are accounted for in the CTV-to-PTV and
13
PRV margin. In order to appropriately account for these uncertainties, quantification of abdominal and thoracic motion is essential to reach high accuracy in pediatric radiotherapy. This thesis particularly focusses on organ motion, which can be described by interfractional position variation (i.e., day-to-day variations of the anatomy) and intrafractional motion (mainly caused by respiration). 1.3 | Organ motion With the introduction of in-room CBCT imaging the interfractional setup error is minimized. However, since the image quality of CBCT is often too poor to clearly visualize the tumor or tumor bed, a surrogate is required for the rigid registration. Therefore, the position verification is then either based on bony structure or implanted markers, which are clearly visible on CBCT images. This means that a residual error between the tumor and the surrogate remains (Figure 1.4). OARs also move relative to the surrogate. Interfractional position variation is often related to variations in bowel- and bladder filling, tumor size, patient weight and treatment-induced tissue changes. This interfractional position variation (after setup verification) needs to be determined in order to be considered when defining the CTV-to-PTV margin. Additionally, the CTV-to-PTV margin accounts for intrafractional motion, which is mainly induced by respiration. Efforts to manage respiration, such as breath holding, gating techniques or real-time monitoring, have been introduced in radiotherapy, but reports in pediatric radiotherapy are scarce [48, 49]. Although also children could potentially benefit from these techniques, they experience radiotherapy already as a stressful procedure [36, 50, 51] and these techniques may cause further distress and anxiety. Additionally, it is questionable if the youngest children (e.g., <8 years) would be able to follow a breath-hold procedure. Furthermore, due to the ALARA principle (keeping doses As Low As Reasonably Achievable), and previously reported radiation risks in children from CT [52, 53], 4DCT is not frequently used in children. Therefore, respiratory motion in children needs to be accounted for in the CTV-to-PTV margin. In case a pre-treatment 4DCT is available, respiratory-induced motion can be measured pre-treatment. Solely accounting for the internal motion then leads to the internal target volume (ITV), which includes the CTV plus an internal margin, covering the entire respiratory-induced motion range. However, this leads to large margins and increased dose to healthy surrounding tissues. An alternative approach, the mid-ventilation based PTV planning, leads to smaller margins, simultaneously accounting for respiratory motion and other geometrical uncertainties [54, 55].
Margin recipe Geometrical uncertainties are a superposition of a systematic and a random error, and form the basis for the CTV-to-PTV margin recipes [56–60]. Systematic errors originate in the treatment preparation phase and therefore affect all treatment fractions. Random errors occur arbitrarily and could have a different effect each single fraction [58], effectively blurring the dose distribution. Thus, the effect of the systematic error on the dose distribution is more significant than that of random errors. Interfractional position variation and intrafractional motion have both systematic and random components. However, when a 4DCT is acquired to assess respiratory motion pre-treatment, the intrafractional motion will only add to the random deviation.
14
Population-based margins are derived from these errors, which are usually determined for specific treatment sites and patient groups. For each patient, a mean error and standard deviation (SD) is measured over a number of fractions. The CTV-to-PTV margin recipe is typically formulated as 2.5 Σ +0.7σ, where the systematic error (denoted as Σ) is the square root of the quadratic sum of standards deviations (SD) of the individual means of the errors and the random error (denoted as σ) is the square root of the quadratic sum of the root mean squares of the individual SDs of the errors [57, 59]. Similar margin definitions are also formulated for OARs to define adequate planning risk volumes (PRV) [31]. In adults, many studies focus on quantification of the organ motion in order to define accurate population-based margins for specific tumor sites [61]. Since childhood cancer is a rare disease, the small population makes it difficult to derive population-based margins for children. More importantly, childhood cancer patients are a highly diverse group, varying from infants to adolescents with different heights and weights. Therefore, margins for children should be defined with more distinction. Additionally, achievements in radiotherapy are mainly focused on adult patients and pragmatically translated and implemented into a pediatric setting [62]. Data on appropriate margin sizes in pediatric radiotherapy is lacking, and clinically used margins are mainly based on experience of the radiotherapist or knowledge based on available adult data. Also, since children are smaller than adults and body compositions differ, it is expected that organ motion also differs. With the introduction of IGRT in children, imaging data has become available to quantify organ motion in children and thus a first step towards being able to derive adequate margins for children. Besides, the benefit of IGRT includes pre-treatment position verification and possibly reduces PTV expansions and dose to OARs [63, 64]. However, in institutes where IGRT is frequently used in children, there was notably variability in PTV expansions for different tumor sites [46]. This stressed the need for a consensus regarding appropriate margin definitions in children, as was also recommended by the PROS [35].
15
Figure 1.4 | Example of position verification using the bony anatomy as surrogate and measuring the residual error (from top to bottom): unaligned overlap of planning CT (in purple) and CBCT (in green), bones aligned (in white), right kidney aligned (note: bones shifted).
Unaligned initial overlap: CBCT + CT
Bony anatomy aligned
Kidney aligned (note: bones shifted)
16
1.4 | Objective and outline of this thesis Since treatment with radiotherapy significantly contributes to the risk of developing adverse events later in life of this vulnerable group of childhood cancer survivors, extremely high accuracy in radiotherapy is essential. However, geometrical uncertainties, due to interfractional position variation (i.e., day-to-day variations of the anatomy) and intrafractional motion (mainly caused by respiration), are present and limit the accuracy. To account for these uncertainties, a safety margin is added to ensure target coverage. However, in children, data on these uncertainties is lacking. Margins in pediatric radiotherapy are based on clinical experience from the radiotherapist or pragmatically translated from adult data and settings. Therefore, the main focus of this thesis is the quantification of interfractional position variation and intrafractional motion in children in order to define appropriate child-specific based margins. Children are treated with abdominal and thoracic radiotherapy for a wide range of primary cancer diagnosis. Especially in the abdominal and thoracic area, tumors and organs are prone to motion. Moreover, the tumor can be in very close proximity to radiosensitive OARs. Therefore, in the first part of the thesis (chapters 2-4) we focus on interfractional position variation of abdominal organs. First, in chapter 2, we quantify interfractional position variation of the kidneys and the diaphragm in children in order to estimate the interfractional component of the safety margin. For comparison with adults, in chapter 3, we analyze interfractional position variation of the kidneys and diaphragm in adults and compare both groups. Additionally, we investigate the possible correlations of continuous values of age, height and weight with interfractional position variation. In chapter 4, we quantify interfractional position variation of several abdominal organs in children. To increase insight on abdominal organ position variation and in order to evaluate if close located organs could function as a surrogate for organ motion, we investigate the possible correlations of position variations between the kidneys, spleen, liver and diaphragm. In the second part of this thesis, intrafractional motion will be investigated. The magnitude and variability of respiratory-induced diaphragm motion in children is described in chapter 5, and a pooled analysis of pediatric and adult data is performed in chapter 6. In order to assess respiratory motion prior to treatment, pre-treatment 4DCT was introduced in pediatric radiotherapy. However, studies on adult patients have indicated that respiratory motion, as measured on 4DCT, is not always representative for respiratory motion during the subsequent treatment course. Therefore, in chapter 7, we analyze whether, in children, respiratory-induced diaphragm motion on a single 4DCT can accurately predict daily respiratory motion. From the interfractional position variation and intrafractional motion, the systematic and random errors can be calculated for the abdominal and thoracic areas, leading towards a first step in assessing appropriate pediatric-data driven margins. Moreover, for techniques like proton and carbon ion therapy, where, compared to photon therapy, the very sharp dose fall-off of protons and carbon ions are even less forgiving for anatomical variations during treatment, accounting for organ motion is paramount [65, 66]. Therefore, quantification and an extensive understanding of organ motion in children is essential for high-accuracy image-guided radiotherapy in children.
17
References 1. Jantien Vrijmoet-Wiersma CM, van Klink JMM, Kolk AM et al. Assessment of Parental
Psychological Stress in Pediatric Cancer: A Review. J. Pediatr. Psychol. 2008; 33(7):694–706. 2. Patterson JM, Holm KE, Gurney JG. The impact of childhood cancer on the family: a qualitative
analysis of strains, resources, and coping behaviors. Psychooncology. 2004; 13(6):390–407. 3. Winther JF, Kenborg L, Byrne J et al. Childhood cancer survivor cohorts in Europe. Acta Oncol.
(Madr). 2015; 54(5):655–668. 4. Cijfers over Kanker. [https://www.cijfersoverkanker.nl/kinderen-en-kanker]. 5. Kaatsch P. Epidemiology of childhood cancer. Cancer Treat. Rev. 2010; 36(4):277–285. 6. Gatta G, Zigon G, Capocaccia R et al. Survival of European children and young adults with cancer
diagnosed 1995-2002. Eur. J. Cancer 2009; 45(6):992–1005. 7. Gatta G, Botta L, Rossi S et al. Childhood cancer survival in Europe 1999–2007: results of
EUROCARE-5—a population-based study. Lancet Oncol. 2014; 15(1):35–47. 8. Robison LL, Hudson MM. Survivors of childhood and adolescent cancer: Life-long risks and
responsibilities. Nat. Rev. Cancer 2014; 14(1):61–70. 9. Oeffinger KC, Mertens AC, Sklar C a et al. Chronic Health Conditions in Adult Survivors of
Childhood Cancer. N. Engl. J. Med. 2006; 355(15):1572–1582. 10. Geenen MM, Cardous-Ubbink MC, Kremer LCM et al. Medical Assessment of Adverse Health
Outcomes in Long-term Survivors of Childhood Cancer. JAMA 2007; 297(24):2705. 11. Danaei G, Vander Hoorn S, Lopez AD et al. Causes of cancer in the world: comparative risk
assessment of nine behavioural and environmental risk factors. Lancet 2005; 366(9499):1784–1793.
12. Lichtenstein P, Holm N V., Verkasalo PK et al. Environmental and Heritable Factors in the Causation of Cancer — Analyses of Cohorts of Twins from Sweden, Denmark, and Finland. N. Engl. J. Med. 2000; 343(2):78–85.
13. Spector LG, Pankratz N, Marcotte EL. Genetic and Nongenetic Risk Factors for Childhood Cancer. Pediatr. Clin. North Am. 2015; 62(1):11–25.
14. Steliarova-Foucher E, Colombet M, Ries LAG et al. International incidence of childhood cancer, 2001–10: a population-based registry study. Lancet Oncol. 2017; 18(6):719–731.
15. Vo KT, Matthay KK, Neuhaus J et al. Clinical, Biologic, and Prognostic Differences on the Basis of Primary Tumor Site in Neuroblastoma: A Report From the International Neuroblastoma Risk Group Project. J. Clin. Oncol. 2014; 32(28):3169–3176.
16. Williams RF, Fernandez-Pineda I, Gosain A. Pediatric Sarcomas. Surg. Clin. North Am. 2016; 96(5):1107–1125.
17. Selo N, Bölling T, Ernst I et al. Acute toxicity profile of radiotherapy in 690 children and adolescents: RiSK data. Radiother. Oncol. 2010; 97(1):119–126.
18. Claude L, Laprie A. Quelles contraintes pour quels organes à risque en radiothérapie chez l’enfant? Cancer/Radiotherapie 2015; 19(6–7):484–488.
19. Children’s Oncology Group. [https://www.childrensoncologygroup.org/]. 20. International Society of Paediatric Oncology. [https://siop-online.org/]. 21. Paediatric Radiation Oncology Society. [http://intpros.org/]. 22. O’Leary M, Krailo M, Anderson JR, Reaman GH. Progress in Childhood Cancer: 50 Years of
Research Collaboration, a Report From the Children’s Oncology Group. Semin. Oncol. 2008; 35(5):484–493.
23. Jairam V, Roberts KB, Yu JB. Historical trends in the use of radiation therapy for pediatric cancers: 1973-2008. Int. J. Radiat. Oncol. Biol. Phys. 2013; 85(3):1–11.
24. Green DM, Kun LE, Matthay KK et al. Relevance of historical therapeutic approaches to the contemporary treatment of pediatric solid tumors. Pediatr. Blood Cancer 2013; 60(7):1083–1094.
25. Beltran C, Pai Panandiker AS, Krasin MJ, Merchant TE. Daily image-guided localization for
18
neuroblastoma. J. Appl. Clin. Med. Phys. 2010; 11(4):162–169. 26. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as
measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
27. Guerreiro F, Seravalli E, Janssens GO et al. Intra- and inter-fraction uncertainties during IGRT for Wilms’ tumor. Acta Oncol. (Madr). 2018; 57(7):941–949.
29. Morgan-Fletcher SL. Prescribing, Recording and Reporting Photon Beam Therapy (Supplement to ICRU Report 50), ICRU Report 62 . ICRU, pp. ix+52, 1999 (ICRU Bethesda, MD) $65.00 ISBN 0-913394-61-0. Br. J. Radiol. 2001; 74(879):294–294.
30. Stroom JC, Heijmen BJM. Geometrical uncertainties, radiotherapy planning margins, and the ICRU-62 report. Radiother. Oncol. 2002; 64(1):75–83.
31. McKenzie A, van Herk M, Mijnheer B. Margins for geometric uncertainty around organs at risk in radiotherapy. Radiother. Oncol. 2002; 62(3):299–307.
32. Jaffray DA, Siewerdsen JH. Cone-beam computed tomography with a flat-panel imager: Initial performance characterization. Med. Phys. 2000; 27(6):1311–1323.
33. Yin F-F, Wong J, Balter J et al. The Role of In-Room kV X-Ray Imaging for Patient Setup and Target Localization Report of AAPM Task Group 104. Data Manag. 2009.
34. Katz ER, Kellerman J, Siegel SE. Behavioral distress in children with cancer undergoing medical procedures: Developmental considerations. J. Consult. Clin. Psychol. 1980; 48(3):356–365.
35. Scott MT, Todd KE, Oakley H et al. Reducing Anesthesia and Health Care Cost Through Utilization of Child Life Specialists in Pediatric Radiation Oncology. Int. J. Radiat. Oncol. Biol. Phys. 2016; 96(2):401–405.
36. Filin A, Treisman S, Peles Bortz A. Radiation Therapy Preparation by a Multidisciplinary Team for Childhood Cancer Patients Aged 3½ to 6 Years. J. Pediatr. Oncol. Nurs. 2009; 26(2):81–85.
37. Woodman H. Put yourself in their shoes the vulnerability of children and their families when attending for radiotherapy treatment: The role of the specialist paediatric radiotherapy radiographer. Radiography 2013; 19(4):311–314.
38. Engvall G, Ångström-Brännström C, Mullaney T et al. It is tough and tiring but it works- children’s experiences of undergoing radiotherapy. PLoS One 2016; 11(4):1–13.
39. Fortney JT, Halperin EC, Hertz CM, Schulman SR. Anesthesia for pediatric external beam radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 1999; 44(3):587–91.
40. Evans P, Chisholm D. Anaesthesia and paediatric oncology. Curr. Anaesth. Crit. Care 2008; 19(2):50–58.
41. Cosgrove VP, Thomas MDR, Weston SJ et al. Physical Characterization of a New Concept Design of an Elekta Radiation Head with Integrated 160-leaf Multi-leaf Collimator. Int. J. Radiat. Oncol. 2009; 75(3):S722–S723.
42. Group IMRTCW. Intensity-modulated radiotherapy: current status and issues of interest. Int. J. Radiat. Oncol. 2001; 51(4):880–914.
43. Otto K. Volumetric modulated arc therapy: IMRT in a single gantry arc. Med. Phys. 2007; 35(1):310–317.
44. Gupta T, Narayan Ca. Image-guided radiation therapy: Physician′s perspectives. J. Med. Phys. 2012; 37(4):174.
45. Hess CB, Thompson HM, Benedict SH et al. Exposure Risks among Children Undergoing Radiation Therapy: Considerations in the Era of Image Guided Radiation Therapy. Int. J. Radiat. Oncol. Biol. Phys. 2016; 94(5):978–992.
46. Alcorn SR, Chen MJ, Claude L et al. Practice patterns of photon and proton pediatric image guided radiation treatment: Results from an International Pediatric Research Consortium. Pract. Radiat. Oncol. 2014; 4(5):336–341.
47. De Jong R, Lens E, Van Herk M et al. Optimizing cone-beam CT presets for children to reduce
19
imaging dose illustrated with craniospinal axis. Radiother. Oncol. 2014; 111:S109–S110. 48. Demoor-Goldschmidt C, Chiavassa S, Josset S et al. Asservissement respiratoire lors d’une
radiothérapie pulmonaire bilatérale pour le sarcome d’Ewing ou le néphroblastome chez des enfants et jeunes adultes : études dosimétrique et clinique de faisabilité. Cancer/Radiotherapie 2017; 21(2):124–129.
49. Claude L, Malet C, Pommier P et al. Active Breathing Control for Hodgkin’s Disease in Childhood and Adolescence: Feasibility, Advantages, and Limits. Int. J. Radiat. Oncol. Biol. Phys. 2007; 67(5):1470–1475.
50. Tyc VL, Klosky JL, Kronenberg M et al. Children’s Distress in Anticipation of Radiation Therapy Procedures. Child. Heal. Care 2002; 31(1):11–27.
51. Bucholtz JD. Comforting children during radiotherapy. Oncol. Nurs. Forum 1994; 21(6):987–94. 52. Slovis TL. Children, computed tomography radiation dose, and the As Low As Reasonably
Achievable (ALARA) concept. Pediatrics 2003; 112(4):971–972. 53. Brody AS, Frush DP, Huda W, Brent RL. Radiation Risk to Children From Computed Tomography.
Pediatrics 2007; 120(3):677–682. 54. Wolthaus JWH, Sonke JJ, van Herk M et al. Comparison of Different Strategies to Use Four-
Dimensional Computed Tomography in Treatment Planning for Lung Cancer Patients. Int. J. Radiat. Oncol. Biol. Phys. 2008; 70(4):1229–1238.
55. Wolthaus JWH, Schneider C, Sonke JJ et al. Mid-ventilation CT scan construction from four-dimensional respiration-correlated CT scans for radiotherapy planning of lung cancer patients. Int. J. Radiat. Oncol. Biol. Phys. 2006; 65(5):1560–1571.
56. Stroom JC, De Boer HCJ, Huizenga H, Visser AG. Inclusion of geometrical uncertainties in radiotherapy treatment planning by means of coverage probability. Int. J. Radiat. Oncol. Biol. Phys. 1999; 43(4):905–919.
57. Bel A, van Herk M, Lebesque J V. Target margins for random geometrical treatment uncertainties in conformal radiotherapy. Med. Phys. 1996; 23(9):1537–1545.
58. Van Herk M, Remeijer P, Rasch C, Lebesque J V. The probability of correct target dosage: Dose-population histograms for deriving treatment margins in radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 2000; 47(4):1121–1135.
59. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 60. van Herk M, Witte M, van der Geer J et al. Biologic and physical fractionation effects of random
geometric errors. Int. J. Radiat. Oncol. 2003; 57(5):1460–1471. 61. Langen KM, Jones DTL. Organ motion and its management. Int. J. Radiat. Oncol. Biol. Phys. 2001;
50(1):265–278. 62. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood
cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
63. Dawson LA, Jaffray DA. Advances in Image-Guided Radiation Therapy. J. Clin. Oncol. 2007; 25(8):938–946.
64. van Herk M. Different Styles of Image-Guided Radiotherapy. Semin. Radiat. Oncol. 2007; 17(4):258–267.
65. Engelsman M, Schwarz M, Dong L. Physics Controversies in Proton Therapy. Semin. Radiat. Oncol. 2013; 23(2):88–96.
66. Timmermann B. Proton Beam Therapy for Childhood Malignancies: Status Report. Klin. Pädiatrie 2010; 222(03):127–133.
PART 1 Interfractional position variation
Chapter 2 Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: a multicenter study
Sophie C. Huijskens, Irma W.E.M. van Dijk, Rianne de Jong, Jorrit Visser, Raquel Dávila
Fajardo, Cécile M. Ronckers, Geert O.R.J. Janssens, John H. Maduro, Coen R.N. Rasch, Tanja
Alderliesten, Arjan Bel
Radiotherapy and Oncology 2015; Volume 117 (3): 425-431.
Doi: 10.1016/j.radonc.2015.09.020
23
24
Abstract
Purpose: To quantify renal and diaphragmatic interfractional motion in order to estimate systematic and random errors, and to investigate the correlation between interfractional motion and patient-specific factors.
Methods: We used 527 retrospective abdominal-thoracic cone beam CT scans of 39 childhood cancer patients (<18 years) to quantify renal motion relative to bony anatomy in the left-right (LR), cranio-caudal (CC) and anterior-posterior (AP) directions, and diaphragmatic motion in the CC direction only. Interfractional motion was quantified by distributions of systematic and random errors in each direction (standard deviations Σ and σ, respectively). Also, correlation between organ motion and height was analyzed.
Results: Inter-patient organ motion varied widely, with the largest movements in the CC direction. Values of Σ in LR, CC, and AP directions were 1.1, 3.8, 2.1 mm for the right, and 1.3, 3.0, 1.5 mm for the left kidney, respectively. The σ in these three directions was 1.1, 3.1, 1.7 mm for the right, and 1.2, 2.9, 2.1 mm for the left kidney, respectively. For the diaphragm we estimated Σ = 5.2 mm and σ = 4.0 mm. No correlations were found between organ motion and height.
Conclusions: The large inter-patient organ motion variations and the lack of correlation between motion and patient-related factors, suggest that individualized margin approaches might be required.
25
2.1 | Introduction
Continuous developments of more effective multimodality treatment strategies for pediatric cancer have led to a steep increase in the number of childhood cancer survivors during the last decades [1]. Inextricably, with the enhanced cancer survival, the incidence of treatment-related adverse events has become evident. Particularly, treatment with radiation therapy (RT) significantly contributes to the risk of developing adverse events [2, 3]. This underscores the need for extremely high accuracy in treatment planning and actual dose delivery to the target volume, to minimize dose to surrounding healthy tissues.
Organ motion is a well-known phenomenon that limits the maximum achievable accuracy. Organ motion is induced by respiration (intrafractional motion) or day-to-day variety of anatomical deviations (interfractional motion) related to e.g., variations in bowel- and bladder filling, tumor size, patient weight and treatment-induced tissue changes [4].
To account for treatment uncertainties, of which organ motion is the most challenging to deal with, the clinical tumor volume (CTV) is extended with an isotropic margin, thereby defining the planning target volume (PTV) [5]. Similar margin definitions are also taken into consideration for organs at risk (OAR) to define adequate planning risk volumes (PRV) [5]. The margin size is based on the systematic and random errors [6, 7] and has extensively been studied for adults [8–10]. A few studies have reported on pediatric organ motion, mainly focusing on cranial RT [11–15]. This treatment set-up allows for adequate patient immobilization. Moreover, intra-cranial organ motion is substantially smaller than that in thoracic-abdominal organs. Only two published studies quantified extra-cranial organ motion, with focus on intrafractional organ motion using 4D computed tomography (CT) and interfractional organ motion using Cone Beam CT (CBCT) [16, 17]. Panandiker et al. found a correlation between patient-specific factors including age and height and intrafractional renal motion [17]. No reports are available on relations between these factors and interfractional organ motion. To date, no clear guidelines are available for defining appropriate margins for pediatric abdominal RT. Children seem to differ from adults with respect to organ motion, and margins currently used may not be optimal in pediatric RT. Therefore, adequate data on organ motion in children is urgently needed in order to define patient-specific margins with the aim to decrease irradiated volume and minimize dose to healthy tissue thus reducing the risk of developing adverse events [2, 3].
Image-guided RT (IGRT) enables quantification of interfractional organ motion, using registration of CBCTs to the reference CT. Since the kidneys are representative for abdominal organ motion and kidneys are also considered as the primary dose-limiting organs for upper abdomen radiotherapy [18], here we focus on renal motion. To correlate interfractional renal and diaphragmatic motion we also analyzed the diaphragm.
In this study, our goal was to quantify renal and diaphragmatic interfractional motion in order to estimate the systematic and random errors and to investigate the possible correlation between interfractional motion and patient-specific factors.
26
2.2 | Methods
Patient population This retrospective multicenter cohort consisted of 39 patients, treated between 2006 and 2014, for whom a reference CT and at least 5 CBCT images of the upper abdomen and/or thorax were available. The median age at diagnosis was 8 years (range 1.6-17.8 years). The median height at diagnosis was 124 cm (range 77-184 cm). The most frequent cancer diagnoses among children younger than 8 years were neuroblastoma (n=11) and Wilms’ tumor (n=5). In children older than 8 years, Ewing sarcoma (n=7) was the most prevalent malignancy. Most (37/39) patients were treated in the supine position and 2 patients were treated under general anesthesia. A full overview of patient characteristics and treatment details, including tumor indication, number of CBCTs, and treatment location is listed in Table 2.1.
Reference CT and CBCT imaging For each patient, a pretreatment CT scan (120 kV, 2.5- or 5 mm slice thickness) was obtained according to institution-based standard protocols for planning purposes. This CT scan was considered as the reference for organ position (refCT). According to institution-based protocols, all patients received half-fan width CBCT imaging during their treatment (Synergy, Elekta Oncology systems, Crawly, UK). The timeframe varied between approximately 35-60 seconds and the degree of circumferential rotation of the CBCT image was 200 or 360 degrees. For each patient, a CBCT was acquired at the first three treatment fractions, followed by daily or weekly CBCTs throughout the treatment depending on the tumor type and its corresponding, institution dependent, treatment protocol. CBCTs were either used for offline verification protocols or enabled direct set-up verification through online registration with the refCT. In this study, we retrospectively analyzed the imaging data, including a total of 39 refCTs and 527 CBCTs, ranging from 5 to 33 CBCTs per patient. For some patients one or both kidneys or diaphragm were not evaluable (i.e., organ not included in image, due to nephrectomy, or insufficiently visible due to artifacts) on the CBCT, resulting in 25 right kidneys (291 CBCTs), 28 left kidneys (331 CBCTs), and 32 diaphragms (422 CBCTs) included in our analyses (Table 2.1).
27
Table 2.1 | Patient characteristics and treatment details. No.
Institute Sex Tumor Age at diagnosis
(years)
Height (cm)
Weight (kg) No. of CBCTs RT location Immobilizationa
Kidneys Diaphragm
1 AMC F Medulloblastoma 12.5 156 37 5 5 Spinal cordb NO 2 AMC M Pineal germinoma 13.4 166 48 12 6 Spinal cordb NO
3c d AMC M Medulloblastoma 7.6 126 24 6 6 Spinal cordb NO 4 AMC F Medulloblastoma 8.3 139 28 7 7 Spinal cordb NO 5 AMC F Medulloblastoma 8.1 130 24 5 - Spinal cordb NO 6 AMC M Pineal germinoma 13.3 175 58 5 - Spinal cordb NO 7d AMC M Pineal germinoma 15.8 159 48 8 8 Spinal cordb NO 8 AMC F B-cel lymphoma 16.9 163 62 - 18 Thorax NO 9 AMC F Ewing sarcoma 15.8 173 50 - 8 Thorax NO
10 AMC M Wilms' tumor 15.6 167 67 7e 7 Abdomen NO 11 AMC M Ewing sarcoma 14.9 177 65 9e 9 Spinal cord NO 12 AMC F Neuroblastoma 9.6 120 23 7 7 Thorax NO 13 AMC M DSRCT 9.9 137 26 10 5 Abdomen NO 14 AMC M PPB 4.0 106 17 - 12 Thorax NO 15 AMC M Ewing sarcoma 17.8 182 81 30 30 Spinal cord NO 16 AMC M Ewing sarcoma 15.0 157 59 - 25 Thorax NO 17 AMC M Ewing sarcoma 16.8 184 62 8e 8 Abdomen NO 18 AMC F Neuroblastoma 5.3 115 24 7 7 Abdomen NO 19 AMC F Hodgkin lymphoma 15.9 163 51 - 6 Thorax NO 20 AMC M EMRS 3.3 106 17 23 23 Abdomen NO 21c UMCN M Wilms' tumorg 3.1 102 17 16e 10 Abdomen NO 22 UMCN F Wilms' tumorg 6.6 119 19 14f - Abdomen NO 23 UMCN M Wilms' tumorg 3.4 98 15 6e 6 Abdomen YES 24 UMCN M Wilms' tumorg 4.5 108 16 6f - Abdomen YES 25 UMCN F Wilms' tumorg 7.9 135 28 14f 6 Abdomen YES 26 UMCN M Ewing sarcoma 15.4 177 65 - 33 Thorax YES 27 UMCN M Ewing sarcoma 6.6 124 23 - 30 Thorax YES 28 UMCN M Ewing sarcoma 15.2 159 54 - 30 Thorax NO 29 UMCN M Neuroblastoma 2.2 99 16 11 11 Abdomen YES 30 UMCN M Neuroblastoma 4.5 108 16 20 20 Abdomen YES 31 UMCN F Neuroblastoma 4.8 117 22 12 12 Abdomen YES 32 UMCN M Neuroblastoma 1.6 77 11 11 11 Abdomen YES 33 UMCN M Neuroblastoma 4.1 107 17 20 20 Abdomen YES 34 UMCN M Neuroblastoma 5.3 112 17 12 12 Abdomen YES 35 UMCN M Neuroblastoma 3.7 105 17 12 12 Abdomen YES 36 UMCN F Neuroblastoma 4.5 108 16 12 12 Abdomen YES 37 UMCG F Sarcoma 3.9 99 15 28e - Abdomen YES 38 UMCG F Neuroblastoma 5.7 115 19 12 - Abdomen YES 39 UMCG M Neuroblastoma 4.5 101 14 10 - Abdomen YES
Abbreviations: AMC = Academic Medical Center; UMCN = University Medical Center Nijmegen; UMCG = University Medical Center Groningen; M = male; F = female; DSRCT = Desmoplastic small round cell tumor; PPB = Pleuropulmonairy blastoma; EMRS = Embryonal Rhabdomyosarcoma a For immobilization a vacuum matrass was used b Spinal cord was part of craniospinal irradiation c Patient 3 and 21 were treated under general anesthesia d Patient 3 and 7 are treated in prone position e Only left kidney evaluable on CBCTs f Only right kidney evaluable on CBCTs g The affected kidney in Wilms’ tumor patients was surgically removed
28
Image registration We collected and stored the imaging data in a centralized database for image analysis. We used Elekta X-ray Volume Imaging (XVI) software (version 4.5; Elekta Oncology Systems) for a two-step rigid registration performed by a single experienced observer. First, a region of interest (ROI) was defined encompassing the vertebral column representative for the chosen area closest to the kidney. The CBCTs were then registered with the refCT using the automatic chamfer match algorithm. This first match of the bony anatomy of each CBCT enabled a consistent quantification of the organs with respect to the bony anatomy, irrespective of daily positioning variations. The kidneys were, prior to registration, delineated as separate ROIs including the whole kidney volume. This enabled to obtain a resolution greater than the slice thickness acquired in the refCT. Second, the kidney was automatically registered using the shaped ROI and a grey value algorithm. Results of the automatic procedure were inspected by the observer and when automatic registration failed, for example due to artifacts, manual registration was conducted (n=81/622 registrations). Registrations resulted in interfractional deviations relative to bony anatomy, expressed as composite vectors in the left-right (LR), cranio-caudal (CC) and anterior-posterior (AP) directions, for the right and left kidney separately. The + and – signs respectively indicate right/caudal/posterior and left/cranio/anterior directions. Registration of the diaphragm was only feasible in the CC direction. The observer registered each CBCT manually. The anterior-posterior plane used for diaphragm registration in each CBCT was chosen at the position of maximum kidney length. Measurements were corrected for rotations by assessing the center of mass (COM) coordinates for each kidney. We equated these coordinates to the refCT to determine the exact distance and direction of the interfractional deviations. Statistical data analysis For each patient, we determined organ specific mean, absolute median and standard deviation (SD) of the interfractional motion relative to bony anatomy for each direction. Inter-patient variation was analyzed by considering the spread of all patients’ interfractional kidney motion in the three orthogonal directions. To investigate the equality of variance between patients in each direction and studied organs, we tested the variance of each single patient versus the overall group variance using an F-test (p < 0.01 was considered as statistically significant). For the patient group as a whole, the results of the interfractional motion are expressed as the group mean, the group systematic error (Σ; the SD of the individual means of all patients) and the group random error (σ; root mean square value of the individual SDs) [6], for each direction. Linear regression analysis We used linear regression analysis to assess the relationship between renal and diaphragmatic motion and patient-specific factors (including age and height) and to investigate the possible correlation between renal and diaphragmatic motion (p < 0.05 was considered as statistically significant). We also divided the patient cohort in two subgroups based on median values and analyzed the effects of age and height on the systematic and random error. Since tumors had been surgically removed before radiation treatment, we have not investigated a possible relation between organ motion and tumor size. All statistical analyses were done using R version 3.2.1. (R Foundation for Statistical Computing, USA).
29
Table 2.2 | Group mean, the group systematic (Σ) and group random errors (σ) in mm in the orthogonal directions for the right- and left kidney and in CC direction for the diaphragm based on 527 CBCTs from 39 children.
(mm)
Right Kidney Left Kidney Diaphragm
LR CC AP LR CC AP CC
Group mean 0.6 0.5 0 -0.6 1.5 0 1.1
Σ 1.1 3.8 2.1 1.3 3.0 1.5 5.2
σ 1.1 3.1 1.7 1.2 2.9 2.1 4.0
Abbreviations: LR = Left-Right; CC = Cranial-Caudal; AP = Anterior-Posterior
Figure 2.1 | Median interfractional motion in children in the studied organs. Adult data derived from [8]. Significant differences were found in CC and AP directions (see discussion). Abbreviations: LR = Left-Right; CC = Cranial-Caudal; AP = Anterior-Posterior.
30
2.3 | Results The mean interfractional motion of the kidneys was smaller than 1.5 mm in our study population (Table 2.2). Figure 2.1 shows absolute median interfractional organ motion in children. The highest median values and largest spreads in interfractional motion in both kidneys were found in the CC direction (Figure 2.2), smaller spreads were found in the LR and AP directions. Interfractional renal motion varied largely between patients (Figure 2.2). In all directions and organs, multiple patients’ (2-5) variance differed significantly (p < 0.01) from the mean group variance. Linear regression analysis Since age and height were highly correlated (R2 > 0.9, graphs not shown) and kidney motion was largest in CC direction, we analyzed only height in relation to organ motion in CC direction. Linear regression analysis did not show a correlation between random errors of renal and diaphragmatic motion in the CC direction and height (Figure 2.3). Significant (p < 0.05), but weak correlations were found between interfractional kidney motion and diaphragmatic motion, except for the LR direction in the right kidney (Figure 2.4). The systematic and random errors in the two subgroups, based on median height, differed less than 1.5 mm.
31
Figure 2.2 | Boxplots for the interfractional motion relative to the refCT for all 39 patients in the 3 orthogonal directions, LR, CC, and AP, for the right (left column) and left (right column) kidney. The zero-line distinguishes the opposite directions, wherein the + and – signs respectively indicate right/caudal/posterior and left/cranial/anterior directions. Boxes: median value and upper and lower quartiles; whiskers: lowest and highest data point within 1.5 x interquartile range; circles: outliers. For some patients, one or both kidneys was not evaluable. See Table 2.1 for patient- and treatment-specific details. Abbreviations: LR = Left-Right; CC = Cranial-Caudal; AP = Anterior-Posterior.
32
Figure 2.3 | Scatter plots with regression lines describing relationships between random errors of renal and diaphragmatic interfractional motion in the CC direction (each point represents the random error for a single patient) and height.
Figure 2.4 | Scatterplots with regression lines of the linear regression analyses describing relationships between renal (top row: right kidney, bottom row: left kidney) interfractional motion in the three orthogonal directions and diaphragmatic interfractional motion in the CC direction. Abbreviations: LR = Left-Right; CC = Cranial-Caudal; AP = Anterior-Posterior.
33
2.4 | Discussion We quantified renal and diaphragmatic interfractional motion in 39 children using daily or weekly CBCTs and we calculated systematic and random errors. Our results show that kidney motion was largest in the CC direction, suggesting that margins should be applied anisotropically rather than isotropically. Moreover, a large variation in mean interfractional deviations and a wide distribution of random errors was found between patients. Therefore, our findings suggest an individualized approach to define appropriate margins in abdominal pediatric RT.
The large variation in interfractional organ motion between patients also suggests that the selection of the reference scan for estimating organ motion might be non-trivial. The refCT yields a short acquisition time over repeatedly changing anatomy during respiration cycles, which might not be the most representative as reference situation [19]. A CBCT has a slower gantry rotation and averages the motion over the observed breathing phases into one blurred 3D image. The timeframe of 35-60 seconds ascertains that the CBCTs show an average of the full range of motion. Therefore, we considered an alternative by calculating the mean interfractional motion based on the first CBCT as a surrogate of the reference scan for all patients. It turned out that differences between the respective calculations based on the refCT and the first CBCT were less than 1 mm, which is practically negligible (data not shown).
For younger and/or more mobile children, institution-based protocols determined whether or not to use a vacuum matrass for immobilization. Using a vacuum matrass might affect the breathing pattern and influence the baseline organ position. We did not analyze the effect of immobilization, since the interfractional organ motion with respect to the bony anatomy, as quantified in our study, is not affected by the immobilization systems used.
In one of the few studies published on extra-cranial pediatric organ motion, Nazmy et al. used the CBCTs of 9 abdominal neuroblastoma patients (mean age 4.1 ±1.6 years) to analyze renal interfractional motion relative to bony anatomy [16]. When we analyzed the 10 neuroblastoma patients (mean age 4.6 ±2.2 years) in our cohort separately, we found a slightly smaller range in interfractional motion in the CC direction (right kidney, range -1 to 8 mm vs. -4 to 10 mm; left kidney, range -4 to 5 mm vs. -4 to 8 mm). As point of interest we used the kidney COM, because it is less sensitive to kidney deformation, whereas Nazmy et al. used the upper pole of the kidney. Due to this method, Nazmy et al. might interpreted kidney deformations as translations, yielding an overestimation of the kidney motion. This is in line with the slightly smaller deviations that we found. Interestingly, all patients included in the neuroblastoma study [16] were treated under general anesthesia (GA), whereas GA was not used in any of the neuroblastoma patients in our study. Instead, a large proportion of young children in our cohort were positioned in a vacuum matrass and treated during their regular nap time as a means to establish immobilization.
Panandiker et al. studied extra-cranial pediatric organ motion using 4DCTs, in which respiratory cycles were captured as series of 8 phases [17]. They focused on renal intrafractional motion in 20 patients (median age 8 years, range 2-18 years), of whom 10 were treated under GA. The authors used an arbitrary baseline in 3D space to define organ motion, rather than bony anatomy as we did in our study. In accordance with our methodology, they used the geometric COM as primary calculation reference point. They reported a relationship between intrafractional kidney motion and age. Unexpectedly, we did not find such relationship for interfractional kidney motion with age. In accordance with Panandiker et al., both renal interfractional motion correlated with diaphragmatic motion. However, Panandiker et al. studied intrafractional motion mainly resulting from respiration,
34
while interfractional motion, as quantified in our study, results also from other anatomical variations, such as inconstant bowel filling. These differences in inter- and intrafractional motion need to be further investigated.
Although we did not find a correlation between interfractional organ motion and height, we found a suggestive difference in renal and diaphragmatic motion between children and adults, when comparing our results with the literature. Wysocka et al. determined the median interfractional motion of the kidneys and diaphragm relative to bony anatomy in 22 adults [8]. However, instead of CBCTs, follow-up CT scans were used for registrations. Since a CT has a short acquisition time, organ position strongly depends on the breathing phase during that time, which could lead to more extreme deviations when compared to CBCT-based analysis. We compared kidney motion under free breathing as reported by Wysocka et al. with our data. To compare the median interfractional motion, we calculated the median values as well, and in accordance with their study, we found the largest displacements in the CC direction (Figure 2.1). The median interfractional displacements of the kidneys and diaphragm were seemingly smaller in the children in our cohort than in their study. This indicates that abdominal interfractional organ motion in children may be different from that in adults, although some caution in interpretation is warranted given the above-stated differences in methodology.
Strengths of our study are the large patient number and the substantial amount of CBCTs (mean 14, range 5-33 per patient). With a range of 5-33, the number of CBCTs differed substantially between patients. Analysis and recalculations in which we weighted the data with the number of CBCTs showed no significant changes. We are the first to focus on the quantification of interfractional renal motion using a large amount of imaging data, acquired during IGRT. This allowed for a good estimation of the group systematic (Σ) and random errors (σ). Subsequently, these values can be used in a margin recipe, such as 2.5 Σ + 0.7 σ [6], to estimate an appropriate CTV-PTV margin. PRV margins can be calculated using 1.3 Σ + 0.5 σ [7]. As a note of caution, our study included only one component of the margin, i.e., the interfractional organ motion, and did not consider other effects (such as delineation uncertainties or set-up variation). Moreover, intrafractional abdominal motion due to respiration during treatment should also be investigated. The retrospective (daily or weekly) CBCTs used in this study can also be used for the quantification of intrafractional abdominal motion [20]. In order to properly calculate the margins, all components of organ motion should be taken into consideration. A better understanding of abdominal organ motion in children during RT is also essential to take full advantage of the state-of-the-art treatment approaches for children, such as intensity modulated RT and intensity modulated proton therapy.
2.5 | Conclusions
Renal and diaphragmatic interfractional motion was largest in the CC direction. Interfractional
motion did not correlate with patient-specific factors and variation between patients was large. This suggests that individualized margin approaches might be required. The outcomes of this study are a first step towards guidelines for pediatric abdominal margin definitions. To provide a full insight into renal and diaphragmatic motion, inter- and intrafractional motion should be further analyzed. Such comprehensive insight into pediatric abdominal organ motion is essential for high accuracy in hitting the target and minimizing dose to healthy tissue.
35
References 1. Curry HL, Parkes SE, Powell JE, Mann JR. Caring for survivors of childhood cancers: The size of
the problem. Eur. J. Cancer 2006; 42(4):501–508. 2. Geenen MM, Cardous-Ubbink MC, Kremer LCM et al. Medical Assessment of Adverse Health
Outcomes in Long-term Survivors of Childhood Cancer. JAMA 2007; 297(24):2705. 3. van Dijk IWEM, Oldenburger F, Cardous-Ubbink MC et al. Evaluation of Late Adverse Events in
Long-Term Wilms’ Tumor Survivors. Int. J. Radiat. Oncol. 2010; 78(2):370–378. 4. Lutkenhaus LJ, Kamphuis M, van Wieringen N et al. Reduction in cardiac volume during
chemoradiotherapy for patients with esophageal cancer. Radiother. Oncol. 2013; 109(2):200–203.
5. Prescribing, Recording and Reporting Photon Beam Therapy (Supplement to ICRU Report 50), ICRU Report 62 . ICRU
6. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 7. McKenzie A, van Herk M, Mijnheer B. Margins for geometric uncertainty around organs at risk
in radiotherapy. Radiother. Oncol. 2002; 62(3):299–307. 8. Wysocka B, Kassam Z, Lockwood G et al. Interfraction and Respiratory Organ Motion During
Conformal Radiotherapy in Gastric Cancer. Int. J. Radiat. Oncol. 2010; 77(1):53–59. 9. Pham D, Kron T, Foroudi F et al. A Review of Kidney Motion under Free, Deep and Forced-
Shallow Breathing Conditions: Implications for Stereotactic Ablative Body Radiotherapy Treatment. Technol. Cancer Res. Treat. 2014; 13(4):315–323.
10. van der Horst A, Wognum S, Dávila Fajardo R et al. Interfractional Position Variation of Pancreatic Tumors Quantified Using Intratumoral Fiducial Markers and Daily Cone Beam Computed Tomography. Int. J. Radiat. Oncol. 2013; 87(1):202–208.
11. Beltran C, Krasin MJ, Merchant TE. Inter- and intrafractional positional uncertainties in pediatric radiotherapy patients with brain and head and neck tumors. Int. J. Radiat. Oncol. Biol. Phys. 2011; 79(4):1266–1274.
13. Beltran C, Naik M, Merchant TE. Dosimetric effect of setup motion and target volume margin reduction in pediatric ependymoma. Radiother. Oncol. 2010; 96(2):216–222.
14. Beltran C, Naik M, Merchant TE. Dosimetric effect of target expansion and setup uncertainty during radiation therapy in pediatric craniopharyngioma. Radiother. Oncol. 2010; 97(3):399–403.
15. Zhu Y, Stovall J, Butler L et al. Comparison of two immobilization techniques using portal film and digitally reconstructed radiographs for pediatric patients with brain tumors. Int. J. Radiat. Oncol. 2000; 48(4):1233–1240.
16. Nazmy MS, Khafaga Y, Mousa A, Khalil E. Cone beam CT for organs motion evaluation in pediatric abdominal neuroblastoma. Radiother. Oncol. 2012; 102(3):388–392.
17. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
18. Dawson LA, Kavanagh BD, Paulino AC et al. Radiation-Associated Kidney Injury. Int. J. Radiat. Oncol. 2010; 76(3):S108–S115.
19. Balter JM, Ten Haken RK, Lawrence TS et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int. J. Radiat. Oncol. 1996; 36(1):167–174.
20. Lens E, Van Der Horst A, Kroon PS et al. Differences in respiratory-induced pancreatic tumor motion between 4D treatment planning CT and daily cone beam CT, measured using intratumoral fiducials. Acta Oncol. (Madr). 2014; 53(9):1257–1264.
Chapter 3 Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference?
Irma W.E.M. van Dijk*, Sophie C. Huijskens*, Rianne de Jong, Jorrit Visser, Raquel Dávila
Purpose: Pediatric safety margins are generally based on data from adult studies; however, adult-based margins might be too large for children. The aim of this study was to quantify and compare interfractional organ position variation in children and adults. Methods: For 35 children and 35 adults treated with thoracic/abdominal irradiation, 850 (range 5–30 per patient) retrospectively collected cone beam CT images were registered to the reference CT that was used for radiation treatment planning purposes. Renal position variation was assessed in three orthogonal directions and summarized as 3D vector lengths. Diaphragmatic position variation was assessed in the cranial-caudal (CC) direction only. We calculated means and SDs to estimate group systematic (Σ) and random errors (σ) of organ position variation. Finally, we investigated possible correlations between organ position variation and patients’ height. Results: Interfractional organ position variation was different in children and adults. Median 3D right and left kidney vector lengths were significantly smaller in children than in adults (2.8, 2.9 mm vs. 5.6, 5.2 mm, respectively; p<.05). Generally, the pediatric Σ and σ were significantly smaller than in adults (p<.007). Overall and within both subgroups, organ position variation and patients’ height were only negligibly correlated. Conclusions: Interfractional renal and diaphragmatic position variation in children is smaller than in adults indicating that pediatric margins should be defined differently from adult margins. Underlying mechanisms and other components of geometrical uncertainties need further investigation to explain differences and to appropriately define pediatric safety margins.
39
3.1 | Introduction As a result of improved treatment modalities, long-term childhood cancer survival has increased remarkably with five year survival rates approaching 80% [1]. Concurrently, increased survival is accompanied by late treatment-related adverse events, especially when radiotherapy has been part of the treatment [2–4]. Developing the optimal radiation treatment plan involves a delicate tradeoff between tumor coverage on the one hand and sparing surrounding healthy tissues (organs at risk; OARs) on the other hand. The radiation treatment volume is determined by the planning target volume (PTV), which is an expansion of the clinical target volume (CTV), thus accounting for geometrical uncertainties due to daily anatomical variations (interfractional organ position variation), breathing (intrafraction motion) and delineation errors [5]. All these components are considered when calculating PTV safety margins using the group systematic (Σ) and the group random error (σ) [6]. Likewise, planning-organ-at-risk volume (PRV) margins can be calculated to avoid OARs [7].
Since pediatric cancer is a rare disease, and data on organ motion in children is scarce [8], safety margins as currently used in pediatric radiation treatment are mainly based on margins as defined for adults. Abdominal organ position variation and its impact on margin size for different tumor sites have so far only been reported for adult patients [9–14] and pediatric patients separately [15–17]. In our previous study, we quantified renal and diaphragmatic interfractional motion in children [17]. Comparison of our results with those reported in literature [12] showed that organ motion was seemingly smaller in our pediatric cohort than in adults [17]. This suggests that adult-based safety margins might be too large for pediatric patients. However, to date no comparative studies have been performed in a cohort including both children and adults. In this study, we aim to quantify interfractional renal and diaphragmatic position variation in children and adults, thus enabling a straightforward comparison of results. 3.2 | Methods and Materials Patient population For this retrospective study, we included children (<18 years) and adults (≥18 years) who had been treated at the AMC radiation oncology department between October 2010 and December 2014. Patients were included if a pretreatment CT scan and at least five cone beam CT scans (CBCTs) of the upper abdomen and/or thorax were available for registration; five registrations were considered to enhance the reliability of the statistical analyses [18]. Thirty-five children who met these criteria had been had been diagnosed with a variety of pediatric tumors (Table 3.1); a part of these data (19/35 children) have been analyzed previously [17]. Based on the inclusion criteria, we randomly selected 35 adults who had been treated in the same time period for esophageal, gastric and pancreatic tumors. We collected information on general anesthesia, and patient characteristics including age at the first radiation treatment fraction, height, weight, primary cancer diagnosis and radiation site.
40
Table 3.1 | Characteristics of included patients Children N = 35
(%) Adults N = 35
(%) Gender Male Female
21 14
(60) (40)
25 10
(71) (29)
Age at 1st RT fraction (years) Mean (median; range) 0-5 6-10 11-18 30-49 50-69 ≥70
10.3 (9.3; 3.1-17.8))
7 14 14
(20) (40) (40)
59.9 (61.0; 34.1-94.0)
8 21 6
(23) (60) (17)
Height (cm) Mean (median; range) a
140 (137; 92-184)
175 (175; 160-203)
Weight (kg) Mean (median; range) a
36.3 (28; 13.0-81.0)
72.3 (72; 50.0-96.0)
Type of primary cancer CNS tumorb
Sarcomac
Neuroblastoma Renal tumord
Othere
Oesophagus Pancreas Stomach
16 9 3 3 4
(46) (26) (9) (9)
(11)
10 13 12
(29) (37) (34)
Radiation Site (Cranio)spinal Thoracic/Mediastinal Abdominal (incl. flank) Upper abdominal
18 9 8
(51) (26) (23)
35
(100)
Total number of CBCTs 374 476 Mean (median; range) 10 (5-30) 14 (5-27)
Abbreviations: CNS = central nervous system aData with respect to age, height, and weight closely represent the normal distribution. bIncluding: anaplastic glioma (n=1), ependymoma (n=1), germinoma pinealis (n=4), medulloblastoma (n=10)
Reference CT and CBCT imaging According to tumor-based treatment protocols, pretreatment CT scans for planning purposes were acquired for all patients. These scans were considered the reference for organ position (i.e., refCT), and included 3D-CTs (for 60 patients) and 4D-CTs (for one child and nine adults) (LightSpeed RT16; General Electric Company, Waukesha, WI, USA). Slice thickness varied from 2.5 to 3.0 mm; 10 children had 3D-CTs with 5.0 mm slice thickness. The 10 breathing phases of the 4D-CT were averaged to simulate a 3D-CT. For all patients, CBCT images were routinely acquired for setup verification before radiation delivery (Synergy, Elekta Oncology Systems, Crawley, UK). In general for children, an accustomed extended no-action level (eNAL) protocol was used [19]. This yields CBCT imaging and online correction at the first three radiation treatment fractions after which the a-priori set-up correction is adjusted, to be checked at the fourth fraction. From the fifth fraction on the eNAL protocol is followed, acquiring weekly imaging, unless eNAL results exceed tolerance limits [19]. In adults treated for esophageal, gastric, and pancreatic tumors, daily CBCTs were acquired for online position verification. Pediatric CBCT acquisition time varied between 35 and 60 s, with a 200° or 360° rotation, respectively. The acquisition time for adult CBCT was 120 s with a 360° rotation. Image registration Imaging data were collected and stored in our database for image analysis. Elekta X-ray volume imaging (XVI) version 4.5 software (Elekta Oncology Systems, Stockholm, Sweden) was used for two-step rigid organ registrations as described previously [17]. First, after defining a region of interest (ROI) including six to seven vertebrae at the level of the diaphragm and the kidneys, the CBCT was registered to the refCT using the automatic chamfer match algorithm for bony anatomy to account for daily setup variations. Bony anatomy registration was then followed by registration of the left and right kidney separately, using distinct ROIs based on delineations including the whole kidney volume, enabling the assessment of smaller values of organ position variation (greater resolution) than the slice thickness acquired in the refCT. For each patient, renal position variation was assessed in the cranial-caudal (CC), left–right (LR) and anterior–posterior (AP) directions. Registration outcomes were visually evaluated and (manually) corrected if necessary. To correct interfractional position variation for rotations, we assessed the center of mass coordinates for the left and right kidney separately, and compared them to the refCT in order to determine the distance and direction of interfractional position variation. Diaphragm position variation was considered as a surrogate for abdominal organ position variation, and also assessed using the two-step rigid registration method. After the automatic registration of the vertebral column as described for renal registration, the diaphragm as one complete structure was manually registered in the CC direction only. Statistical analysis Per patient, the mean and SD of interfractional position variation were calculated in three directions for the kidneys, and in the CC direction only for the diaphragm. We calculated median 3D vector lengths as a summarized measure of renal position variation, and used the non-parametric test for independent samples to test for differences between children and adults (significance level p<.05). Subsequently, the group means, group systematic errors (Σ; the SD of patients’ individual means), and group random errors (σ; the root mean square of the individual SDs) for children and adults were
42
estimated [6]. To investigate the systematic set-up error for patient positioning, we used the one-sample t-test to test whether group means significantly differed from 0 (i.e., reference value of organ position in the refCT).
Since not all data fitted the normal distribution (tested with the Shapiro–Wilk’s test for normality), the non-parametric Levene’s test for equality of variance was used to analyze the difference of Σ. To test for the difference of σ, the Mann–Whitney U-test was used. This test was also used to investigate possible differences in organ position variation in children of similar ages treated with and without anesthesia. Since a mutual dependency between renal and diaphragmatic position variation can be assumed, and seven directions were tested (i.e., CC, LR, AP in both kidneys and CC only in the diaphragm), we adjusted p values using the Bonferroni correction to reduce the chance of finding significant differences by coincidence [20]. Differences were considered to be significant if test outcomes showed a p value<.007 (i.e., 0.05/7).
To investigate the possible correlation between renal and diaphragmatic position variation and patient height, Spearman’s rank correlation coefficients (ρ) were calculated for the whole cohort, for pediatric and adult patients separately, and for the children and adults with overlapping heights (significance level p<.05). Additionally, the correlation with BMI (weight (kg)/(height (m))2) was investigated. We did not include age in our analyses, since children and adults physically differ; adolescents can be taller than small adults.
Data were analyzed using statistical software SPSS version 22.0 for Windows (SPSS, INC, Chicago, IL, USA) and the R software package version 3.2.1 (R Foundation for Statistical Computing, Vienna, Austria).
43
Figure 3.1 | Boxplots showing the distributions of the individual means and SDs of the kidneys and diaphragm in the analyzed directions in children (blue) and adults (red). Boxes: median value and upper and lower quartiles; whiskers: lowest and highest data point within 1.5 x interquartile range; circles: outliers. *Significant differences (Bonferroni adjusted p<.007). CC: cranial-caudal; LR: left–right; AP: anterior–posterior.
44
3.3 | Results
Patient population The mean ages (range) of children and adults were 10.3 (3.1–17.8) and 59.9 (34.1–94.0) years, respectively (Table 3.1). Mean heights (range) in children and adults were 140 cm (92–184 cm) and 175 cm (160–203 cm), respectively. Thirty-three of 35 children and all adults were treated in supine head-first position. Anesthesia was required in five children (mean age 5.2 (range, 3.1–7.6) years; Figure 3.3); seven children of similar ages (mean 5.8 (range 3.3–7.9) years) were treated without anesthesia. Overall, 850 CBCT images were successfully registered (range 5–30 per patient). In some patients one or both kidneys and the diaphragm could not be registered due to various reasons (see Appendix 3). Organ position variation The upper panel of Figure 3.1 shows the distributions of the individual means and the group means in children and adults for all organs in all directions. The one-sample t-test showed that the group means (Table 3.2) were not significantly different from the reference value 0 (adjusted p>.007), meaning that the systematic set-up error for patient positioning was minimal. The lower panel of Figure 3.1 shows the distributions of the individual SDs in the two patient groups. In both groups, renal position variation was largest in the CC direction. The median 3D vector lengths of the right and left kidney were significantly smaller in children as compared to those in adults (Figure 3.2: 2.8, 2.9 mm vs. 5.6, 5.2 mm, respectively; p<.05).
Table 3.2 presents the values of Σ and σ. For children, Σ values were significantly smaller in the CC direction of the right kidney, the CC and AP directions of the left kidney, and the CC direction of the diaphragm when compared to adults (adjusted p<.007). Pediatric σ values were significantly smaller than those for adults for all directions of the right kidney, and the CC and AP directions of the left kidney (adjusted p<.007). Organ position variation in the five children treated with anesthesia was not significantly different compared to seven children of similar age ranges (3.1–7.6 and 3.3–7.9 years, respectively) who had no anesthesia (adjusted p>.007). Correlation between organ position variation and patient height and BMI Correlations between interfractional organ position variation and height were negligible, taking all patients into account (Figure 3.3). Spearman’s ρ varied from 0.277 (correlation between the CC right kidney position variation and height) to 0.393 (correlation between AP right kidney position variation and height). Spearman’s ρ for the pediatric and adult patients separately indicated only negligible, non-significant correlations (Supplementary Figure 3.1(A,B)). Likewise, correlations between interfractional organ position variation and height in the subgroup of children and adults with overlapping heights were negligible and non-significant (Supplementary Figure 3.2). Finally, correlations with BMI were also found to be negligible, though significant (Supplementary Figure 3.3).
45
Figure 3.2 | Boxplots of 3D vector lengths of interfractional renal position variation in children (blue) and adults (red). Boxes: median value and upper and lower quartiles; whiskers: lowest and highest data point within 1.5 x interquartile range; circles: outliers. * Significant differences (p < .05).
46
Table 3.2 | The group mean, group systematic (Σ) and group random error (σ) for children and adults.
Mean (mm)
Right kidney Left kidney Diaphragm
CC LR AP CC LR AP CC
Children 0.7 -0.1 -0.4 1.0 -0.4 -0.9 -0.8
Adults 2.0 0.7 -0.6 -0.4 -0.1 -3.3 2.1
Systematic error Σ (mm)
Right kidney Left kidney Diaphragm
CC LR AP CC LR AP CC
Children 3.6 1.3 1.8 3.1 1.4 0.9 3.4
Adults 6.6 1.9 3.2 6.5 2.1 2.2 9.9
Random error σ (mm)
Right kidney Left kidney Diaphragm
CC LR AP CC LR AP CC
Children 2.9 1.0 1.5 2.5 1.5 2.0 3.7
Adults 5.0 1.5 2.9 5.0 1.5 2.1 4.9
Abbreviations: CC = cranial-caudal, LR = left-right, AP = anterior-posterior
Note: The calculations of the group systematic (Σ) and random error (σ) are based on the distributions of the individual means and SDs as presented in Figure 3.1.
47
Figure 3.3 | Distributions of the individual random errors (σ) of interfractional renal and diaphragmatic position variation in the CC, LR and AP directions. Spearman’s ρ indicate correlations between organ position variation and patients’ height. Dots (blue), triangles (dark blue) and circles (red) represent, respectively, pediatric patients treated without and with anesthesia, and adult patients. CC: cranial-caudal; LR: left–right; AP: anterior–posterior.
48
3.4 | Discussion Thus far, adult [9–14] and pediatric [15–17] studies separately reported on abdominal organ motion, which complicates a straightforward comparison of results due to different methodologies. This study is the first to quantify and compare interfractional organ position variation in a relatively large number of pediatric and adult cancer patients. Our results show that renal and diaphragmatic position variation was generally smaller in children than in adults; the median 3D vector lengths of the right and left kidney were significantly smaller in children as compared to adults.
Diaphragm position variation was assessed by manual registration of the diaphragm as one complete structure, which forced us to make concessions regarding the best fit. Due to different anatomy of the left and right upper abdomen, with the spleen in the upper far left part of the abdomen and the largest part of the liver in the upper right part of the abdomen, it is conceivable that the position variation of the left and right diaphragm domes could be different. Considering the left and right diaphragm dome position variation as a surrogate for, respectively, left and right sided abdominal organ position variation, might, therefore, lead to the need of distinct definitions of PTV and PRV margins for left and right sided tumors and OARs. In a future study, therefore, it would be worthwhile to investigate the diaphragm position variation considering the left and right diaphragm domes separately. Further, the findings thereof could be compared to the findings for the left and right kidney position variations.
Since interfractional organ position variation was investigated relative to the bony anatomy, it is not to be expected that treatment position variations caused the differences between children and adults. Daily anatomical variations due to differences in organ filling and the amount of air in stomach or bowel were present in the patients who showed extreme values of organ position variation (Figure 3.1). When we repeated our analysis excluding these outliers, the difference in mean position variation of the right kidney in LR direction in pediatrics vs. adults was found to be significant as well, whereas other results did not change.
In this study, we registered CBCTs with 3D- and averaged 4D-CTs. The acquisition times of pediatric and adult CBCTs were 35–60 and 120 s, respectively; both included a sufficient number of breathing cycles to ascertain an average of the full range of motion. This is comparable with an averaged 4D-CT, whereas a 3D-CT is acquired in a relatively shorter time frame than a CBCT. Therefore, a 3D-CT might not sufficiently represent the anatomy during a radiation treatment course [21]. However, differences in results are expected to be practically negligible; when we based our interfractional organ position variation calculations on the first CBCT instead of the refCT, registration outcomes were similar, as shown in our previous study on organ motion in children [17]. Some children had 3D-CTs with 5.0 mm slice thickness, which could have led to an overestimation of interfractional organ position variation in the CC-direction.
The differences in organ position variation resulted in generally smaller Σ and σ values for kidneys and diaphragm in children than in adults. The availability of a considerable amount of CBCTs (5–30 per patient) enhanced the reliability of our analyses, and allowed for good estimations of Σ and σ. When entering the Σ and σ values from Table 3.2 in the PTV or PRV margin recipes [6, 7], renal safety margins could be calculated, resulting in smaller PTV and PRV margins for children than for adults. However, such calculations are only valid for geometrical uncertainties due to interfractional organ position variation, without taking other components, such as intrafractional motion and delineation errors into account.
49
Pham et al. conducted a systematic review on the degree of intrafractional renal motion associated with various breathing conditions in children and adults. In concordance with our results, they summarized that under free breathing the mean right and left kidney motion in children was considerably smaller than in adults [10].
The use of general anesthesia to immobilize young children during radiotherapy varies. In our cohort five children needed general anesthesia. Organ position variations in this small group did not differ from outcomes in children of similar ages treated without anesthesia. Pai Panandiker et al. studied intrafractional renal motion in 20 pediatric patients and found that renal motion in patients undergoing anesthesia was smaller than in patients who did not need anesthesia. They also concluded that renal motion correlated with age and height [16]. In our previous study, it appeared that pediatric age and height were interdependent factors (correlation coefficient R2 > 0.9); therefore, only the correlation between organ position variation and height was investigated whereby no significant correlations were found for interfractional motion [17]. In this study, we explicitly chose to include height, and not age, in the correlation analyses since adolescents can be taller than small adults. We found only negligible correlations between organ position variation and height; the low ρ values indicate that only a small proportion of interfractional organ position variation can be attributed to patients’ height. This means that other, underlying mechanisms cause differences in organ position variation in children and adults. There is a substantial difference in age between the pediatric and adult group (i.e., pediatrics: maximum age, 17.8 years; adults: minimum age, 34.1 years). Different physique including stature, body fat and elasticity of tissues and organs of the (pediatric and adult) patients might affect organ position variation; for example, the extent of tissue and organ elasticity could have influenced organ deformation. These and other – yet unknown – factors might have hampered establishing a possible correlation between position variation and patient height in our cohort. In addition, although our study includes a relatively large number of patients, even larger cohorts might be needed to uncover with statistical significance correlations between organ position variation and patient characteristics.
Interfractional organ position variation is an important component of geometrical uncertainties. Our results show that smaller PTV margins could be applied in children. However, it is important to further investigate other components, such as intrafraction motion and delineation errors to generate an all-encompassing definition of pediatric safety margins.
3.5 | Conclusions
In conclusion, we showed that interfractional renal and diaphragmatic position variation in children is smaller than in adults. Underlying, yet unknown mechanisms need to be investigated to explain these differences. Nevertheless, these results indicate that pediatric safety margins should be defined differently from adult margins. Other components of geometrical uncertainties have to be investigated to appropriately define pediatric PTV margins.
50
References 1. Gatta G, Botta L, Rossi S et al. Childhood cancer survival in Europe 1999–2007: results of
EUROCARE-5—a population-based study. Lancet Oncol. 2014; 15(1):35–47. 2. Geenen MM, Cardous-Ubbink MC, Kremer LCM et al. Medical Assessment of Adverse Health
Outcomes in Long-term Survivors of Childhood Cancer. JAMA 2007; 297(24):2705. 3. Oeffinger KC, Mertens AC, Sklar C a et al. Chronic Health Conditions in Adult Survivors of
Childhood Cancer. N. Engl. J. Med. 2006; 355(15):1572–1582. 4. van Dijk IWEM, Oldenburger F, Cardous-Ubbink MC et al. Evaluation of Late Adverse Events in
138(1):146. 6. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 7. McKenzie A, van Herk M, Mijnheer B. Margins for geometric uncertainty around organs at risk
in radiotherapy. Radiother. Oncol. 2002; 62(3):299–307. 8. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood
cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
9. NamysŁ-Kaletka A, Wydmanski J, Tukiendorf A et al. Influence of interfraction motion on margins for radiotherapy of gastric cancer. Br. J. Radiol. 2015. doi:10.1259/bjr.20140610.
10. Pham D, Kron T, Foroudi F et al. A Review of Kidney Motion under Free, Deep and Forced-Shallow Breathing Conditions: Implications for Stereotactic Ablative Body Radiotherapy Treatment. Technol. Cancer Res. Treat. 2014; 13(4):315–323.
11. van der Horst A, Wognum S, Dávila Fajardo R et al. Interfractional Position Variation of Pancreatic Tumors Quantified Using Intratumoral Fiducial Markers and Daily Cone Beam Computed Tomography. Int. J. Radiat. Oncol. 2013; 87(1):202–208.
12. Wysocka B, Kassam Z, Lockwood G et al. Interfraction and Respiratory Organ Motion During Conformal Radiotherapy in Gastric Cancer. Int. J. Radiat. Oncol. 2010; 77(1):53–59.
13. Brandner ED, Wu A, Chen H et al. Abdominal organ motion measured using 4D CT. Int. J. Radiat. Oncol. Biol. Phys. 2006. doi:10.1016/j.ijrobp.2005.12.042.
14. Jin P, van der Horst A, de Jong R et al. Marker-based quantification of interfractional tumor position variation and the use of markers for setup verification in radiation therapy for esophageal cancer. Radiother. Oncol. 2015. doi:10.1016/j.radonc.2015.10.005.
15. Nazmy MS, Khafaga Y, Mousa A, Khalil E. Cone beam CT for organs motion evaluation in pediatric abdominal neuroblastoma. Radiother. Oncol. 2012; 102(3):388–392.
16. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
17. Huijskens SC, van Dijk IWEM, de Jong R et al. Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: A multicenter study. Radiother. Oncol. 2015; 117(3):425–431.
18. Smitsmans MHP, Pos FJ, de Bois J et al. The Influence of a Dietary Protocol on Cone Beam CT–Guided Radiotherapy for Prostate Cancer Patients. Int. J. Radiat. Oncol. 2008; 71(4):1279–1286.
19. de Boer HCJ, Heijmen BJM. eNAL: An Extension of the NAL Setup Correction Protocol for Effective Use of Weekly Follow-up Measurements. Int. J. Radiat. Oncol. 2007; 67(5):1586–1595.
20. Ashby D. Practical statistics for medical research. Douglas G. Altman, Chapman and Hall, London, 1991. No. of pages: 611. Price: £32.00. Stat. Med. 1991; 10(10):1635–1636.
21. Balter JM, Ten Haken RK, Lawrence TS et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int. J. Radiat. Oncol. 1996; 36(1):167–174.
51
Appendix 3
CBCT to CT organ registration
In some patients one or both kidneys or the diaphragm could not be registered. Reasons for non-
registration were for example nephrectomy in Wilms’ tumor patients, a small field of view (FOV) so
that the concerning organ was not included in the cone beam CT scans (CBCTs), or insufficient visibility
due to (breathing) artefacts. In a number of children treated with craniospinal irradiation, diaphragm
and kidneys were registered in separate thoracic and abdominal images, respectively. This resulted in
analyses of 24 right and 29 left kidneys (in 344 CBCTs), and 30 diaphragms (in 357 CBCTs) in the
children. For the adults we analyzed 32 right and 35 left kidneys, and 27 diaphragms in 476 CBCTs with
a FOV large enough to include both the diaphragm and kidneys.
52
Supp
lem
enta
ry T
able
3.1
. Med
ian,
IQR
and
rang
e (m
m) o
f org
an p
ositi
on v
aria
tion
in c
hild
ren
and
adul
ts
Ch
ildre
n
Adul
ts
M
edia
n (IQ
R)
(min
– m
ax)
M
edia
n (IQ
R)
(min
- m
ax)
Righ
t kid
ney
CC
-0.4
(-0
.7 –
2.8
) (-1
0.2
– 9.
1)
1.
7 (-2
.1 –
6.6
) (-1
3.3
– 14
.5)
LR
-0.1
(-5
.4 –
5.5
) (-4
.2 –
2.3
)
0.9
(-0.7
– 2
.3)
(-3.3
– 3
.5)
AP
-2.6
(-1
.5 -0
.4)
(-4.1
– 4
.3)
-0
.4
(-2.0
– 0
.9)
(-8.1
– 1
0.4)
Left
kid
ney
CC
1.3
(-1.7
– 3
.4)
(-4.8
– 7
.3)
0.
3 (-4
.3 –
3.9
) (-1
9.5
– 14
.7)
LR
-0.2
(-1
.4 –
0.6
) (-3
.5 –
1.9
)
-0.0
(-1
.4 –
1.3
) (-4
.3 –
4.2
)
AP
-0.2
(-1
.0 –
0.4
) (-2
1.1
– 1.
6)
-0
.2
(-1.9
– 1
.0)
(-4.9
– 6
.5)
Diap
hrag
m
CC
0.2
(-4.0
– 1
.7)
(-9.2
– 5
.6)
2.
3 (-4
.0 –
8.6
) (-1
8.3
– 26
.9)
Abbr
evia
tions
: IQ
R =
inte
rqua
rtile
rang
e, m
in =
min
imum
, max
= m
axim
um, C
C =
cran
ial-c
auda
l, LR
= le
ft-r
ight
, AP
= an
terio
r-po
ster
ior
53
Supplementary Figure 3.1A | Pediatric patients: Distributions of the individual random errors (σ) of interfractional renal and diaphragmatic position variation in the CC, LR, and AP directions. Spearman’s ρ indicate correlations between organ position variation and patients’ height. Dots (blue) and triangles (dark blue) represent pediatric patients treated without and with anesthesia, respectively. Abbreviations: CC = cranial-caudal, LR = left-right, AP = anterior-posterior
54
Supplementary Figure 3.1B | Adult patients: Distributions of the individual random errors (σ) of interfractional renal and diaphragmatic position variation in the CC, LR, and AP directions. Spearman’s ρ indicate correlations between organ position variation and patients’ height. Abbreviations: CC = cranial-caudal, LR = left-right, AP = anterior-posterior
55
Supplementary Figure 3.2 | Subgroup of children and adults with overlapping heights. Distributions of the individual random errors (σ) of interfractional renal and diaphragmatic position variation in the CC, LR, and AP directions. Spearman’s ρ indicate correlations between organ position variation and patients’ height. Dots (blue) and circles (red) represent pediatric and adult patients, respectively. Abbreviations: CC = cranial-caudal, LR = left-right, AP = anterior-posterior
56
Supplementary Figure 3.3 | Distributions of the individual random errors (σ) of interfractional renal and diaphragmatic position variation in the CC, LR, and AP directions. Spearman’s ρ indicate correlations between organ position variation and patients’ BMI. Dots (blue) and circles (red) represent pediatric and adult patients, respectively. Abbreviations: CC = cranial-caudal, LR = left-right, AP = anterior-posterior, BMI = Body Mass Index (calculated as weight (kg) / height2 (meter))
57
Chapter 4 Abdominal organ position variation in children during image-guided radiotherapy
Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Dirk te Lindert,
Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel
Radiation Oncology 2018; Volume 13 (1): 173.
Doi: 10.1186/s13014-018-1108-9.
59
Abstract
Purpose: Interfractional organ position variation might differ for abdominal organs and this could have consequences for defining safety margins. Therefore, the purpose of this study is to quantify interfractional position variations of abdominal organs in children in order to investigate possible correlations between abdominal organs and determine whether position variation is location-dependent. Methods: For 20 children (2.2–17.8 years), we retrospectively analyzed 113 CBCTs acquired during the treatment course, which were registered to the reference CT to assess interfractional position variation of the liver, spleen, kidneys, and both diaphragm domes. Organ position variation was assessed in three orthogonal directions and relative to the bony anatomy. Diaphragm dome position variation was assessed in the cranial-caudal (CC) direction only. We investigated possible correlations between position variations of the organs (Spearman’s correlation test, ρ), and tested if organ position variations in the CC direction are related to the diaphragm dome position variations (linear regression analysis, R2) (both tests: significance level p<0.05). Differences of variations of systematic (Σ) and random errors (σ) between organs were tested (Bonferroni significance level p<0.004). Results: In all directions, correlations between liver and spleen position variations, and between right and left kidney position variations were weak (ρ≤0.43). In the CC direction, the position variations of the right and left diaphragm domes were significantly, and stronger, correlated with position variations of the liver (R2=0.55) and spleen (R2=0.63), respectively, compared to the right (R2=0.00) and left kidney (R2=0.25). Differences in Σ and σ between all organs were small and insignificant. Conclusions: No (strong) correlations between interfractional position variations of abdominal organs in children were observed. From present results, we concluded that diaphragm dome position variations could be more representative for superiorly located abdominal (liver, spleen) organ position variations than for inferiorly located (kidneys) organ position variations. Differences of systematic and random errors between abdominal organs were small, suggesting that for margin definitions, there was insufficient evidence of a dependence of organ position variation on anatomical location.
60
4.1 | Introduction
Continuous developments in pediatric cancer treatment using multimodality strategies, including surgery, chemotherapy, and radiotherapy have led to increasing numbers of childhood cancer survivors [1]. Inevitably, the occurrence of treatment associated adverse events has also increased. Treatments including radiotherapy significantly contribute to the risk of developing adverse events.
Children are treated with abdominal and thoracic radiotherapy for a wide range of primary cancer diagnoses, including Wilms’ tumor, neuroblastoma, and Ewing sarcoma. Moreover, treatment of the craniospinal axis and lung metastasis involve irradiation of the abdominal and thoracic region. The anatomical locations of these tumors and adjacent organs at risk (OARs) vary; target volumes can be in very close proximity to the lungs, diaphragm, liver, spleen, and kidneys. As a result, healthy tissues and OARs are unavoidably exposed to radiation when irradiating the tumor [2, 3]. Although adequate tumor dose coverage is the primary goal in radiotherapy, sparing the vital and long-term functions of adjacent organs is also of great concern. Especially in children, who have a relative long life expectancy when surviving cancer, organs are still growing and have low tolerance to radiation [4, 5]. To ensure adequate tumor dose coverage while minimizing radiation dose to surrounding healthy tissues, knowledge about the extent of target and organ motion, particularly present in the abdominal and thoracic area, is needed. Thus, quantifying the motion of vital and sensitive organs such as the liver, spleen, and kidneys is essential.
These abdominal organs move with every breathing cycle (intrafraction motion) and from day-to-day (interfraction motion). Intra- and interfractional motion of the tumor and OARs are incorporated by expanding the clinical target volume and OARs volumes to the planning target volume (PTV) and planning risk volumes (PRVs), respectively [6]. In adults, many studies have quantified motion of various organs, enabling to define accurate margins for PTVs and PRVs. Despite the increasing number of publications on pediatric organ motion [7–15], data is still limited and no consensus has been reached in pediatric radiotherapy to define PTV or PRV margins for abdominal tumors or OARs. Therefore, PTV margins for children are currently pragmatically based on available adult data and PRV margins are often not used in pediatric radiotherapy. Due to different anatomical locations (e.g., right vs. left side of the abdomen, (retro)peritoneum, adjacent to the vertebrae), or abdominal processes (e.g., intestinal peristaltic or air pockets), abdominal organ motion might be location-dependent, as was discussed before in Van Dijk et al. [14]. This could lead to differences in PTV and PRV margins depending on the anatomical location.
The most commonly used PTV margin recipe is from van Herk et al. (2.5 ∑ + 0.7 σ), where the systematic (∑) and random (σ) component are based on quadratically adding the systematic/random
errors that occur during treatment (√∑𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝟐𝟐 + ∑𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊
𝟐𝟐 and √𝝈𝝈𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝟐𝟐 + 𝝈𝝈𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊𝒊
𝟐𝟐 ) [16]. Previous studies
mainly reported on intrafractional organ motion, focusing on respiratory-induced abdominal organ motion through various phases of the breathing cycle as measured on a single four-dimensional computed tomography (4DCT) [9, 11, 15] or 4D magnetic resonance imaging (4DMRI) [12, 17]. Although organ motion seems to be more prone to respiratory motion than to day-to-day position variations, Guerreiro et al. showed that in a homogenous group of 15 children, interfractional abdominal organ motion was larger than intrafraction motion (∑𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎 𝜎𝜎𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 > ∑𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 𝑎𝑎𝑎𝑎𝑎𝑎 𝜎𝜎𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖) [15]. In addition, Huijskens et al. showed that for respiratory-induced diaphragm motion in children the systematic error was found to be smaller than the random error (∑𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 < 𝜎𝜎𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖) [8]. This seems to indicate that the systematic component of the PTV and PRV margins is predominated by the day-
61
to-day systematic (i.e., interfractional) variations (∑𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 ). Moreover, van Herk’s margin recipe shows that the systematic component weighs more than the random component [16]. Therefore, quantification and a comprehensive understanding of interfractional abdominal organ motion is essential for high-accuracy image-guided radiotherapy.
Most studies on abdominal organ motion have focused only on the quantification of the interfractional component [7, 10, 15, 18], without investigating location-dependency, or possible correlations between organ position variations. Whenever possible, resection of a tumor takes place before radiation treatment and usually surgical clips are placed to localize the remaining tumor bed. If not, an anatomical structure close to the target could function as a surrogate for localization and position variation. However, such a strategy will only be successful when there is a clear understanding of the correlations between the tumor and the anatomical surrogate. In addition, radiation treatment might also lead in the future towards adaptive strategies in children. However, often, certain organs are not directly visible on daily cone beam CTs (CBCTs), due to artefacts, smaller field of view or, especially in children, low dose imaging protocols. Moreover, markers are rather not placed in children and online evaluation of the positions of organs is thus mostly unfeasible in clinical practice. Here as well, another close anatomical structure might be considered as a surrogate. For instance, when the diaphragm, being very well visible on CBCT images, is used as a surrogate for the assessment of abdominal organ position. Some adult studies have shown reliable correlations of the diaphragm with abdominal organs [19–22], while other studies show weak correlations [20, 23–25]. This is mostly depending on the tumor site and therefore, outcomes cannot be generalized for adults. For children, correlations between the diaphragm and abdominal organs has not been extensively studied. It is therefore crucial to have a clear understanding of the correlation between the tumor or organ and a particular surrogate.
Therefore, the aim of this study was to increase the insight on interfractional position variation of abdominal organs in children. We investigated possible correlations between abdominal organs and determined whether position variation is location-dependent. Additionally, we investigated whether diaphragm position variation could be a surrogate for abdominal organ position variation, by analyzing the right and left diaphragm domes separately.
62
4.2 | Methods
Patient population
For this retrospective study, we included 20 patients younger than 18 years, treated for various tumors at our radiation oncology department between December 2010 and September 2017 (Table 4.1). Patients were included if a pre-treatment CT scan and multiple CBCT scans of the abdomen or thorax were available, in which the liver, spleen, kidneys, and right and left diaphragm domes were visible (Figure 4.1).
Table 4.1 | Patient characteristics
No. Sex Tumor Age at
diagnosis (years)
Height (cm)
Weight (kg)
No. of CBCTs RT location
1 F Sarcoma 11.5 155 38 5 Thorax 2 M Medulloblastoma 6.6 110 18 5 Spinal corda
3 F Hodgkin lymphoma 16.5 166 49 5 abdomen 4 M Medulloblastoma 14.1 175 36 5 Spinal cord 5 M Medulloblastoma 8.3 128 25 5 Spinal cord 6 F Medulloblastoma 6.7 117 20 2 Spinal cord 7 M Ewing sarcoma 16.8 184 62 8 Thorax 8 M Medulloblastoma 6.7 129 24 5 Spinal cord 9 M Spinal metastesis 2.6 90 12 8 Thorax
10 F Medulloblastoma 7 118 22 6 Spinal cord 11 M Anaplastic glioma 7.9 132 31 5 Spinal cord
12b M Medulloblastoma 5.1 109 17 8 Spinal cord 13 F Neuroblastoma 5.3 115 24 6 Abdomen 14 M Sarcoma 10.9 142 37 5 Thorax 15 M DSRCT 9.9 137 26 5 Abdomen 16 M Neuroblastoma 4.7 118 22 6 Abdomen
17b M Medulloblastoma 4.9 105 18 6 Spinal cord 18 M Ewing sarcoma 17.9 182 81 7 Thorax 19 F Osteosarcoma 15.1 159 53 5 Thorax 20 M Neuroblastoma 2.2 90 15 6 Abdomen
Abbreviations: M = male; F = female; DSRCT = desmoplastic small round cell tumor; a Spinal cord was part of craniospinal irradiation b Patients 12 and 17 were treated under general anesthesia.; this did not influence interfractional organ position variations
63
Imaging data
For each patient, a pre-treatment CT scan (120 kV, 2.5- or 5 mm slice thickness) was acquired for planning purposes (LightSpeed RT16; General Electric Company, Waukesha, WI, USA). This scan was considered as the reference CT (refCT) scan and included original organ delineations, as used for clinical practice (Figure 4.1). For all patients, CBCT images (1 mm slice thickness, 1 mm in-plane resolution) were routinely acquired using the CBCT scanner mounted on the Elekta Synergy linac (Elekta AB, Stockholm, Sweden) as part of the position verification protocol. This yields CBCT imaging at the first three treatment fractions, followed by daily or weekly CBCT acquisitions, depending on the treatment protocol. To be consistent, we included for all patients the first three CBCTs and thereafter weekly acquired CBCTs. All CBCTs were acquired with 120 kV, 10 mA, and 10 or 40 ms exposure time per projection. The scanning time of the CBCT scan varied between 35–60s, and the degree of circumferential rotation was 200 or 360 degrees. In this study, we retrospectively analyzed the imaging data, including a total of 20 refCTs and 113 CBCTs.
Imaging registration
Elekta X-ray Volume Imaging software (XVI 3.0; Elekta AB, Stockholm, Sweden) was used for a two-step rigid registration for each organ separately (example shown in Figure 4.1). First, a region of interest (ROI) was defined in the refCT, including the 12th thoracic through the 4th lumbar vertebra (from the lowest part of the kidneys up to the diaphragm domes). The CBCTs were then registered to the refCT using the automatic chamfer match algorithm [26]. Second, this bony anatomy-based match was followed by registration of each organ separately (i.e., liver, spleen, right kidney, left kidney) with a grey value algorithm [26], based on shaped ROIs defined by the delineated organs including (at least 2/3rd of) the whole organ volume. This enabled the assessment of organ position variation smaller than the slice thickness of the acquired refCT. Automatic registration outcomes (translations and rotations) were visually checked (by SCH/DTL) and manually corrected if necessary. Results were corrected for rotations as follows. First, we assessed the center of mass (COM) coordinates for each organ. Then, we equated these coordinates to the refCT to determine the exact magnitude and direction of the interfractional position variation. By calculating the difference of the magnitude and sign of the COM coordinates of each organ on CBCTs and refCT, registrations resulted in interfractional position variation relative to bony anatomy, expressed as composite vectors in the left-right (LR), cranio-caudal (CC) and anterior-posterior (AP) directions. The + and – signs respectively indicate right/caudal/posterior and left/cranial/anterior directions. For the diaphragm, the bony anatomy-based automatic chamfer match was followed by manual registrations of the right- and left-sided diaphragm dome separately in the CC direction only (by SCH/DTL).
64
Figure 4.1 | A) Delineated organs (right kidney: purple, left kidney: blue, liver: yellow, spleen: pink) on the reference CT. Diaphragm domes are not delineated. Arrows indicate mutual correlations investigated. B) Example of the two-step rigid registration (from top to bottom): unaligned overlap of reference CT and CBCT, bones aligned, right kidney aligned (note: bones shifted).
65
Statistical Analysis
For each patient, organ specific mean and standard deviation (SD) of the interfractional position variation relative to the bony anatomy were determined in the three orthogonal directions, and in the CC direction only for the right and left diaphragm domes. Furthermore, over all patients, we estimated per organ the group mean (i.e., mean of the individual means), the group systematic error (Σ; the SD of the individual means of all patients), and the group random error (σ; the root mean square of the individual SDs of all patients).
To evaluate whether organ position variation is location-dependent, we compared contralateral and superiorly and inferiorly located abdominal organs separately (indicated in Figure 4.1). Since not all data fitted a normal distribution (tested with the Shapiro-Wilk’s test), differences between contralateral organs’ systematic and random errors, were separately tested (i.e., right diaphragm dome vs. left diaphragm dome, liver vs. spleen, right kidney vs. left kidney) with the Levene’s test (for Σ) and Mann-Whitney U-test (for σ). Also, differences in Σ and σ between superiorly and inferiorly located abdominal organs were tested (i.e., liver vs. right kidney, spleen vs. left kidney). Since differences were tested in fourteen combinations (i.e., LR, CC, AP for four organs, and CC only for both diaphragm domes), we adjusted p values according to the Bonferroni correction. Differences were considered to be significant if test outcomes showed a p value<0.004 (i.e., 0.05/14).
We used the Spearman’s correlation test (significance level p<0.05) to investigate the possible correlations in position variations between contralateral organs.
Additionally, to test if right- and left-sided organ position variations in the CC direction are related to the position variations of the superiorly located right- and left-sided diaphragm dome respectively, we used linear regression analysis (significance level p<0.05). Both tests were also performed for each individual patient.
All statistical analyses were done using R version 3.2.1. (R Foundation for Statistical Computing, USA).
Table 4.2 | The group systematic (Σ) and group random errors (σ) in mm in the orthogonal directions for the right kidney, left kidney, liver, and spleen and in CC direction for the diaphragm.
Abbreviations: LR = left–right; CC = cranial–caudal; AP = anterior–posterior
66
4.3 | Results Mean organ position variation was smaller than 1.0 mm (range: -6.9 to 7.4 mm) for the abdominal organs in all orthogonal directions and smaller than 1.8 mm (range: -4.0 to 7.8 mm) for the diaphragm domes in the CC direction. For all organs and across all fractions, ranges of position variations were largest in the CC direction (varying from 10.6 to 13.0 mm) and smallest in the LR direction (varying from 4.1 to 11.1 mm) (Figure 4.2). Overall, kidney position variations were smaller than position variations of the liver and spleen (Figure 4.2). Table 4.2 presents the values of the group mean, systematic and random error per organ in each direction, mainly showing average systematic error in decreasing order of CC (3.2 mm, SD=0.3 mm), AP (1.9 mm, SD=0.9 mm), and LR (1.7 mm, SD=0.5 mm) direction, and average random error also in decreasing order of CC (3.0 mm, SD =0.5 mm), AP (2.1 mm, SD=0.6mm), and LR (1.9 mm, SD=0.7mm) direction. Differences of the systematic error between right- and left-sided organs were insignificant (p≥0.004), as were the differences of the random error between right- and left-sided organs (p≥0.004) (Supplementary Table 4.1). For superiorly and inferiorly located organs, significant but small differences were found between the liver and the right kidney in the AP direction (p=0.002 for Σ), and in the LR direction (p=0.000 for σ). Also, the random error of the spleen and the left kidney was significantly different in the AP direction (p=0.001) (Supplementary Table 4.1).
A moderate and statistically significantly correlation between the position variations of the right and left diaphragm domes was found (ρ=0.63, p=0.00) (Figure 4.3a). The position variations of the liver and spleen in the LR and CC direction were weakly, but statistically significantly correlated (ρ=0.23, p=0.02 and ρ=0.40, p=0.00, respectively) (Figure 4.3b). Position variations of the right and left kidney were weakly, but statistically significant correlated in the LR and AP directions (ρ=-0.43, p=0.00 and ρ=0.23, p=0.01, respectively) (Figure 4.3c). Correlations within each individual patient were similar to the overall group outcomes.
Linear regression analysis showed that right and left diaphragm dome position variations in the CC direction were significantly correlated with position variations of the liver (R2=0.55, p=0.00) and spleen (R2=0.63, p=0.00), respectively. In the CC direction, no (strong) correlation was found between right and left diaphragm dome position variations and the position variations of the right (R2=0.003, p=0.60) and left kidney (R2=0.25, p=0.00) , respectively (Figure 4.4).
67
Figure 4.2 | Boxplots showing the distributions of the individual means (upper panel) and SDs (lower panel) of the interfractional position variations found for right- (light grey) and left-sided (dark grey) organs for all 20 patients. Horizontal bars, boxes, and whiskers represent medians, 50th percentiles (inter quartile range (IQR)), and the highest (lowest) value within 1.5xIQR, respectively. Circles denote outliers. *Significant differences (Bonferroni corrected p<0.004). Abbreviations: LR = left–right; CC = cranial–caudal; AP = anterior–posterior.
68
Figure 4.3 | For all CBCT scans, scatterplots describing relations (Spearman’s ρ and p-value) between right and left diaphragm position variations in the CC direction only (a) and right- (x-axis) and left-sided (y-axis) organ interfractional position variations separately (b; liver and spleen, c; right and left kidney), in the three orthogonal directions. Abbreviations: LR = left–right; CC = cranial–caudal; AP = anterior–posterior.
Figure 4.4 | Scatterplots with regression lines of the linear regression analyses describing relationships for each CBCT between right- and left-sided interfractional organ position variation (y-axis) and diaphragmatic position variation in the CC direction (x-axis). Abbreviations: CC = cranial–caudal.
69
4.4 | Discussion In this study, we quantified interfractional position variation of multiple abdominal organs in 20 children during radiotherapy and evaluated if organ position variation is mutually related and location-dependent. We found weak correlations between the position variations of contralateral organs. In the CC direction, right and left diaphragm dome position variations correlated moderately with the position variations of the liver and spleen, respectively. However, correlations between the position variations of the diaphragm domes and those of both kidneys were negligible. Furthermore, the largest magnitude of organ position variations was observed in the CC direction, followed by the AP and LR directions. We found that differences between group systematic and random errors of abdominal organs were small and insignificant. This comprehensive analysis of organ position variations at different anatomical locations increases the insight in possible consequences for margin definitions, which has not been reported on for children so far.
Nazmy et al. studied interfractional position variation of the liver and kidneys in 9 children (mean age: 4.1 years, SD = 1.6 years) using reference CT and CBCT scans [10]. They also found that, in the CC direction, the liver showed more motion than the kidneys. However, their range of observed position variations of the left kidney was smaller than that of the right kidney. In contrary, when we analyzed patients in our cohort of similar age (n=6; range 2.2-5.3 years) we found slightly larger position variations of the left kidney compared to the right kidney. Although, this comparison involves small sample sizes, a possible explanation might be the different methodology in choosing the point of interest. Nazmy et al. used the upper pole of the kidneys whereby kidney deformations might have been interpreted as translations, resulting in an overestimation of motion. We used the COM as point of interest because it is less sensitive to organ deformations. Although data on organ deformation would provide useful additional information on organ motion characteristics, analyzing organ deformation was outside the scope of the current study.
Our results are comparable to findings of Guerreiro et al. who used a similar methodology as we did [15]. They quantified interfractional position variations of the spleen, liver, and the healthy kidney in patients (n=15, mean age: 4 years) with Wilms’ tumors. Their ranges of interfractional position variation, and the systematic and random errors were generally somewhat smaller than our results, which could be explained by the fact that their cohort consisted of younger patients (age range 1-8 years). However, when we analyzed patients in our cohort of similar age (n=10; range 2.2-7.8 years), the systematic and random errors for all organs and directions in our cohort remained somewhat larger (for Σ; mean difference 1.0 mm, SD=0.6 mm, for σ; mean difference 0.7 mm, SD= 0.7 mm).
Using a 3DCT as a reference point to estimate interfractional position variation is arguable. The 3DCT represents a ‘snapshot’ of repeatedly changing organ positions during the respiratory cycle [27]. A CBCT captures in 35-60 seconds several complete respiratory cycles and averages the motion over the observed breathing phases into one blurred 3D image. To investigate the possible effect of respiratory motion differences on the 3DCT and the CBCTs, we recalculated our measurements using the first CBCT scan as the reference scan instead of the 3DCT. Differences between the respective calculations based on the refCT and the first CBCT were negligible (<1mm). Also, although projection images could enable the quantification of intrafractional motion as well [28], the low dose CBCT protocols that we used for most children [29] unavoidably result in poorer quality of projection images. Therefore, we were not able to distinguish organs on the two-dimensional projection images of these CBCT scans in order to investigate intrafractional motion of the liver, spleen, and kidneys.
70
The outliers shown in Figure 4.2 represent the SD values of the right kidney and spleen position variations of three patients. For one patient, in which the field of view of the CBCT scan was smaller than its refCT, the whole right kidney was visible on the refCT but remained only half visible on the CBCT scan, and registration was performed using an adjusted sub-volume of the kidney. Additionally, in this patient the distance of the COM of the right kidney to the treatment planning isocentre on the refCT was relatively large (>10mm), resulting in a large deviation in organ position variation. For two other patients, the two-step rigid registration for the spleen yielded large rotations (>15 degrees), resulting in large ranges of position variations. However, a sensitivity analysis, excluding these three cases, did not change our results.
The liver and spleen are contralateral organs that substantially differ in tissue composition and function. However, regarding their position variations, differences were small and position variations of both organs were moderately correlated with the position variations of the diaphragm domes. In contrary, the position variations of both kidneys were smaller and showed weak correlations with the diaphragm dome position variations. This might be due to their more inferior and retroperitoneal location. Further, visual inspection showed that the kidneys seem more prone to deformations than the liver and spleen, probably due to their different tissue composition. Therefore, although in the CC direction only, diaphragm position variations, seem to particularly be more representative for position variations of OARs in the upper abdomen than for OAR position variations in the lower abdomen.
The weak to moderate (ρ<0.4), however significant, correlations of position variations between right- and left-sided abdominal organs suggest that organs move only somewhat in similar directions. Therefore, for future online strategies, close located anatomical structures are not recommended as suitable surrogates. However, the overall magnitude of motion is small, and differences of systematic and random errors of the various abdominal organs are small and insignificant, hence negligible. Therefore, regarding margin definitions, there was insufficient evidence of a dependence of organ position variation on anatomical location. Additionally, although differences between abdominal organ position variations were small, overall position variation was largest in the CC direction and smallest in the LR direction. This suggests that margins should be applied anisotropically rather than isotropically. Note, however, that the diaphragm was measured in the CC direction only.
Knowledge about patient’s day-to-day anatomical variation is furthermore valuable when (automating) selecting similar patients from a database of patients’ CT scans for, e.g., automating treatment planning or dose reconstruction [30–34], because this provides a lower bound on the achievable precision of selection.
Besides, as recommended by the Paediatric Radiation Oncology Society (PROS), consensus needs to be reached regarding appropriate margin definitions in children [35]. With increasing data, knowledge on organ motion during radiotherapy in children is expanding. However, due to generally small patient numbers and different methodologies in separates studies, definitive statements regarding margin definitions cannot be made yet. Therefore, close collaborations between research groups, and pooling of data might contribute to achieving consensus on margin definitions. A summarized all-encompassing overview of all published data so far, including inter- and intrafractional organ motion, could provide a basis for this. Especially, with more proton and carbon therapy facilities in development, aiming for high-precision radiotherapy and the need for the assessment of the anatomical variations in children, induced by organ motion, becomes even more important.
71
4.5 | Conclusions
No (strong) correlations between interfractional position variations of abdominal organs in
children were observed. Differences of systematic and random errors between abdominal organs were small, suggesting that for margin definitions, there was insufficient evidence of a dependence of organ position variation on anatomical location. From present results, we concluded that diaphragm dome position variations could be more representative for superiorly located abdominal (liver, spleen) organ position variations than for inferiorly located (kidneys) organ position variations.
72
References
1. Curry HL, Parkes SE, Powell JE, Mann JR. Caring for survivors of childhood cancers: The size of the problem. Eur. J. Cancer 2006; 42(4):501–508.
2. Bölling T, Könemann S, Ernst I, Willich N. Late effects of thoracic irradiation in children. Strahlentherapie und Onkol. 2008; 184(6):289–295.
3. Bölling T, Willich N, Ernst I. Late effects of abdominal irradiation in children: A review of the literature. Anticancer Res. 2010; 30(1):227–232.
4. Selo N, Bölling T, Ernst I et al. Acute toxicity profile of radiotherapy in 690 children and adolescents: RiSK data. Radiother. Oncol. 2010; 97(1):119–126.
5. Claude L, Laprie A. Quelles contraintes pour quels organes à risque en radiothérapie chez l’enfant? Cancer/Radiotherapie 2015; 19(6–7):484–488.
6. Prescribing, Recording and Reporting Photon Beam Therapy (Supplement to ICRU Report 50), ICRU Report 62 . ICRU
7. Huijskens SC, van Dijk IWEM, de Jong R et al. Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: A multicenter study. Radiother. Oncol. 2015; 117(3):425–431.
8. Huijskens SC, van Dijk IWEM, Visser J et al. Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy. Radiother. Oncol. 2017; 123(2):263–269.
9. Kannan S, Teo BKK, Solberg T, Hill-Kayser C. Organ motion in pediatric high-risk neuroblastoma patients using four-dimensional computed tomography. J. Appl. Clin. Med. Phys. 2017; 18(1):107–114.
10. Nazmy MS, Khafaga Y, Mousa A, Khalil E. Cone beam CT for organs motion evaluation in pediatric abdominal neuroblastoma. Radiother. Oncol. 2012; 102(3):388–392.
11. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
12. Uh J, Krasin MJ, Li Y et al. Quantification of Pediatric Abdominal Organ Motion With a 4-Dimensional Magnetic Resonance Imaging Method. Int. J. Radiat. Oncol. Biol. Phys. 2017; 99(1):227–237.
13. Demoor-Goldschmidt C, Chiavassa S, Josset S et al. Asservissement respiratoire lors d’une radiothérapie pulmonaire bilatérale pour le sarcome d’Ewing ou le néphroblastome chez des enfants et jeunes adultes : études dosimétrique et clinique de faisabilité. Cancer/Radiotherapie 2017; 21(2):124–129.
14. van Dijk IWEM, Huijskens SC, de Jong R et al. Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference? Acta Oncol. 2017; 56(8):1065–1071.
15. Guerreiro F, Seravalli E, Janssens GO et al. Intra- and inter-fraction uncertainties during IGRT for Wilms’ tumor. Acta Oncol. (Madr). 2018; 57(7):941–949.
16. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 17. Pai Panandiker AS, Winchell A, Rolen M et al. 4DMRI Provides More Accurate Renal Motion
Estimation for IMRT in Young Children. Int. J. Radiat. Oncol. 2013; 87(2):S599. 18. Beltran C, Pai Panandiker AS, Krasin MJ, Merchant TE. Daily image-guided localization for
neuroblastoma. J. Appl. Clin. Med. Phys. 2010; 11(4):162–169. 19. Yang J, Cai J, Wang H et al. Is Diaphragm Motion a Good Surrogate for Liver Tumor Motion? Int.
J. Radiat. Oncol. 2014; 90(4):952–958. 20. Xi M, Liu M-Z, Li Q-Q et al. Analysis of abdominal organ motion using four-dimensional CT. Chin.
J. Cancer 2009; 28(9):989–993. 21. Cerviño LI, Jiang Y, Sandhu A, Jiang SB. Tumor motion prediction with the diaphragm as a
surrogate: a feasibility study. Phys. Med. Biol. 2010; 55(9):N221–N229. 22. Dawson LA, Eccles C, Bissonnette J-P, Brock KK. Accuracy of daily image guidance for
73
hypofractionated liver radiotherapy with active breathing control. Int. J. Radiat. Oncol. 2005; 62(4):1247–1252.
23. Siva S, Pham D, Gill S et al. An analysis of respiratory induced kidney motion on four-dimensional computed tomography and its implications for stereotactic kidney radiotherapy. Radiat. Oncol. 2013; 8(1):248.
24. Lens E, van der Horst A, Versteijne E et al. Considerable pancreatic tumor motion during breath-holding. Acta Oncol. (Madr). 2016; 55(11):1360–1368.
25. Stevens CW, Munden RF, Forster KM et al. Respiratory-driven lung tumor motion is independent of tumor size, tumor location, and pulmonary function. Int. J. Radiat. Oncol. 2001; 51(1):62–68.
26. Roche A, Malandain G, Pennec X, Ayache N. The correlation ratio as a new similarity measure for multimodal image registration. Proc. Int. Conf. Med. Image Comput. Comput. Interv., 1998:1115–1124.
27. Balter JM, Ten Haken RK, Lawrence TS et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int. J. Radiat. Oncol. 1996; 36(1):167–174.
28. Zijp L, Sonke J-J, van Herk M. Extraction of the respiratory signal from sequential thorax cone-beam X-ray images. onf Use Comput Radiat Ther (ICCR)., 2004:507–509.
29. De Jong R, Lens E, Van Herk M et al. Optimizing cone-beam CT presets for children to reduce imaging dose illustrated with craniospinal axis. Radiother. Oncol. 2014; 111:S109–S110.
30. Deshpande RR, DeMarco J, Sayre JW, Liu BJ. Knowledge-driven decision support for assessing dose distributions in radiation therapy of head and neck cancer. Int. J. Comput. Assist. Radiol. Surg. 2016; 11(11):2071–2083.
31. Ng A, Nguyen T-N, Moseley JL et al. Reconstruction of 3D lung models from 2D planning data sets for Hodgkin’s lymphoma patients using combined deformable image registration and navigator channels. Med. Phys. 2010; 37(3):1017–1028.
32. Wang Z, van Dijk IWEM, Wiersma J et al. Are age and gender suitable matching criteria in organ dose reconstruction using surrogate childhood cancer patients’ CT scans? Med. Phys. 2018; 45(6):2628–2638.
33. Schmidt M, Lo JY, Grzetic S et al. Semiautomated head-and-neck IMRT planning using dose warping and scaling to robustly adapt plans in a knowledge database containing potentially suboptimal plans. Med. Phys. 2015; 42(8):4428–4434.
34. Virgolin M, van Dijk IWEM, Wiersma J et al. On the feasibility of automatically selecting similar patients in highly individualized radiotherapy dose reconstruction for historic data of pediatric cancer survivors. Med. Phys. 2018; 45(4):1504–1517.
35. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
Appe
ndix
4
Supp
lem
enta
ry T
able
4.1
| P
-val
ues f
or d
iffer
ence
s of g
roup
syst
emat
ic e
rror
s (Σ)
and
gro
up ra
ndom
err
ors (σ)
; tes
ted
for d
iffer
ence
s be
twee
n rig
ht- a
nd le
ft-sid
ed o
rgan
s and
supe
riorly
vs.
infe
riorly
loca
ted
orga
ns in
the
orth
ogon
al d
irect
ions
.
Ri
ght K
idne
y vs
. Le
ft K
idne
y
Li
ver v
s. Sp
leen
Righ
t Kid
ney
vs.
Live
r
Le
ft K
idne
y vs
.
Sple
en
Ri
ght
Diap
hrag
m v
s.
Left
Dia
phra
gm
LR
CC
AP
LR
CC
AP
LR
CC
AP
LR
CC
AP
CC
Σ a
0.66
0.
46
0.17
0.58
0.
97
0.67
0.14
0.
34
0.00
2*
0.
42
0.82
0.
08
0.
66
σ b
0.05
0.
05
0.21
0.89
0.
38
0.91
4
0.00
* 0.
27
0.07
0.02
0.
77
0.00
1*
0.
62
a Diff
eren
ces t
este
d w
ith th
e Le
vene
’s te
st
b Diff
eren
ces t
este
d w
ith th
e M
ann-
Whi
tney
U-t
est
* ad
just
ed p
val
ues a
ccor
ding
to th
e Bo
nfer
roni
cor
rect
ion
(Sig
nific
ance
leve
l : p
val
ue<0
.004
(i.e
., 0.
05/1
4))
PART 2 Intrafractional motion
Chapter 5 Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy
Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel
Radiotherapy and Oncology 2017; Volume 123 (2): 263-269
Doi: 10.1016/j.radonc.2017.03.016
79
Abstract
Purpose: To analyse the variability of respiratory motion during image-guided radiotherapy in paediatric cancer patients and to investigate possible relationships thereof with patient-specific factors. Methods: Respiratory-induced diaphragm motion was retrospectively analysed on 480 cone beam CTs acquired during the treatment course of 45 children (<18 years). The cranial-caudal positions of the right diaphragm dome top in exhale and inhale phases were manually selected in the projection images. The difference in position between both phases defines the amplitude. The cycle time equalled inspiratory plus expiratory time. We analysed the variability of the intra- and interfractional respiratory motion and studied possible correlations between respiratory-induced diaphragm motion and age, height, and weight. Results: Over all patients, mean amplitude and cycle time were 10.7 mm (range 4.1-17.4 mm) and 2.9 s (range 2.1-3.9 s). Intrafractional variability was larger than interfractional variability (2.4 mm vs. 1.4 mm and 0.5 s vs. 0.4 s for amplitude and cycle time, respectively). Correlations between mean amplitude and patient-specific factors were significant but weak (p<0.05, ρ ≤ 0.45). Conclusions: Large ranges of amplitude and cycle time and weak correlations confirm that respiratory motion is patient-specific and requires an individualised approach to account for. Since interfractional variability was small, we suggest that a pre-treatment 4DCT in children could be sufficiently predictive to quantify the respiratory motion.
80
5.1 | Introduction
High-precision image-guided radiotherapy (IGRT) is extremely important in children [1–9], since anatomical deviations during the RT course can yield a miss of the tumour. A large dose deposition in the surrounding healthy tissues induces the risk of (late) side effects [10–12]. To ensure adequate tumour coverage while minimizing dose to healthy surrounding tissues, it is crucial to comprehensively quantify the geometric uncertainties.
In clinical practice, to account for these geometric uncertainties, the clinical target volume (CTV) and organs at risk (OARs) are expanded with a safety margin, defining the planning target volume (PTV) and planning OARs volumes (PRVs), respectively [13]. Solely accounting for the internal organ motion leads to the internal target volume (ITV), which includes the CTV plus an internal margin (IM), covering the entire tumour motion range. However, this leads to large margins and increased dose to healthy surrounding tissues. An alternative approach, the mid-ventilation based PTV planning, leads to smaller margins, simultaneously accounting for respiratory motion and other geometrical uncertainties [14–16].
The optimal safety margin to account for respiratory motion in children is yet unknown, since quantitative studies in children are scarce. A few studies have reported on accuracy of RT for cranial and head and neck tumours in children [1–5], indicating that appropriate patient immobilization and a safety margin <3.5 mm are sufficient to account for target motion. However, literature on extra-cranial RT is limited [6–8]. Recently, we quantified day-to-day position variation of the kidneys and diaphragm in children [7] and compared this to adults [17]. This interfractional renal and diaphragm position variation in children was notably smaller than in adults [7, 17]. However, mainly due to respiration during the treatment, abdominal and thoracic tumours are more prone to intrafractional organ motion, which limits the accuracy of RT.
Motion compensating techniques, such as breath-hold, beam gating, tracking, and abdominal compression are well studied and applied in adults [18]. Some of these techniques are not patient-friendly, require intensive training, and imply an increased workload in the clinic. Although, paediatric RT could potentially benefit from these techniques as well, children experience RT already as a stressful procedure [19–21] and these techniques may cause further distress and anxiety. Additionally, it is questionable if the youngest children (e.g. ≤ 8 years) would be able to follow a breath-hold procedure. Therefore, these techniques are currently not frequently used in paediatric RT.
Also, due to the ALARA principle (keeping doses as low as reasonable achievable), and previously reported radiation risks in children from computed tomography (CT) [22–25], four-dimensional (4D) CT is not commonly used in children. Additionally, respiratory patterns vary from day-to-day and a single 4DCT might not be a good representation for daily respiratory motion during the whole treatment course [26]. Therefore, to account for respiratory motion in children the entire range of motion should be considered when defining the safety margins. Knowledge on variability in respiratory motion is essential for adequate margin definitions, which is even more important in proton therapy than in photon therapy [27–29].
The only study on respiratory motion in children quantified renal and diaphragmatic intrafractional motion using 4DCTs and showed a strong correlation between age and diaphragm motion [6]. However, since the measurements were performed in a single 4DCT, no data were available on the variability of respiratory motion during and between multiple fractions throughout the treatment course. Also, no data were available about possible relationships between respiratory motion variabilities and patient-specific factors, such as age, height, and weight. Knowledge about such
81
relationships might be beneficial for developing a more individualised approach and precise margin definitions.
Daily or weekly cone-beam (CB)CT scans acquired during IGRT for position verification enable quantification of inter- and intrafractional organ motion. However, methodologies differ in large extent. For the quantification of respiratory motion the 2D fluoroscopic projection images of the CBCT scans are used. Several methods using 2D projection images for quantification of respiratory motion have been investigated in multiple adult-based studies [30–33]. The Amsterdam Shroud (AS) method is a frequently used and reliable method for detection of respiratory-induced diaphragm motion [32, 34, 35]. Our study is the first to extract the respiratory signal of children from 2D projection images using an adapted version of the AS method, thus enabling the assessment of the variability of respiratory motion within and between fractions throughout a paediatric treatment course.
The purpose of this study was to quantify and analyse respiratory-induced diaphragm motion, since the diaphragm is well visible in CBCT images and the respiratory-induced motion of the diaphragm is assumed to correlate with upper abdominal and thoracic target volumes and OARs. Additionally, our relatively large patient number (n=45) enabled investigation of possible correlations between respiratory-induced diaphragm motion and aforementioned patient-specific factors.
5.2 | Methods
Patient data
We retrospectively analysed the CBCT data of 45 children who received IGRT between December 2010 and May 2016. Patients were included when the diaphragm was visible on upper abdominal or thoracic CBCTs. Mediastinal surgery, causing diaphragmatic dysfunction, was an exclusion criterion because the remaining diaphragmatic motion was minimal and therefore not representative for normal free breathing. Median age at the start of RT was 11 years (range 2–18 years). Median height was 148 cm (range 90–186 cm). Most patients (43/45) were treated in supine position and 7 children (age range 2–11 years) were treated under general anaesthesia (GA). Immobilization devices are not used in our institute for abdominal and thoracic RT. A full overview of patient characteristics and treatment details, including tumour indication and RT location is listed in Table 5.1. CBCT acquisition
For each patient, CBCTs (Synergy, Elekta Oncology systems, Crawly, UK) for position verification were acquired for the first three treatment fractions, followed by daily or weekly CBCTs according to an accustomed extended no-action level (eNAL) protocol [36], and corresponding to tumour type and patient-specific needs. The acquisition parameters for all CBCTs were 120 kV, 10 mA and 10 or 40 ms exposure time per projection. The circumferential rotation varied from 200 to 360 degrees and the timeframe varied between 35 s and 120 s, resulting in a variation in number of projection images per CBCT (180 to 760). A single projection was acquired in 180 ms. In this study, we included a total of 480 CBCTs (median 7, range 4-32 CBCTs per patient).
82
Table 5.1 | Patient characteristics and radiotherapy (RT) location.
No. Gender Tumour Age at
start RT (years)
Height (cm)
Weight (kg) RT location Number
of CBCT
1 M Ewingsarcoma 10.9 137 28 Thorax 6 2 F Ewingsarcoma 11.6 155 38 Thorax 27 3 M Medulloblastoma 11.0 158 46 CSI 4 4 M Medulloblastoma 10.9 150 33 CSI 5 5 F Ewingsarcoma 17.9 164 53 Thorax 17 6a M Medulloblastoma 6.6 110 18 CSI 5 7 F Ewingsarcoma 12.5 151 70 Thorax 32 8 M Medulloblastoma 14.1 175 36 CSI 9 9 F Medulloblastoma 8.2 139 28 CSI 7
10 F Neuroblastoma 6.2 103 18 Thorax 23
11 M CCSK 8.6 126 23 Abdomen 5 12 M Ewingsarcoma 16.8 184 62 Thorax 8
13a,b M Medulloblastoma 7.6 126 24 CSI 6 14 F Germinoma 9.3 139 43 CSI 6 15 M Medulloblastoma 6.7 129 24 CSI 5 16 F Hodgkin lymphoma 15.9 163 51 Thorax 6 17 F Medulloblastoma 7.0 118 22 CSI 7 18 M Medulloblastoma 11.0 143 28 CSI 4 19b M Pineal germinoma 15.8 159 48 CSI 17 20 M Pineal germinoma 13.3 166 48 CSI 6 21 M Medulloblastoma 6.2 123 23 CSI 10 22 M Ewingsarcoma 14.7 177 65 Thorax 10 23 F Ewingsarcoma 15.8 173 50 Thorax 8 24 25a
F F
B-cel lymphoma Ependymoma
16.8 5.6
163 113
62 20
Thorax Spinal
12 21
26 F Ewingsarcoma 13.1 148 43 Thorax 10 27 M ERMS 16.8 186 64 Thorax 5 28 M Glioma 7.8 132 31 CSI 14 29 F Medulloblastoma 14.7 159 50 CSI 9 30a M Medulloblastoma 5.1 109 17 CSI 8 31 F Neuroblastoma 5.3 115 24 Abdomen 7 32 F Medulloblastoma 12.5 156 37 CSI 6 33 M RMS 10.9 142 37 Thorax 25 34 M Wilms’ tumour 15.6 167 67 Abdomen 7 35 M Germinoma 16.9 172 68 CSI 6 36 M DSRCT 9.9 137 26 Abdomen 5 37 M ERMS 3.3 106 17 Abdomen 24 38 M Neuroblastoma 4.7 118 22 Thorax 6 39 F Ewingsarcoma 17.4 154 78 Thorax 16 40a M Medulloblastoma 4.9 105 18 CSI 10 41 M Ewingsarcoma 17.8 182 81 Thorax 28 42 F Osteosarcoma 15.1 159 53 Thorax 6 43a M ERMS 10.7 132 28 Abdomen 4 44a M Neuroblastoma 2.2 90 15 Abdomen 6 45 M RMS 14.4 172 59 Thorax 12
Abbreviations: RT = radiotherapy; CBCT = cone beam computed tomography; M = male; F = female; CCSK = Clear Cell Sarcoma Kidney; (E)RMS = Embryonal Rhabdomyosarcoma; DSRCT = Desmoplastic small round cell tumour; CSI = Craniospinal Irradiation a Patients were treated under general anaesthesia b Patients were treated in prone position
83
Diaphragm Tracking
In four consecutive steps, we extracted the respiratory-induced diaphragm motion from CBCTs using an adapted version of the AS method [32] (example shown in Supplementary Figure 5.1). First, a region of interest (ROI) was defined on a single projection image, including the top of the right diaphragm. For each projection image, within the selected ROI, the derivative of the grey values along the cranial-caudal (CC) direction was calculated and the resulting pixel values along each line (perpendicular to the CC direction) were summed, creating a one-dimensional image. Accumulating all one-dimensional images created a two-dimensional AS image. Along the horizontal axis of this image, representing the projection images, we manually selected the projection images corresponding to end-exhale and end-inhale positions of the diaphragm. In each of those selected projection images, we then manually determined the CC position of the diaphragm dome top. Subsequently, the pixel coordinate corresponding to the position of the diaphragm dome top was translated to millimetres relative to the patients’ planned isocenter and was also corrected for the geometry of the CBCT scanner [31]. This resulted in a patient- and CBCT-dependent timeframe describing the CC position of the diaphragm in end-exhale and end-inhale phases (peaks) over the course of CBCT acquisition.
Respiratory analysis
Details concerning the extraction of the respiratory-induced diaphragm motion characteristics and analysis thereof are presented in a schematic overview (Supplementary Figure 5.1). The amplitude was defined as the difference between the diaphragm position in the end-inhale and end-exhale phase. The cycle time equals inspiratory time plus expiratory time. Results were analysed per fraction and per patient, in order to evaluate the outcomes over the whole patient group.
For each fraction, we calculated the mean amplitude and intrafractional variability (the standard deviation (SD) over all amplitudes). For each patient, we calculated the mean amplitude, interfractional variability (i.e., the SD over mean amplitudes from each fraction), and the intrafractional variability (root mean square of the SD from each fraction). For the whole patient group, we calculated the group mean amplitude by averaging the patients’ mean amplitude, the group interfractional variability by averaging the patients’ interfractional variabilities, and the group intrafractional variability by averaging the patients’ intrafractional variabilities. Also, interpatient variability was calculated, expressed as the SD over all patients’ mean amplitude. All of this was also computed for the cycle time. Statistical Analysis To test for normality of our data we used the Shapiro-Wilks test. Since not all data fitted the normal distribution, we used the Spearman’s correlation test (significance level p < 0.05) to test for possible relationships between respiratory-induced diaphragm motion characteristics and patient-specific factors (age, height, and weight). We tested if respiratory-induced diaphragm motion characteristics of the patients treated under GA (n=7, age range 2-11 years) differed from patients treated without GA in a similar age range (n=12, age range 3-10 years) (Mann-Whitney-U test, significance level p<0.05).
84
Figure 5.1 | Scatter plots describing relations (Spearman’s q and p-value) between respiratory-induced diaphragm motion characteristics and age. Dots and triangles represent patients treated without general anaesthesia (GA), and patients treated under GA, respectively (significance level: p < 0.05). Note: y-axes differ in range.
85
5.3 | Results Over all patients, average amplitude and cycle time were 10.7 mm (range 4.1–17.4 mm) and 2.9 s (range 2.1–3.9 s). Intrafractional variabilities were larger (2.4 mm amplitude, 0.5 s cycle time) than interfractional variabilities (1.4 mm amplitude, 0.4 s cycle time). Interpatient variabilities were 2.9 mm and 0.5 s for amplitude and cycle time, respectively.
An overview of the findings regarding the investigation of the possible correlation between the mean, inter- and intrafractional amplitude and cycle time variabilities, on the one hand, and age, height, and weight, on the other hand, is given in Figure 5.1 and Supplementary Figure 5.2. A significant, but weak correlation was found between mean amplitude, on the one hand, and age, height, and weight, on the other hand, (ρ = 0.40, 0.45, 0.33, respectively). Correlations between inter- and intrafractional variability of amplitude and cycle time, on the one hand, and age, height, and weight, on the other hand, were very weak and insignificant.
Intrafractional amplitude variability was significantly smaller in children treated under GA (1.6 mm) than in children of similar ages treated without GA (2.4 mm), whereas other respiratory-induced diaphragm motion characteristics did not differ (Figure 5.2).
Figure 5.2 | Boxplots describing differences between patients treated under general anaesthesia (GA, n = 7) and patients without GA within the same age range (no GA, n = 12). Boxes: median value and upper and lower quartiles; whiskers: lowest and highest data point within 1.5 x interquartile range; circles: outliers. Note: y-axes differ in range. *Significant differences (p<0.05).
86
5.4 | Discussion We quantified and analysed respiratory-induced diaphragm motion and its variability within and between fractions during complete RT courses in 45 children treated with IGRT. Over all patients, interfractional variability was smaller than intrafractional variability. We found large ranges of respiratory-induced diaphragm amplitude motion and cycle time, indicating substantial differences between patients. Moreover, no clinically significant correlations were found between respiratory-induced diaphragm motion characteristics and patient-specific factors (age, height, and weight). Our study is the first to include such a large number of children and the availability of the corresponding daily or weekly acquired CBCTs comprised an extensive dataset. This allowed for an all-encompassing analysis of respiratory-induced diaphragm motion in children throughout their treatment course, which has not been reported for children so far.
In our institute, paediatric CBCTs are acquired using lower imaging doses than used for adults [23]. We considered using an automatic method to detect the diaphragm dome, as used by Rit et al. [32], however, lower imaging doses result in less projection images and reduced image quality, which hampers the use of an automatic method to detect the diaphragm dome. Therefore, in our study the diaphragm dome was manually tracked in the end-inhale and end-exhale phases in the main direction of respiratory motion (CC direction). Most patients (37/45) received half-fan width CBCTs, resulting in few projection images at the angle where the diaphragm domes are not superimposed. Manually tracking enabled us to include the projection images in which left and right diaphragm domes were superimposed, thus including as many data as possible. Although the top of the diaphragm was clearly visible and easy to distinguish, it was a time-consuming method. Therefore, diaphragm tracking was done by one observer, and inter-observer analysis was not feasible. With a range of 4–32, the number of CBCTs differed substantially between patients. Analysis of our group measurements when corrected for the number of CBCTs showed no significant changes (amplitude <0.5 mm and cycle time <0.1 s). Furthermore, respiratory-induced diaphragm motion does not necessarily correlate with tumour motion [37, 38]. Therefore, using the diaphragm as a surrogate for abdominal and thoracic organ motion has some inaccuracies and induces uncertainties that need to be taken into account for treatment planning by defining safety margins for tumours or OARs. Additionally, since respiratory-induced motion is only one component of geometrical uncertainties, other uncertainties due to interfractional motion, delineation errors, and setup variations, also need to be accounted for.
Panandiker et al. studied intrafractional organ motion using 4DCT, in which the respiratory motion was binned in a series of 8 phases [6]. Maximal diaphragmatic motion between inspiration and expiration phases was quantified. Probably due to their analysis method, they found that age and diaphragmatic motion were highly correlated, in contrast to our results. They divided their patients in a younger (2-8 years, n=11) and an older age group (9-18 years, n=9), and found a mean respiratory-induced diaphragm CC motion of 5.1 mm and 9.6 mm in the younger and older age group, respectively. When we analysed our results stratified in similar age categories, we found these values to be 9.4 mm for our younger group (n=16) and 11.6 mm for our older age group (n=29). However, all younger patients in Panandiker’s study had been treated under GA, while in our study only 7 patients were treated under GA (mean respiratory-induced diaphragm CC motion 8.9 mm). Moreover, extracting the respiratory motion from a 4DCT divided in 8 bins is a coarser method and more sensitive to artefacts, compared to our method. Our measurements were done in daily or weekly CBCTs, which comprise more data and enabled an analysis of respiratory motion during different days, while a 4DCT is a single moment in time.
87
Similar as was found in adults [32], we found larger intrafractional motion variability compared to interfractional motion variability. This indicates that the respiratory motion during each fraction is more irregular than the average respiratory motion over all fractions.
Currently, due to lack of quantitative paediatric-based data, margins in abdominal and thoracic paediatric RT are generally based on data from adults. Our previous study showed that interfractional organ motion was smaller in children than in adults [7, 17]. Additionally, our present results show that respiratory-induced diaphragm amplitude motion in children is also smaller than reported in adults (10.7 mm in children vs. 16.4 mm in adults) [32]. This suggests that current clinically used margins might be too large and that development of guidelines for accurate margin definition in children is urgently needed, as was also recently recommended by the international paediatric radiation oncology society (PROS) [9]. One of the difficulties in developing guidelines, however, is that the amplitude of respiratory-induced diaphragm motion is not known prior to treatment. Although we did not find consistent correlations between respiratory motion characteristics and patient-specific factors, an even larger cohort might be needed to investigate if correlations in different age, height, or weight groups in children could be found in order to estimate prior to treatment what margin size would be appropriate based on those patient-specific factors. However, since we found large interpatient variability, indicating that respiration is patient-specific, accounting for respiratory motion probably requires a more individualized approach, such as 4DCT-controlled RT. Moreover, we found small interfractional motion variability, meaning that the respiratory-induced diaphragm motion per patient was stable over the course of treatment. Therefore, at least a single measurement of respiratory-induced organ motions, as by use of a 4DCT, at the start of treatment might be sufficiently predictive for respiratory-induced organ motion over the course of treatment and can therefore be used during treatment planning to compensate for this uncertainty. Additionally, the need for more accurate imaging with 4DCT, leading to more precise treatment and thus less risk of developing adverse events, should be weighed against the slightly increased imaging dose to the patient. For highly mobile targets in the thorax and upper abdominal region, application of 4DCT is becoming standard in our clinic. Breath-hold might not be effective in eliminating organ motion completely [39]. Moreover, although children are well capable of breath-holding [40] implementing this technique for young children might be a challenge. Alternatively, regularization of breathing might be a more promising option, with some kind of feedback of the breathing signal [41].
The PROS calls for more research on the implementation of modern technologies in paediatric RT [9]. Due to lack of financial support this has been difficult and therefore efforts in adults RT have been pragmatically transferred into a paediatric setting [9]. Even more than in photon therapy, proton therapy demands for high-precision localization of the target volume due to the sharp dose fall-off [29]. More data and knowledge on the internal physiological motion in children is therefore highly needed for optimal abdominal and thoracic proton therapy. Other modern radiation technologies, such as MR-guided RT and respiration-gated RT, are promising in predicting and monitoring the respiratory-induced organ motion over time and during treatment [42–44].
Another solution for better estimation of respiratory motion for margin definitions could be to guide patients’ respiration with an active video system [45]. We found that patients treated under GA had a significantly lower intrafractional amplitude variability than patients within the same age range treated without GA. To reduce intrafractional variability in patients treated without GA, the use of an active video system might be considered to calm the patients’ respiration and reach a more stable respiratory amplitude and cycle time. Previous studies also indicated that familiarizing patients with the RT equipment, staff and process reduces the anxiety and distress [20], which could be an important
88
reason for large intrafractional variabilities. Also, shifting the child’s attention via distraction methods such as e.g., a rewarding system, video games, or storytelling, has shown to increase comfort and reduce anxiety [46].
5.5 | Conclusions
In conclusion, our study provides novel information on respiratory-induced diaphragm motion in children during IGRT. Respiratory motion is patient-specific and requires an individualized approach to account for. Therefore, we suggest performing 4DCT in children, which enables to quantify the individual respiratory motion prior to treatment and to define individualized safety margins, offering an improved potential for tumour control with less dose to healthy tissue.
89
References
1. Beltran C, Krasin MJ, Merchant TE. Inter- and Intrafractional Positional Uncertainties in Pediatric Radiotherapy Patients With Brain and Head and Neck Tumors. Int. J. Radiat. Oncol. 2011; 79(4):1266–1274.
2. Beltran C, Naik M, Merchant TE. Dosimetric effect of setup motion and target volume margin reduction in pediatric ependymoma. Radiother. Oncol. 2010; 96(2):216–222.
3. Beltran C, Naik M, Merchant TE. Dosimetric effect of target expansion and setup uncertainty during radiation therapy in pediatric craniopharyngioma. Radiother. Oncol. 2010; 97(3):399–403.
5. Zhu Y, Stovall J, Butler L et al. Comparison of two immobilization techniques using portal film and digitally reconstructed radiographs for pediatric patients with brain tumors. Int. J. Radiat. Oncol. 2000; 48(4):1233–1240.
6. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
7. Huijskens SC, van Dijk IWEM, de Jong R et al. Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: A multicenter study. Radiother. Oncol. 2015; 117(3):425–431.
8. Nazmy MS, Khafaga Y, Mousa A, Khalil E. Cone beam CT for organs motion evaluation in pediatric abdominal neuroblastoma. Radiother. Oncol. 2012; 102(3):388–392.
9. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
10. Geenen MM, Cardous-Ubbink MC, Kremer LCM et al. Medical Assessment of Adverse Health Outcomes in Long-term Survivors of Childhood Cancer. JAMA 2007; 297(24):2705.
11. van Dijk IWEM, Oldenburger F, Cardous-Ubbink MC et al. Evaluation of Late Adverse Events in Long-Term Wilms’ Tumor Survivors. Int. J. Radiat. Oncol. 2010; 78(2):370–378.
12. Oeffinger KC, Mertens AC, Sklar C a et al. Chronic Health Conditions in Adult Survivors of Childhood Cancer. N. Engl. J. Med. 2006; 355(15):1572–1582.
14. Lens E, Van Der Horst A, Versteijne E et al. Dosimetric advantages of midventilation compared with internal target volume for radiation therapy of pancreatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 2015; 92(3):675–682.
15. Wolthaus JWH, Schneider C, Sonke JJ et al. Mid-ventilation CT scan construction from four-dimensional respiration-correlated CT scans for radiotherapy planning of lung cancer patients. Int. J. Radiat. Oncol. Biol. Phys. 2006; 65(5):1560–1571.
16. Wolthaus JWH, Sonke JJ, van Herk M et al. Comparison of Different Strategies to Use Four-Dimensional Computed Tomography in Treatment Planning for Lung Cancer Patients. Int. J. Radiat. Oncol. Biol. Phys. 2008; 70(4):1229–1238.
17. van Dijk IWEM, Huijskens SC, de Jong R et al. Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference? Acta Oncol. (Madr). 2017; 56(8):1065–1071.
18. Keall PJ, Mageras GS, Balter JM et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76a). Med. Phys. 2006; 33(10):3874–3900.
19. Tyc VL, Klosky JL, Kronenberg M et al. Children’s Distress in Anticipation of Radiation Therapy Procedures. Child. Heal. Care 2002; 31(1):11–27.
20. Filin A, Treisman S, Peles Bortz A. Radiation Therapy Preparation by a Multidisciplinary Team for Childhood Cancer Patients Aged 3½ to 6 Years. J. Pediatr. Oncol. Nurs. 2009; 26(2):81–85.
90
21. Bucholtz JD. Comforting children during radiotherapy. Oncol. Nurs. Forum 1994; 21(6):987–94. 22. Brody AS, Frush DP, Huda W, Brent RL. Radiation Risk to Children From Computed Tomography.
Pediatrics 2007; 120(3):677–682. 23. De Jong R, Lens E, Van Herk M et al. Optimizing cone-beam CT presets for children to reduce
imaging dose illustrated with craniospinal axis. Radiother. Oncol. 2014; 111:S109–S110. 24. Slovis TL. Children, computed tomography radiation dose, and the As Low As Reasonably
Achievable (ALARA) concept. Pediatrics 2003; 112(4):971–972. 25. Miglioretti DL, Johnson E, Williams A et al. The Use of Computed Tomography in Pediatrics and
the Associated Radiation Exposure and Estimated Cancer Risk. JAMA Pediatr. 2013; 167(8):700. 26. Lens E, van der Horst A, Kroon PS et al. Differences in respiratory-induced pancreatic tumor
motion between 4D treatment planning CT and daily cone beam CT, measured using intratumoral fiducials. Acta Oncol. (Madr). 2014; 53(9):1257–1264.
27. Rombi B, Vennarini S, Vinante L et al. Proton radiotherapy for pediatric tumors: review of first clinical results. Ital. J. Pediatr. 2014; 40(1):74.
28. Timmermann B. Proton Beam Therapy for Childhood Malignancies: Status Report. Klin. Pädiatrie 2010; 222(03):127–133.
29. Engelsman M, Schwarz M, Dong L. Physics Controversies in Proton Therapy. Semin. Radiat. Oncol. 2013; 23(2):88–96.
30. Kavanagh A, Evans PM, Hansen VN, Webb S. Obtaining breathing patterns from any sequential thoracic x-ray image set. Phys. Med. Biol. 2009; 54(16):4879–4888.
31. Marchant TE, Amer AM, Moore CJ. Measurement of inter and intra fraction organ motion in radiotherapy using cone beam CT projection images. Phys. Med. Biol. 2008; 53(4):1087–1098.
32. Rit S, van Herk M, Zijp L, Sonke J-J. Quantification of the Variability of Diaphragm Motion and Implications for Treatment Margin Construction. Int. J. Radiat. Oncol. 2012; 82(3):e399–e407.
33. Chen M, Bai J, Siochi RAC. A new method of diaphragm apex motion detection from 2D projection images of mega-voltage cone beam CT. Phys. Med. Biol. 2013; 58(3):715–733.
34. Zijp L, Sonke J-J, van Herk M. Extraction of the respiratory signal from sequential thorax cone-beam X-ray images. onf Use Comput Radiat Ther (ICCR)., 2004:507–509.
35. Balter JM, Ten Haken RK, Lawrence TS et al. Uncertainties in CT-based radiation therapy treatment planning associated with patient breathing. Int. J. Radiat. Oncol. 1996; 36(1):167–174.
36. de Boer HCJ, Heijmen BJM. eNAL: An Extension of the NAL Setup Correction Protocol for Effective Use of Weekly Follow-up Measurements. Int. J. Radiat. Oncol. 2007; 67(5):1586–1595.
37. Seppenwoolde Y, Shirato H, Kitamura K et al. Precise and real-time measurement of 3D tumor motion in lung due to breathing and heartbeat, measured during radiotherapy. Int. J. Radiat. Oncol. 2002; 53(4):822–834.
38. Liu HH, Balter P, Tutt T et al. Assessing Respiration-Induced Tumor Motion and Internal Target Volume Using Four-Dimensional Computed Tomography for Radiotherapy of Lung Cancer. Int. J. Radiat. Oncol. 2007; 68(2):531–540.
39. Lens E, van der Horst A, Versteijne E et al. Considerable pancreatic tumor motion during breath-holding. Acta Oncol. (Madr). 2016; 55(11):1360–1368.
40. Kuhns LR, Berger PE, Zierolf E, Spielman S. Duration of voluntary apnea in children. Radiology 1979; 132(2):355.
41. To DT, Kim JP, Price RG et al. Impact of incorporating visual biofeedback in 4D MRI. J. Appl. Clin. Med. Phys. 2016; 17(3):128–137.
42. Lagendijk JJW, Raaymakers BW, van Vulpen M. The Magnetic Resonance Imaging–Linac System. Semin. Radiat. Oncol. 2014; 24(3):207–209.
43. Seregni M, Paganelli C, Lee D et al. Motion prediction in MRI-guided radiotherapy based on interleaved orthogonal cine-MRI. Phys. Med. Biol. 2016; 61(2):872–887.
44. Kubo HD, Hill BC. Respiration gated radiotherapy treatment: a technical study. Phys. Med. Biol. 1996; 41(1):83–91.
45. Kini VR, Vedam SS, Keall PJ et al. Patient training in respiratory-gated radiotherapy. Med. Dosim.
91
2003; 28(1):7–11. 46. DuHamel KN, Redd WH, Vickberg SM. Behavioral interventions in the diagnosis, treatment and
rehabilitation of children with cancer. Acta Oncol. 1999; 38(6):719–34.
92
Appendix 5
Supplementary Figure 5.1 | Schematic overview of diaphragm motion tracking and respiratory-induced diaphragm motion characteristics and analysis.
a) A region of interest (ROI) was selected including the top of the right diaphragm (blue box). b) The derivative of the grey values along the cranial-caudal (CC) direction (indicated by v) of the ROI was taken and
the pixel values along each line (perpendicular to the CC direction, indicated by u) were summed, creating a one-dimensional image. This was repeated for all projection images.
c) Accumulating all one-dimensional images created a two-dimensional Amsterdam Shroud image. Subsequently, the pixel coordinates corresponding to the position of the diaphragm top were translated to millimeters relative to the patients’ planned isocenter and were also corrected for the geometry of the CBCT scanner, resulting in a respiratory signal as shown in d).
d) The amplitude was defined as the difference between the diaphragm position in the end-inhale and end-exhale phase. The cycle time equals inspiratory time plus expiratory time. Results were analysed per fraction,
e) per patient (over multiple fractions) and over the whole study population.
93
Supplementary Figure 5.2A | Scatter plots showing the relations (Spearman’s ρ and p-value) between respiratory-induced diaphragm motion characteristics and height. Dots and triangles represent patients treated without general anaesthesia (GA), and patients treated under GA, respectively (significance level: p<0.05). Note: y-axes differ in range.
94
Supplementary Figure 5.2B | Scatter plots showing the relations (Spearman’s ρ and p-value) between respiratory-induced diaphragm motion characteristics and weight. Dots and triangles represent patients treated without general anaesthesia (GA), and patients treated under GA, respectively (significance level: p<0.05). Note: y-axes differ in range.
95
Chapter 6 The effectiveness of 4DCT in children and adults: a pooled analysis
Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Coen R.N. Rasch, Tanja
Alderliesten, Arjan Bel
Journal of Applied Clinical Medical Physics 2019; Volume 20(1): 276-283
Doi: 10.1002/acm2.12488
97
Abstract
Purpose: While four-dimensional computed tomography (4DCT) is extensively used in adults, reluctance remains to use 4DCT in children. Day-to-day (interfractional) variability and irregular respiration (intrafractional variability) have shown to be limiting factors of 4DCT effectiveness in adults. In order to evaluate 4DCT applicability in children, the purpose of this study is to quantify inter- and intrafractional variability of respiratory motion in children and adults. The pooled analysis enables a solid comparison to reveal if 4DCT application for planning purposes in children could be valid. Methods: We retrospectively included 90 patients (45 children and 45 adults), for whom the diaphragm was visible on abdominal/thoracic free-breathing cone beam CTs (480 pediatric, 524 adult CBCTs). For each CBCT, the cranial-caudal position of end-exhale and end-inhale positions of the right diaphragm dome were manually selected in the projection images. The difference in position between both phases defines the amplitude. Cycle time equaled inspiratory plus expiratory time. We analyzed the variability of the inter- and intrafractional respiratory-induced diaphragm motion. Results: Ranges of respiratory motion characteristics were large in both children and adults (amplitude: 4-17mm vs. 5-24mm, cycle time 2.1-3.9s vs. 2.7-6.5s). The mean amplitude was slightly smaller in children than in adults (10.7mm vs. 12.3mm; p=0.06). Interfractional amplitude variability was statistically significantly smaller in children than in adults (1.4mm vs. 2.2mm; p=0.00). Mean cycle time was statistically significantly shorter in children (2.9s vs. 3.6s; p=0.00). Additionally, intrafractional cycle time variability was statistically significantly smaller in children (0.5s vs. 0.7s; p=0.00). Conclusions: Overall variability is smaller in children than in adults, indicating that respiratory motion is more regular in children than in adults. This implies that a single pre-treatment 4DCT could be a good representation of daily respiratory motion in children and will be at least equally beneficial for planning purposes as it is in adults.
98
6.1 | Introduction Precise knowledge of organ motion is extremely important for high-precision image-guided radiotherapy, aiming for an optimal balance between accurate target coverage and minimizing dose to surrounding healthy tissues. As the field of radiotherapy is expanding rapidly, with proton and carbon ion therapies, the need for high accuracy is of increasing importance [1]. This holds especially in pediatric radiotherapy, where dose to healthy surrounding tissues is associated with a highly unfavorable increased risk of developing adverse events later in life [2]. Particularly, respiratory-induced organ motion is one of the main challenges to deal with during abdominal radiotherapy. Continuous developments and research have focused on accounting for respiratory-induced organ motion in radiotherapy [3–5].
Typically, safety margins surrounding the tumor and organs at risk are determined to account for inter- and intrafractional organ motion, setup variations, and delineation errors [6–8]. In adults, pre-treatment four-dimensional computed tomography (4DCT) is often acquired to quantify respiratory-induced organ motion in order to assess the intrafractional component of the safety margin. With the 4DCT technique, the image acquisition is related to the patient’s respiration and is binned in a number of uniform respiratory phases [9]. This results in a series of reconstructed 3DCT scans representing the entire respiratory cycle, thereby encompassing the full range of respiratory-induced organ motion. However, day-to-day (interfractional) variability and irregular respiration (i.e., intrafractional variability) have shown to be limiting factors of the 4DCT technique and application for treatment planning in adults [10, 11]. First of all, 4DCT acquisition captures a single time-point while there might be variability of respiratory motion during different treatment days. Adult studies have investigated the predictive value of a single 4DCT for a variety of treatment sites and have reported both positive and negative on it [12–15]. Besides, the 4DCT images are often subject to motion artifacts mostly resulting from irregular respiration, which causes misidentification of the respiratory cycles. Although these limitations are present, 4DCT provides useful information for planning purposes and is routinely applied for highly mobile tumors in adults. However, in pediatric radiotherapy a 4DCT is not commonly applied, since the 4DCT acquisition requires extra patient training and treatment time. Additionally, a 4DCT involves a slightly higher imaging dose and due to the ALARA (keeping doses As Low As Reasonable Achievable) principle, reluctance remains to use 4DCT in the pediatric population. Nevertheless, for mobile targets in the thoracic and abdominal region, such as neuroblastomas, Wilms’ tumors and lung metastases, imaging with 4DCT might yield a more precise treatment. This lowers the risk of adverse effects, but the additional imaging should be weighed against increased imaging dose to the patient.
We recently quantified respiratory-induced diaphragm motion, as a surrogate for motion of upper abdominal and thoracic target volumes and organs at risk, in 45 children [16] and concluded that this respiratory-induced diaphragm motion was smaller and more regular in children than previously reported by another group in adults [17], indicating that a pre-treatment 4DCT could also be promising in pediatric radiotherapy. However, respiratory-induced diaphragm motion in adults was quantified using a partly different methodology in lung cancer patients [17]. Since these patients suffered from lung pathologies, it is likely that respiration was affected and comparison with our pediatric data [16] was inconclusive. A solid comparison of respiratory characteristics in children and adults that could affect 4DCT image quality and its effectiveness requires the same clinical image-guided practice, analysis and similar tumor sites, i.e. excluding adults with lung tumors.
99
Therefore, in this study we aimed to quantify the inter- and intrafractional variability of respiratory-induced diaphragm motion in adults, using the exact same methodology as used in our pediatric study [16]. The pooled analysis of pediatric and adult data enables a methodologically consistent comparison of respiratory motion characteristics in children and adults in order to reveal if 4DCT application for planning purposes in children could be valid. 6.2 | Methods
Patient data
Pediatric data was available from our previous study, where respiratory-induced diaphragm motion was retrospectively analyzed during the treatment course of 45 children (median age 11; range 2-18 years) [16]. We collected information on general anesthesia (GA, n=7), and patient characteristics including age at the first radiation treatment fraction, height, weight, primary cancer diagnosis and radiation site. A detailed overview of pediatric patient characteristics is given in Huijskens et al. [16] and added in Supplementary Table 6.1. For this pooled analysis, 45 adults (median age 63; range 34-93 years), treated at our institute within the same period (2010-2016) as the pediatric group, were randomly included when the diaphragm was visible in upper abdominal or thoracic free-breathing cone beam computed tomography (CBCT) scans. To prevent bias when comparing the respiration pattern in the adult group to that of the pediatric group, lung cancer patients were excluded. Thus, the selection yielded esophageal (n=13), gastric (n=17), and pancreatic (n=15) cancer patients. Supplementary Table 6.2 provides a detailed overview of the adult patient characteristics. In our institute, abdominal compression to control respiratory motion is neither used in children, nor in adults. A general overview of all patient characteristics can be found in Table 6.1. CBCT acquisition
In our pediatric study, a total of 480 pediatric CBCTs (median 7; range 4–32 per patient) were included [16]. Acquisition parameters for pediatric CBCTs were 120 kV, 10 mA, and 10 or 40 ms exposure time per projection. The rotation varied from 200 (n=35) to 360 (n=10) degrees, resulting in a variation in number of projection images per CBCT (180–760). Adult patients received daily or weekly CBCT imaging (Synergy, Elekta Oncology systems, Crawly, UK) prior to treatment for position verification, totaling 524 CBCTs (median 11; range 2-30 per patient). Acquisition parameters for adult CBCTs were 120 kV, 10 mA, and 10 or 40 ms exposure time per projection. For all adults, the rotation yielded 360 degrees, resulting in approximately 760 projection images per CBCT.
100
Table 6.1 | Characteristics of included patients Children N = 45
(%) Adults N = 45
(%) Gender Male Female
28 17
(62) (38)
31 14
(69) (31)
Age at 1st RT fraction (years) Mean (median; range) 0-5 6-10 11-18 30-49 50-69 ≥70
11 (11; 2-18)
6 12 27
(13) (27) (60)
61 (63; 34-93)
9 26 10
(20) (58) (22)
Height (cm) Mean (median; range)
144 (148; 90-186)
175 (175; 134-203)
Weight (kg) Mean (median; range)
40 (37; 15-81)
71 (69; 52-134)
Type of primary cancer CNS tumora
Sarcomab
Neuroblastoma Renal tumorc
Otherd
Oesophagus Pancreas Stomach
20 16 4 2 3
(44) (36) (9) (4) (7)
13 15 17
(29) (33) (38)
Radiation Site (Cranio)spinal Thoracic/Mediastinal Abdominal (incl. flank) Upper abdominal
20 18 7
(44) (40) (16)
45
(100)
Total number of CBCTs 480 524 Mean (median; range) 11 (7; 4-32) 12 (11; 2-30) Rotation (degrees) 200 360 Acquisition Parameters 120kV, 10 mA, 10/40 ms
a Including: anaplastic glioma (n=1), ependymoma (n=1), germinoma pinealis (n=4), medulloblastoma (n=14). b Including: Ewing sarcoma (n=10), rhabdomyosarcoma (n=5), osteosarcoma (n=1). c Wilms’ tumor (n = 1), clear cell carcinoma (n=1). d Including: lymphoma (n=2), desmoplastic small round cell tumor (n=1)
101
Diaphragm Tracking
Identical to the methodology used for the pediatric group [16] an adapted version of the Amsterdam Shroud (AS) method [17, 18] was used to track diaphragm motion from CBCT imaging. For each CBCT, a two-dimensional AS image was created. Along the horizontal axis of this image, representing the projection images, we manually selected the projection images corresponding to end-exhale and end-inhale positions of the diaphragm. In each of those selected projection images, we then manually determined the cranial-caudal (CC) position of the top of the diaphragm (i.e., corresponding to the peak-to-peak position variation). Subsequently, the pixel coordinate corresponding to the position of the top of the diaphragm was translated to millimeters relative to the patients’ planned isocentre, by including a magnification correction to account for the difference in scale between the imaging panel and the isocentre [19]. Additionally, we corrected for the geometry of the CBCT scanner [19]. This resulted in a respiratory pattern describing the CC position of the diaphragm in end-exhale and end-inhale phases over the course of CBCT acquisition (detailed overview shown in Supplementary Figure 6.1).
Respiratory analysis
The amplitude was defined as the difference in CC position of the diaphragm between end-exhale and end-inhale phases. The cycle time described the time between two consecutive end-inhale positions. Day-to-day variation was expressed as interfractional variability (i.e., the SD over mean amplitudes from each fraction), and irregular breathing was expressed as intrafractional variability (root mean square of the SDs from each fraction).
For the adult patients, we calculated the same parameters as in our pediatric study; mean amplitude, interfractional variability, and intrafractional variability (see schematic overview; Supplementary Figure 6.1). For the whole patient group, including both children and adults, we calculated the group mean amplitude by averaging the patients’ mean amplitude, the group interfractional variability by averaging the patients’ interfractional variabilities, and the group intrafractional variability by averaging the patients’ intrafractional variabilities. Calculations of these respiratory parameters were also computed for the cycle time. Statistical Analysis Since not all data fitted the normal distribution (tested with the Shapiro-Wilks test), differences in mean amplitude, mean cycle time, and inter- and intrafractional variabilities in children and adults were tested for significance with the Mann-Whitney U test, considering p<0.05 significant. This comparison also provides insight into possible explanations on respiratory-induced motion based on continuous values of age, height and weight. Therefore, we used the Spearman’s correlation test (significance level p<0.05) to test for possible relationships between respiratory-induced diaphragm motion parameters and patient-specific factors (age, height, and weight). For the pediatric group separately, we tested with the Mann-Whitney U-test (significance level p<0.05) whether respiratory parameters of children treated under GA (n=7, age range 2-11 years) differed from children treated without GA in a similar age range (n=12, age range 3-10).
102
Figure 6.1 | Boxplots showing the distributions of the individual means and standard deviations of the amplitude (upper row, a, b, and c) and cycle time (bottom row, d, e, and f) of respiratory-induced diaphragm motion in children (blue) and adults (red). Boxes: median value and upper and lower quartiles; whiskers: lowest and highest data point within 1.5 x interquartile range; circles: outliers. Note: y-axes differ in range. * significant differences (p<0.05)
103
6.3 | Results The differences in respiratory-induced diaphragmatic motion parameters between children and adults are summarized in Figure 6.1. The mean amplitude was slightly smaller in children than in adults (average: 10.7 mm vs. 12.3 mm, range: 4.1-17.4 mm vs. 5.1-24.4 mm), however statistically insignificant (p=0.06). Interfractional amplitude variability was statistically significantly smaller in children than in adults (average: 1.4 mm vs. 2.2 mm, range: 0.3-3.9 mm vs. 0.4-7.2 mm; p=0.00). Mean cycle time was statistically significantly shorter in children (average: 2.9 s vs. 3.6 s, range: 2.1-3.9 s vs. 2.7-6.5 s; p=0.00), since children breath faster than adults. Additionally, intrafractional cycle time variability was statistically significantly smaller in children (0.5 s vs. 0.7 s, range: 0.2-1.5 s vs. 0.2-3.2 s; p=0.00). The intrafractional amplitude variability was significantly smaller in children treated under GA (1.6 mm) than in children of similar ages treated without GA (2.4 mm), other respiratory-induced diaphragm motion characteristics did not differ (Figure 6.2). The repeated analysis, with exclusion of the children treated under GA, did not change our results, when comparing children and adults.
Possible relationships between respiratory-induced diaphragm motion parameters and patient-specific factors, over continuous values of age, height, and weight, are given in Figure 6.2 and Supplementary Figure 6.2. All correlations of mean amplitude and mean cycle time with age, height, and weight were statistically significant (p<0.05). However, values of Spearman’s correlation coefficients were low (amplitude ρ≤0.30; cycle time ρ≤0.50). Correlations of inter- and intrafractional amplitude and cycle time variabilities with patient-specific factors were all statistically insignificant, except for interfractional amplitude variability with age (p=0.00; ρ=0.36) and weight (p=0.04; ρ=0.22), and intrafractional cycle time variability and age (p=0.00; ρ=0.34).
104
Figure 6.2 | Scatter plots describing relations (Spearman’s ρ and p-value) between the amplitude (upper row, a, b, and c) and cycle time (bottom row, d, e, and f) of respiratory-induced diaphragm motion and age (significance level: p<0.05). Dots (light blue), open circles (dark blue) and triangles (red) represent respectively pediatric patients treated without and with anesthesia, and adult patients. Note: y-axes differ in range.
105
6.4 | Discussion Respiratory motion characteristics of 90 patients, including 45 children and 45 adults, analyzed with identical methodology in 1004 CBCTs [16], were compared to reveal the effectiveness of 4DCT application for planning purposes in pediatric radiotherapy. This comprehensive dataset shows small but statistically significant differences in respiratory-induced diaphragm motion in children and adults. We found that respiratory motion in children during the treatment course is more regular, indicating that a 4DCT will be at least equally beneficial for planning purposes as it is in adults. Additionally, large ranges of mean amplitude and mean cycle time in both children and adults confirm that respiratory motion is patient-specific and requires an individualized approach (e.g., based on 4DCT) to account for. This was emphasized by weak correlations between all respiratory parameters and the patient-specific factors.
Unexpectedly, we found that interfractional variability of the amplitude was statistically significantly smaller in children than in adults, meaning that over all fractions the respiratory amplitude was more stable in children. This could be explained by the fact that, since it is known that patients experience radiotherapy as a stressful procedure [20], more attention in the clinic is paid towards comforting the child and reducing anxiety [21–23], while this is less introduced in the clinic for adults. This might have also led to a more constant cycle time during each fraction in children than in adults, as shown by the smaller intrafractional variability of cycle time in children.
These present results indicate that a single pre-treatment measurement of respiratory-induced motion with 4DCT could be a good representation for motion in children during radiotherapy. Moreover, large variation in amplitude and cycle time in both children and adults confirms that a more individualized approach with 4DCT can be effective for children as well. Recently, discussion is ongoing on radiation risks in children from medical imaging [24–28]. However, although 4DCT increases imaging dose 2-4 times compared to 3DCT [29, 30], it provides more detailed information on organ motion, leading to more precise treatment planning and potentially minimizing dose to healthy tissues. Recently, for this aim a pediatric-specific 4DCT scanning sequence and protocol was developed [31]. In our institute, we recently introduced 4DCT for children and applied the same 4DCT protocol as used for adults. However, when feasible, parameters were adjusted to achieve lower imaging doses. Similarly, in our institute, a low-dose protocol for CBCT imaging was developed and implemented for pediatric patients [32]. For all adults, CBCT imaging was acquired with 360 degrees rotation while pediatric CBCTs were acquired at lower imaging doses with 200-360 degrees rotations. This resulted in a variation in number of projection images (180 to 760) between children and adults. However, for each patient, a sufficient amount of projection images was available for tracking diaphragm motion, representing sufficient breathing cycles (approximately 10-30) for the calculation of our parameters.
Although the differences in respiratory motion characteristics between children and adults are smaller, our present results on intrafractional organ motion are in line with findings from our previous study, in which we demonstrated that interfractional abdominal organ motion in children differed from that in adults [33]. This underscores the need, also in children, for a more individualized approach using 4DCT to define safety margins. However, all-encompassing safety margins cannot be defined solely based on inter- and intrafractional motion. Setup variations and delineation errors [6], should also be taken into account, but to our knowledge have not been reported on for pediatric radiotherapy. Additionally, since respiratory-induced diaphragm motion is used as a surrogate for abdominal organ or tumor motion, uncertainties need to be taken into account when these results are used for treatment planning purposes. Therefore, assumptions regarding potential margin reduction and
106
dosimetric impact should be interpreted with caution. Nevertheless, as recently recommended by the Paediatric Radiation Oncology Society (PROS) [34], consensus needs to be reached towards accurate margin definition in pediatric radiotherapy. With this study, we take another step closer towards developing guidelines for the appropriate approach to define accurate safety margins in pediatric radiotherapy.
5.5 | Conclusions
In conclusion, these present results indicate that for children, a single pre-treatment measurement of respiratory-induced motion with 4DCT could be effective and could provide a good representation for intrafractional motion during radiotherapy. Moreover, large variation in amplitude and cycle time in both children and adults, confirms that 4DCT could be used for a more precise and individualized approach in pediatric radiotherapy, thereby aiming for more accurate safety margins and minimizing the risk of adverse events.
107
References
1. Engelsman M, Schwarz M, Dong L. Physics Controversies in Proton Therapy. Semin. Radiat. Oncol. 2013; 23(2):88–96.
2. Geenen MM, Cardous-Ubbink MC, Kremer LCM et al. Medical Assessment of Adverse Health Outcomes in Long-term Survivors of Childhood Cancer. JAMA 2007; 297(24):2705.
3. Keall PJ, Mageras GS, Balter JM et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76a). Med. Phys. 2006; 33(10):3874–3900.
4. Liu HH, Balter P, Tutt T et al. Assessing Respiration-Induced Tumor Motion and Internal Target Volume Using Four-Dimensional Computed Tomography for Radiotherapy of Lung Cancer. Int. J. Radiat. Oncol. 2007; 68(2):531–540.
5. Kubo HD, Hill BC. Respiration gated radiotherapy treatment: a technical study. Phys. Med. Biol. 1996; 41(1):83–91.
6. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 7. Morgan-Fletcher SL. Prescribing, Recording and Reporting Photon Beam Therapy (Supplement
to ICRU Report 50), ICRU Report 62 . ICRU, pp. ix+52, 1999 (ICRU Bethesda, MD) $65.00 ISBN 0-913394-61-0. Br. J. Radiol. 2001; 74(879):294–294.
8. McKenzie A, van Herk M, Mijnheer B. Margins for geometric uncertainty around organs at risk in radiotherapy. Radiother. Oncol. 2002; 62(3):299–307.
9. Rietzel E, Pan T, Chen GTY. Four-dimensional computed tomography: Image formation and clinical protocol. Med. Phys. 2005; 32(4):874–889.
10. Watkins WT, Li R, Lewis J et al. Patient-specific motion artifacts in 4DCT. Med. Phys. 2010; 37(6Part1):2855–2861.
11. Yamamoto T, Langner U, Loo BW et al. Retrospective Analysis of Artifacts in Four-Dimensional CT Images of 50 Abdominal and Thoracic Radiotherapy Patients. Int. J. Radiat. Oncol. 2008; 72(4):1250–1258.
12. Ge J, Santanam L, Noel C, Parikh PJ. Planning 4-Dimensional Computed Tomography (4DCT) Cannot Adequately Represent Daily Intrafractional Motion of Abdominal Tumors. Int. J. Radiat. Oncol. 2013; 85(4):999–1005.
13. Lens E, van der Horst A, Kroon PS et al. Differences in respiratory-induced pancreatic tumor motion between 4D treatment planning CT and daily cone beam CT, measured using intratumoral fiducials. Acta Oncol. (Madr). 2014; 53(9):1257–1264.
14. Rankine L, Wan H, Parikh P et al. Cone-Beam Computed Tomography Internal Motion Tracking Should Be Used to Validate 4-Dimensional Computed Tomography for Abdominal Radiation Therapy Patients. Int. J. Radiat. Oncol. 2016; 95(2):818–826.
15. Jin P, Hulshof MCCM, van Wieringen N et al. Interfractional variability of respiration-induced esophageal tumor motion quantified using fiducial markers and four-dimensional cone-beam computed tomography. Radiother. Oncol. 2017; 124(1):147–154.
16. Huijskens SC, van Dijk IWEM, Visser J et al. Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy. Radiother. Oncol. 2017; 123(2):263–269.
17. Rit S, van Herk M, Zijp L, Sonke J-J. Quantification of the Variability of Diaphragm Motion and Implications for Treatment Margin Construction. Int. J. Radiat. Oncol. 2012; 82(3):e399–e407.
18. Zijp L, Sonke J-J, van Herk M. Extraction of the respiratory signal from sequential thorax cone-beam X-ray images. onf Use Comput Radiat Ther (ICCR)., 2004:507–509.
19. Marchant TE, Amer AM, Moore CJ. Measurement of inter and intra fraction organ motion in radiotherapy using cone beam CT projection images. Phys. Med. Biol. 2008; 53(4):1087–1098.
20. Tyc VL, Klosky JL, Kronenberg M et al. Children’s Distress in Anticipation of Radiation Therapy Procedures. Child. Heal. Care 2002; 31(1):11–27.
21. DuHamel KN, Redd WH, Vickberg SMJ. Behavioral interventions in the diagnosis, treatment and rehabilitation of children with cancer. Acta Oncol. (Madr). 1999; 38(6):719–734.
22. Filin A, Treisman S, Peles Bortz A. Radiation Therapy Preparation by a Multidisciplinary Team
108
for Childhood Cancer Patients Aged 3½ to 6 Years. J. Pediatr. Oncol. Nurs. 2009; 26(2):81–85. 23. Bucholtz JD. Comforting children during radiotherapy. Oncol. Nurs. Forum 1994; 21(6):987–94. 24. Andronikou S. Response to Dr. Frush. Pediatr. Radiol. 2017; 47(1):122–123. 25. Andronikou S. Letting go of what we believe about radiation and the risk of cancer in children.
Pediatr. Radiol. 2017; 47(1):113–115. 26. Cohen M. Reply to Dr. Andronikou: Disavowing the ALARA concept in pediatric imaging.
Pediatric radiology. Pediatr. Radiol. 2017; 47:116–7. 27. Frush DP. Reply to Dr. Andronikou: Holding on to informed use of diagnostic imaging using
ionizing radiation. Pediatr. Radiol. 2017; 47(1):119–121. 28. Doss M. Disavowing the ALARA concept in pediatric imaging. Pediatr. Radiol. 2017; 47(1):118–
118. 29. Mori S, Ko S, Ishii T, Nishizawa K. Effective Doses in Four-Dimensional Computed Tomography
for Lung Radiotherapy Planning. Med. Dosim. 2009. doi:10.1016/j.meddos.2008.08.002. 30. Hubbard P, Callahan J, Cramb J et al. Audit of radiation dose delivered in time-resolved four-
dimensional computed tomography in a radiotherapy department. J. Med. Imaging Radiat. Oncol. 2015; 59(3):346–352.
31. Lavan N CE, Burland H, Ockwell C, Saran FH MH. Abstracts From the 48th Congress of the International Society of Paediatric Oncology (SIOP) Dublin, Ireland. Pediatr. Blood Cancer 2016; 63(Suppl 3):S5-321.
32. De Jong R, Lens E, Van Herk M et al. Optimizing cone-beam CT presets for children to reduce imaging dose illustrated with craniospinal axis. Radiother. Oncol. 2014; 111:S109–S110.
33. van Dijk IWEM, Huijskens SC, de Jong R et al. Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference? Acta Oncol. (Madr). 2017; 56(8):1065–1071.
34. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
109
Appendix 6
Supplementary Figure 6.1 | Schematic overview of diaphragm motion tracking and respiratory-induced diaphragm motion characteristics and analysis, acquired from [16].
a) A region of interest (ROI) was selected including the top of the right diaphragm (blue box). b) The derivative of the grey values along the cranial-caudal (CC) direction (indicated by v) of the ROI was taken and
the pixel values along each line (perpendicular to the CC direction, indicated by u) were summed, creating a one-dimensional image. This was repeated for all projection images.
c) Accumulating all one-dimensional images created a two-dimensional Amsterdam Shroud image. Subsequently, the pixel coordinates corresponding to the position of the diaphragm top were translated to millimeters relative to the patients’ planned isocenter and were also corrected for the geometry of the CBCT scanner, resulting in a respiratory signal as shown in d).
d) The amplitude was defined as the difference between the diaphragm position in the end-inhale and end-exhale phase. The cycle time equals inspiratory time plus expiratory time. Results were analysed per fraction,
e) per patient (over multiple fractions) and over the whole study population.
110
Supplementary Figure 6.2A | Scatter plots describing relations (Spearman’s ρ and p-value) between the amplitude (upper row, a, b, and c) and cycle time (bottom row, d, e, and f) of respiratory-induced diaphragm motion and height (significance level: p<0.05). Dots (light blue), open circles (dark blue) and triangles (red) represent respectively pediatric patients treated without and with anesthesia, and adult patients. Note: y-axes differ in range.
111
Supplementary Figure 6.2B | Scatter plots describing relations (Spearman’s ρ and p-value) between the amplitude (upper row, a, b, and c) and cycle time (bottom row, d, e, and f) of respiratory-induced diaphragm motion and weight (significance level: p<0.05). Dots (light blue), open circles (dark blue) and triangles (red) represent respectively pediatric patients treated without and with anesthesia, and adult patients. Note: y-axes differ in range.
1 M Ewing sarcoma 10.9 137 28 Thorax 6 2 F Ewing sarcoma 11.6 155 38 Thorax 27 3 M Medulloblastoma 11.0 158 46 CSI 4 4 M Medulloblastoma 10.9 150 33 CSI 5 5 F Ewing sarcoma 17.9 164 53 Thorax 17 6a M Medulloblastoma 6.6 110 18 CSI 5 7 F Ewing sarcoma 12.5 151 70 Thorax 32 8 M Medulloblastoma 14.1 175 36 CSI 9 9 F Medulloblastoma 8.2 139 28 CSI 7
10 F Neuroblastoma 6.2 103 18 Thorax 23
11 M CCSK 8.6 126 23 Abdomen 5 12 M Ewing sarcoma 16.8 184 62 Thorax 8
13a,b M Medulloblastoma 7.6 126 24 CSI 6 14 F Germinoma 9.3 139 43 CSI 6 15 M Medulloblastoma 6.7 129 24 CSI 5 16 F Hodgkin lymphoma 15.9 163 51 Thorax 6 17 F Medulloblastoma 7.0 118 22 CSI 7 18 M Medulloblastoma 11.0 143 28 CSI 4 19b M Pineal germinoma 15.8 159 48 CSI 17 20 M Pineal germinoma 13.3 166 48 CSI 6 21 M Medulloblastoma 6.2 123 23 CSI 10 22 M Ewing sarcoma 14.7 177 65 Thorax 10 23 F Ewing sarcoma 15.8 173 50 Thorax 8 24 25a
F F
B-cell lymphoma Ependymoma
16.8 5.6
163 113
62 20
Thorax Spinal
12 21
26 F Ewing sarcoma 13.1 148 43 Thorax 10 27 M ERMS 16.8 186 64 Thorax 5 28 M Glioma 7.8 132 31 CSI 14 29 F Medulloblastoma 14.7 159 50 CSI 9 30a M Medulloblastoma 5.1 109 17 CSI 8 31 F Neuroblastoma 5.3 115 24 Abdomen 7 32 F Medulloblastoma 12.5 156 37 CSI 6 33 M RMS 10.9 142 37 Thorax 25 34 M Wilms’ tumour 15.6 167 67 Abdomen 7 35 M Germinoma 16.9 172 68 CSI 6 36 M DSRCT 9.9 137 26 Abdomen 5 37 M ERMS 3.3 106 17 Abdomen 24 38 M Neuroblastoma 4.7 118 22 Thorax 6 39 F Ewing sarcoma 17.4 154 78 Thorax 16 40a M Medulloblastoma 4.9 105 18 CSI 10 41 M Ewing sarcoma 17.8 182 81 Thorax 28 42 F Osteosarcoma 15.1 159 53 Thorax 6 43a M ERMS 10.7 132 28 Abdomen 4 44a M Neuroblastoma 2.2 90 15 Abdomen 6 45 M RMS 14.4 172 59 Thorax 12
Abbreviations: CBCT = cone beam computed tomography; M = male; F = female; CCSK = Clear Cell Sarcoma Kidney; (E)RMS = (Embryonal) Rhabdomyosarcoma; DSRCT = Desmoplastic small round cell tumour; CSI = Craniospinal Irradiation a Patients were treated under general anaesthesia (GA) b Patients were treated in prone position
114
Supplementary Table 6.2 | Adult patient characteristics No. Gender Tumour Age at start RT (years) Height (cm) Weight (kg) Number of CBCT
1 F Gastric carcinoma 70 169 53 12 2 M Pancreatic cancer 53 203 80 2
3 F Oesaphagus carcinoma 69 169 62 7
4 F Gastric carcinoma 45 175 62 7 5 M Pancreatic cancer 69 183 66 15
6 M Gastric carcinoma 63 175 72 25
7 F Pancreatic cancer 52 165 67 15
8 M Pancreatic cancer 71 180 60 11
9 M Pancreatic cancer 70 180 65 11
10 F Pancreatic cancer 74 157 54 15
11 F Oesaphagus carcinoma 68 164 66 8
12 M Pancreatic cancer 68 175 62 11
13 M Oesaphagus carcinoma 44 184 134 12
14 M Gastric carcinoma 52 173 75 6
15 M Oesaphagus carcinoma 66 180 71 2
16 M Gastric carcinoma 66 189 68 11
17 M Pancreatic cancer 60 174 62 25
18 M Oesaphagus carcinoma 93 168 81 14
19 F Gastric carcinoma 55 176 54 8
20 M Pancreatic cancer 66 171 80 15
21 M Oesaphagus carcinoma 65 188 87 12
22 M Gastric carcinoma 56 188 79 25
23 M Oesaphagus carcinoma 42 169 70 5
24 M Gastric carcinoma 52 176 68 6
25 M Pancreatic cancer 42 183 78 15
26 M Gastric carcinoma 62 176 72 6
27 M Gastric carcinoma 52 166 74 6
28 M Gastric carcinoma 68 173 60 5
29 F Gastric carcinoma 41 161 67 6
30 F Oesaphagus carcinoma 34 177 96 7
31 M Gastric carcinoma 43 185 84 5
32 M Oesaphagus carcinoma 70 172 73 10
33 M Oesaphagus carcinoma 46 174 68 9
34 F Pancreatic cancer 77 167 64 23
35 M Oesaphagus carcinoma 67 178 72 9
36 F Pancreatic cancer 67 134 52 12
37 F Pancreatic cancer 58 170 60 14
38 M Gastric carcinoma 61 179 70 6
39 M Gastric carcinoma 62 171 69 25
40 M Pancreatic cancer 78 184 83 30 41 M Pancreatic cancer 80 181 73 11
42 M Gastric carcinoma 51 192 70 7
43 F Oesaphagus carcinoma 46 168 70 24 44 F Oesaphagus carcinoma 75 165 69 7 45 M Gastric carcinoma 68 170 54 7
Abbreviations: CBCT = cone beam computed tomography; M = male; F = female, RT = radiotherapy
115
Chapter 7 Predictive value of pediatric respiratory-induced diaphragm motion quantified using pre-treatment 4DCT and CBCTs Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Coen R.N. Rasch, Tanja
Alderliesten, Arjan Bel
Radiation Oncology 2018; Volume 13(1): 198
Doi: 10.1186/s13014-018-1143-6
117
Abstract
Purpose: In adults, a single pre-treatment four-dimensional CT (4DCT) acquisition is often used to account for respiratory-induced target motion during radiotherapy. However, studies have indicated that a 4DCT is not always representative for respiratory motion. Our aim was to investigate whether respiratory-induced diaphragm motion in children on a single pre-treatment 4DCT can accurately predict respiratory-induced diaphragm motion as observed on daily cone beam CTs (CBCTs). Methods: Twelve patients (mean age 14.5yrs; range 8.6–17.9yrs) were retrospectively included based on visibility of the diaphragm on abdominal or thoracic imaging data acquired during free breathing. A 4DCT for planning purposes and daily/weekly CBCTs (total 125; range 4–29 per patient) acquired prior to dose delivery were available. The amplitude, corresponding to the difference in position of the diaphragm in cranial-caudal direction in end-inspiration and end-expiration phases, was extracted from the 4DCT (A4DCT). The amplitude in CBCTs (ACBCT) was defined as displacement between averaged in- and expiration diaphragm positions on corresponding projection images, and the distribution of ACBCT was compared to A4DCT (one-sample t-test, significance level p<0.05). Results: Over all patients, the mean A4DCT was 10.4 mm and the mean ACBCT 11.6 mm. For 9/12 patients, A4DCT differed significantly (p<0.05) from ACBCT. Differences >3 mm were found in 69/125 CBCTs (55%), with A4DCT mostly underestimating ACBCT. For 7/12 patients, diaphragm positions differed significantly from the baseline position. Conclusion: Respiratory-induced diaphragm motion determined on 4DCT does not accurately predict the daily respiratory-induced diaphragm motion observed on CBCTs, as the amplitude and baseline position differed statistically significantly in the majority of patients. Regular monitoring of respiratory motion during the treatment course using CBCTs could yield a higher accuracy when a daily adaptation to the actual breathing amplitude takes place.
118
7.1 | Introduction
Respiratory motion during radiotherapy may lead to uncertainties in radiation dose delivery, and accounting for it is challenging. Continuous efforts in the field have led to innovative methods to deal with respiratory motion, such as breath holding, beam gating or tracking, or online visualizing of respiratory motion (e.g., magnetic resonance (MR) guidance) [1]. These techniques increase treatment time and clinical workload, and often require patient training which might lead to additional patient distress and anxiety. Although children could also benefit from these techniques [2, 3], it is known that they experience radiotherapy already as a stressful procedure [4–6], and therefore the use of these techniques remains limited in pediatric radiotherapy.
Respiratory motion is more commonly accounted for by the use of an internal margin (IM) that encompasses the clinical target volume (CTV), defining an internal target volume (ITV) [7]. This leads to unfavorable large margins, thereby increasing dose to surrounding healthy tissues. The mid-ventilation based planning target volume (PTV) approach accounts for both respiratory motion and day-to-day geometrical variations and achieves smaller margins [8–10]. No matter which approach is used, in order to assess respiratory motion, a pre-treatment respiratory-correlated four-dimensional computed tomography (4DCT) is essential.
In adults, a single pre-treatment 4DCT is used extensively to assess respiratory motion [11–13]. Our previous study showed that respiratory-induced diaphragm motion throughout the treatment course was more stable in children than previously reported by others in adults [14]. This implies that a single measurement could be more representative in children than in adults and suggests that a pre-treatment 4DCT in children could be at least equally beneficial as it is in adults [14]. To our knowledge, only few institutes have clinically introduced 4DCT for pediatric radiotherapy planning purposes and reported on this [15–17]. Generally, the conclusion was that 4DCT is an effective tool to accurately determine respiratory-induced organ motion (e.g., liver, spleen, kidneys) for pediatric specific cases, providing the data on respiratory motion needed to define margins, thereby stressing the need for individualized margins [15–17]. However, studies on adult patients have also indicated that respiratory motion, as measured on 4DCT, is not always representative for respiratory motion during the subsequent treatment course [11–13, 18]. Therefore, a single pre-treatment measurement for planning purposes might be a misrepresentation and could lead to under- or overestimating respiratory motion, yielding insufficient target coverage or undesired dose to organs at risk (OARs) [19]. In previous pediatric studies, respiratory-induced organ motion was only measured within a single 4DCT per patient [15–17], without assessing how representative the 4DCT is for respiratory-induced motion during the treatment course. To assess the daily respiratory-induced motion during the treatment course in children, cone beam CT (CBCT) scans acquired for position verification can be used [14]. With increasing use of 4DCT in pediatric radiotherapy, assessment of the predictive value of the measurements on 4DCT is essential to take full advantage of this technique.
Therefore, the aim of this study was to investigate if respiratory-induced diaphragm motion during radiotherapy in children, as a surrogate for respiratory-induced abdominal motion, can be accurately predicted by a single measurement based on a pre-treatment 4DCT. We also analyzed possible time trends in respiratory-induced diaphragm motion over the complete treatment course. Finally, to investigate if measurements on CBCTs could be predictive for respiratory-induced diaphragm motion that continues post-acquisition (i.e., the actual respiratory motion during dose delivery), we quantified and compared respiratory-induced diaphragm motion on two CBCTs acquired within one treatment session with an interval of minutes.
119
7.2 | Methods
Patient data
From November 2014 to December 2017, fourteen patients (mean age 14.5 years, range 8.6-17.9 years) had a 4DCT scan during free breathing for treatment planning purposes and multiple CBCT scans acquired for position verification during the treatment course. Patients were included when the complete diaphragm was visible on the 4DCT and CBCTs. Two of the fourteen eligible patients were excluded from this retrospective study; one patient received general anesthesia (GA) and the imaging data of another patient showed severe motion artefacts. Patient characteristics are listed in Table 7.1.
Table 7.1 | Patient characteristics No. Sex Tumor type Age at 4DCT
(years) Height (cm) Weight (kg) No. of CBCTs
1 M Ewingsarcoma 10.7 137 28.0 7 2 F Ewingsarcoma 16.3 162 66.5 8 3a F Ewingsarcoma 17.9 163 52.6 12 4a M Osteosarcoma 14.9 186 71.6 5 5b F Ewingsarcoma 12.5 151 70.0 29 6b M ERMS 16.1 182 57.6 19 7b M Ewingsarcoma 14.3 182 55.6 6 8 M CCS 8.6 125 23.0 5 9 F Non Hodgkin 17.1 178 86.0 11 10 F Ewingsarcoma 14.8 153 57.0 7 11b M RMS prostate 16.7 186 64.0 4 12 M Non-RMS 14.4 172 59.0 12 Abbreviations: M = male; F = female; 4DCT = four-dimensional computed tomography; CBCT = cone beam CT; (E)RMS = embryonal rhabdomyosarcoma; CCS = clear cell sarcoma a Patients had within multiple treatment sessions repetitive CBCTs (only the 1st and 2nd were included in the analysis) b Patients had two CBCTs within one treatment session
120
4DCT
The 4DCTs (LightSpeed RT16 system, General Electric Company, Waukesha WI, USA) were acquired during free breathing (using the Varian RPM system v1.7.3). The respiratory cycle was divided into ten phase bins, resulting in ten phase scans (slice thickness 2.5mm). Velocity (Velocity, version 3.1, Varian Medical Systems, Palo Alto, CA, USA) was used to perform a two-step rigid registration. The end-inspiration phase scan (i.e., the 0% phase scan) was used as a reference and was registered to the other nine phase scans. In one patient (no. 7), the 90% phase scan served as the reference, due to motion artefacts on the 0% phase scan. For all other patients and phase scans, no (severe) motion artefacts were seen that could have hampered the registrations. The right diaphragm domes were matched manually in the cranial-caudal (CC) direction, using translations only. The obtained translations resulted in the excursion of the right diaphragm dome throughout the respiratory cycle in the CC direction. The difference between the most extreme translations, typically the 0% to the 50% or 60% phase scan, was defined as the amplitude (A4DCT) (Figure 7.1A).
CBCT
For each patient, CBCT scans during free breathing (Synergy, Elekta Oncology systems, Crawly, UK) for position verification were daily and/or weekly acquired according to a customized extended no-action level (eNAL) protocol [20], totalling 125 CBCT scans (range 4-29 per patient). Six of the 12 patients had multiple CBCTs within one treatment session (total 13, range 2-5, not included in the 125), depending on their treatment protocol (e.g., stereotactic or spinal cord irradiation), or in one case a second CBCT was necessary due to artefacts. These artefacts, however, still allowed sufficient number of useable projection images for evaluation of respiratory-induced diaphragm motion. For all CBCTs, a single projection was acquired in 180ms and the energy was 120kV, tube current 10mA and 10 or 40ms exposure time per projection. The circumferential rotation varied from 200 to 360 degrees and the acquisition time varied between 35s and 120s, resulting in a variation in number of projection images per CBCT (180 to 760).
The methodology to extract the respiratory-induced diaphragm motion has been described previously [14]. In short, for each CBCT, an adapted version of the Amsterdam Shroud (AS) method was used to create an AS image [21], allowing for manual selection of the projection images corresponding to the end-inspiration and end-expiration positions of the right diaphragm dome. In each of those selected projection images, we then manually determined the CC position of the top of the right diaphragm dome. Pixel coordinates were corrected for the scanner geometry and translated relative to the patients’ isocenter [22]. This resulted in a patient- and CBCT-dependent timeframe describing the CC position of the diaphragm in end-inspiration and end-expiration phases (peaks) over the course of CBCT acquisition. The amplitude was defined as the displacement between averaged end-inspiration and averaged end-expiration diaphragm positions (ACBCT) (Figure 7.1B).
121
Statistical analysis
In order to determine whether respiratory-induced diaphragm motion observed in a single pre-treatment 4DCT accurately predicted respiratory-induced diaphragm motion during the treatment course, we tested for each patient separately whether ACBCT over all CBCTs differed from A4DCT using a one-sample t-test (significance level p<0.05). For each patient, we calculated the absolute differences between A4DCT and each ACBCT, and determined for which fractions the difference was larger than 3 mm (4DCT slice thickness is 2.5 mm).
To investigate possible time trends for each patient, we applied a linear regression analysis on ACBCT over the course of treatment. We also calculated the interfractional variability of ACBCT per patient (i.e., the SD over ACBCT). The diaphragm position on the averaged pre-treatment 4DCT (this is the averaged scan, based on all phases of the 4DCT) was considered as a reference (i.e., baseline), for which a rigid registration on bony anatomy was taken into account. We then calculated for each patient separately shifts of the diaphragm position in CC direction (i.e., the average position of the diaphragm during one CBCT) over the course of treatment. We used a one-sample t-test to test
Figure 7.1A | Left) Rigid registration of the cranial-caudal position of the right diaphragm (inside the red box) in all breathing phases. Right) The difference between the most extreme translations was defined as the amplitude (A4DCT).
Figure 7.1B | Amsterdam Shroud (AS) method to manually track the CC diaphragm position in CBCT projection images. a) Region of interest (red box). b) CC gradient filter applied and sum of all pixels creates a 1D image. C) This is repeated for all projection images, creating a 2D image. d) Detection of diaphragm positions in inhale and exhale breathing phases. e) Pixel coordinates translated to CC position. The amplitude was defined as the displacement between averaged end-inspiration and end-expiration diaphragm positions (ACBCT).
122
whether diaphragm positions on CBCT significantly differed from the baseline diaphragm position on the averaged 4DCT.
Additionally, six patients received multiple CBCTs during one fraction. For above mentioned analysis, only the first CBCT was included in the analysis. To validate if respiratory-induced diaphragm motion measurements on CBCTs could be predictive for respiratory-induced motion that continues post-acquisition (i.e., the actual respiratory motion during dose delivery), we compared the amplitude measured on the first and second CBCT (ACBCT(1) to ACBCT(2); paired t-test, significance level p<0.05). These are acquired within 4-10 minutes, which is a representative time interval between CBCT acquisition and start of dose delivery. R Software package version 3.2.1. (R foundation for statistical Computing, Austria) was used for all statistical analysis. 7.3 | Results Over all patients, the mean A4DCT was 10.4 mm (SD = 4.3 mm) and the mean ACBCT was 11.6 mm (SD = 5.7 mm). For 9 out of 12 patients, A4DCT differed statistically significantly (p<0.05) from ACBCT (Figure 7.2). Underestimation of A4DCT compared to ACBCT was found in 76% of the measurements (95/125 CBCTs), and was observed in 11 out of 12 patients. Hence, overestimation was found in 24% of the measurements (30/125 CBCTs), and was observed in 3 out of 12 patients. Differences >3 mm were found in 69 of the 125 CBCTs (55%).
For each patient, we plotted ACBCT over time of the treatment course (Supplementary Figure 7.1) where day 0 is the day of 4DCT acquisition. We found that 8 of the 12 trend lines had a negative slope. Absolute slopes larger than 0.1 mm/day were observed in 4 patients with only few data points (4 to 6 points for patients 4, 7, 8, and 11). For the other patients, with more data points, we observed by both visual inspection and linear fits, no obvious time trend (absolute slopes ranged from 0.00-0.09 mm/day). Overall, interfractional variability of ACBCT was 2.2 mm (range 0.7-4.4 mm; individual values shown in Figure 7.3). For 7 out of 12 patients, averaged diaphragm positions in CC direction observed on CBCTs differed statistically significantly (mean 7.4 mm, SD = 5.9 mm; p<0.05) from the baseline diaphragm position as measured on the averaged 4DCT (Figure 7.3).
Patients 3 and 4 had on multiple days additional CBCT scans. Patients 5, 6, 7, and 11 had only on one day an additional CBCT. The subsequent CBCTs were acquired within a time interval of 4-10 minutes. Over all six patients, ACBCT(2) was significantly different from ACBCT(1) (mean difference 2.9 mm, SD = 2.5 mm, p=0.002) (Figure 7.4). However, patient 4 (in Figure 7.4 indicated by the green cross symbol) showed significant deviations from the group measurements. We performed a sensitivity analysis by excluding this patient from the analysis. Although the average difference was now 1.7 mm (SD = 1.4 mm), ACBCT(2) remained significantly different from ACBCT(1) (p=0.033).
123
Figure 7.2 | Respiratory-induced diaphragm motion on 4DCT (blue solid lines) and CBCT data (boxplots) of 12 children during image-guided radiotherapy. Patients with * showed no significant differences (p>0.05).
124
Figure 7.3 | Open dots and dashed lines represent the baseline diaphragm position on the planning 4DCT (averaged 4DCT). Black dots represent the average positions of the diaphragm on each CBCT. Whiskers represent measured end-inspiration and end-expiration positions on 4DCT and CBCTs for each patient plotted as function of days. Day 0 is the day of 4DCT acquisition (open dots), IV = interfractional variability, * indicates a significant difference from baseline (p<0.05).
125
7.4 | Discussion In this study we investigated if respiratory-induced diaphragm motion in children during radiotherapy could be accurately predicted based on a 4DCT scan acquired prior to the start of radiotherapy. We compared the amplitude of the diaphragm displacement on 4DCT and daily/weekly CBCTs, enabling an encompassing analysis of pre-treatment respiratory-induced motion and during complete radiotherapy courses in children. This also enabled to investigate possible time trends and day-to-day variations. Our study showed that for the majority of patients (9/12 patients) respiratory-induced diaphragm motion on 4DCT differed significantly from measurements on CBCTs. Also, respiratory-induced diaphragm motion derived from CBCTs acquired within an interval of minutes was statistically significantly different. No obvious time trends in respiratory-induced diaphragm motion over the course of treatment were found, but significant baseline shifts of the diaphragm position were seen in 7/12 patients. These findings suggest that respiratory-induced diaphragm motion as measured on 4DCT was not representative for respiratory motion during the treatment course.
Although acquisition of 4DCT and CBCT scans differs, the amplitude was quantified in a similar way. During 4DCT acquisition, one breathing cycle is included per table position. This represents only a short time period, and amplitude between consecutive breathing cycles varies [14]. This uncertainty would be of a similar size (2.2 mm) as measured during the consecutive CBCTs. Additionally, a 4DCT scan is binned into 10 3D-breathing-phase scans corresponding to 10 phases of the respiratory cycle using phase binning, which already underestimates the diaphragm motion slightly [23]. For 4DCT, we quantified the amplitude as the maximal displacement between the most extreme diaphragm positions. On the other hand, CBCT acquisition time varied between 35 and 120 seconds and thus
Figure 7.4 | Significant difference (indicated by *) between amplitudes measured on the first and second CBCT acquired within one treatment session. Each different color and symbol represent different patients.
126
included more breathing cycles compared to the 4DCT. In the amplitude calculations for both types of CT scans, we averaged the diaphragm positions on the CBCT scan for in- and exhale phases and the difference between these averaged inhale and exhale diaphragm position defined the amplitude on CBCT. This is a slightly different approach as used by others [13, 24], who binned projection images corresponding to in- and exhale phases for (4D-) CBCT reconstructions, thereby averaging (e.g. blurring) the actual diaphragm positions on the reconstructed image. While different approaches have their advantages and limitations, for the comparison of 4DCT and CBCT data, in this study we chose to average the actual diaphragm positions at end-inspiration and end-expiration as measured on the corresponding projection image. This guarantees that all projection images are taken into account, and represents a realistic view of the actual motion happened.
Since we used respiratory-induced diaphragm motion as a surrogate for respiratory-induced abdominal motion, our outcomes cannot be directly applied for calculating safety margins. This was shown by Panandiker et al. who assessed intrafractional renal and diaphragm motion on free-breathing 4DCTs in 20 children, and concluded that measuring diaphragm motion alone does not reliably quantify renal motion [15]. Adult studies reported both positive and negative on using the diaphragm as a reliable surrogate for tumor or organ motion [25–27]. Two other pediatric studies have reported on intrafractional abdominal organ and tumor motion using 4DCT scans and concluded that 4DCT is an effective tool to accurately determine respiratory-induced organ motion for pediatric specific cases, leading to the desired more individualized treatment approach [16, 17]. However, in these studies, correlations of respiratory-induced organ motion with diaphragm motion were not investigated. Since respiratory-induced diaphragm motion does not necessarily correlate with tumor motion, using the diaphragm as a surrogate for abdominal and thoracic organ motion could induce some inaccuracies and uncertainties that need to be taken into account for treatment planning purposes.
A 4DCT involves a slightly higher imaging dose compared to a 3DCT and due to the ALARA principle (keeping doses as low as reasonably achievable) and previously reported radiation risks in children from CT scans [28–30], reluctance remains to use 4DCT in the pediatric population. It would be interesting to investigate the possible correlation between external thorax vertical displacement and the internal longitudinal diaphragm motion in children. In case of a strong and clear correlation, which was found for adults [31], the possibility of using an external reliable surrogate for internal respiratory-induced organ motion could decrease additional imaging dose. Since daily imaging dose adds to the total treatment dose, minimizing additional dose has to be carefully considered. Ultimately, the additional imaging dose in the pediatric population should be balanced with better treatment planning and delivery, in order to minimize dose to the healthy surrounding tissues. Especially, 4DMRI shows to be a promising tool for future image- and MR-guided pediatric radiotherapy, providing superior soft tissue contrast and higher resolution in CC direction, while avoiding ionizing radiation doses [32, 33]. The measured amplitude of respiratory-induced diaphragm motion on 4DCT was on average larger than the 6-17 mm range reported in literature [15, 16, 32]. However, patients in our cohort had an older age at treatment (mean 14.5 years, range 8.6-17.9 years) than those in other studies (ranges 1-20 years), and we excluded a patient treated under GA. Two studies divided their cohort into 2 groups based on age (>9 years); when we compared our results to their older age groups (n=9; mean 12.3 years [15] and n=18; mean 15.3 years [32]), we saw a similar range of diaphragm motion. Although different ranges of diaphragm motion have been found for younger versus older children [15, 32], no clinically significant correlation has been found in studies investigating possible relationships between
127
respiratory-induced diaphragm motion and age [14, 17]. The same holds for patients treated under GA; differences in amplitude of respiratory-induced diaphragm motion in patients treated with- or without GA were insignificant [14, 17, 32]. Outcomes reported in adult studies that investigated the predictive value of measurements done in the 4DCT are not consistent; some studies found that measurements in 4DCT did not accurately predict respiratory-induced motion as seen on daily/weekly CBCT images [11–13, 18], while others concluded that respiratory-induced motion as measured in 4DCT was representative for the daily motion during the treatment course [24, 34]. These differences mostly depended on the tumor location, considering that abdominally located tumors could also be affected by abdominal processes, while thoracically located tumors are situated closer to the mediastinum. Interestingly, in those adult studies where respiratory-induced motion measured in the 4DCT was not representative, measurements overestimated daily respiratory-induced motion [11, 13], while in our pediatric cohort, the 4DCT mostly underestimated the daily respiratory-induced diaphragm motion. To account for respiratory-induced motion using such a single measurement could possibly lead to insufficient target coverage. Therefore, our results suggest monitoring of respiratory motion with CBCT on a more regular basis, and adapt treatment plans to the actual breathing amplitude when necessary.
For 7 out of 12 patients, the averaged CC positions of the diaphragm during the treatment course differed significantly from the baseline diaphragm position as measured on the 4DCT, introducing a systematic interfractional position variation. The patient number in the present study is low and some patients only had a few CBCTs. This means that measurements regarding baseline-shifts could have been random. However, these results emphasize the benefit and need for daily imaging and monitoring to enable baseline positioning correction.
Nevertheless, present and previous results also confirm that respiratory motion in children varies from day-to-day and even within consecutive breathing cycles [14, 17]. The measured respiratory-induced diaphragm motion on CBCTs acquired within a 4-10 minute interval showed significant differences, meaning that the actual respiratory-induced motion during dose delivery can again be different than measured on the CBCT. However, this analysis was only based on a small number of repetitive CBCTs (n=13) evaluated in six patients. Future studies should involve larger imaging datasets for evaluation of measurements on CBCTs for predicting respiratory-induced motion in children. In addition, as mentioned above, it would be interesting to asses respiratory-induced motion online using CBCTs acquired at, for example, the first three treatment fractions. This would enable to identify which patients deviate from their pre-treatment measurements on 4DCT and might benefit from an adaptive approach in order to maintain appropriate tumor dose coverage.
7.5 | Conclusions
In conclusion, respiratory-induced diaphragm motion in children determined on 4DCT does not accurately predict the daily respiratory motion observed on CBCTs, as the amplitude differed statistically significantly in the majority of patients. Our results show the limitations of using a single pre-treatment 4DCT to take the patient-specific respiratory-induced diaphragm motion for treatment planning purposes into account. Regular monitoring of respiratory motion during the treatment course using CBCTs could yield a higher accuracy when a daily adaptation to the actual breathing amplitude takes place.
128
References 1. Keall PJ, Mageras GS, Balter JM et al. The management of respiratory motion in radiation
oncology report of AAPM Task Group 76a). Med. Phys. 2006; 33(10):3874–3900. 2. Claude L, Malet C, Pommier P et al. Active Breathing Control for Hodgkin’s Disease in Childhood
and Adolescence: Feasibility, Advantages, and Limits. Int. J. Radiat. Oncol. Biol. Phys. 2007; 67(5):1470–1475.
3. Demoor-Goldschmidt C, Chiavassa S, Josset S et al. Asservissement respiratoire lors d’une radiothérapie pulmonaire bilatérale pour le sarcome d’Ewing ou le néphroblastome chez des enfants et jeunes adultes : études dosimétrique et clinique de faisabilité. Cancer/Radiotherapie 2017; 21(2):124–129.
4. Filin A, Treisman S, Peles Bortz A. Radiation Therapy Preparation by a Multidisciplinary Team for Childhood Cancer Patients Aged 3½ to 6 Years. J. Pediatr. Oncol. Nurs. 2009; 26(2):81–85.
5. Bucholtz JD. Comforting children during radiotherapy. Oncol. Nurs. Forum 1994; 21(6):987–94. 6. Tyc VL, Klosky JL, Kronenberg M et al. Children’s Distress in Anticipation of Radiation Therapy
Procedures. Child. Heal. Care 2002; 31(1):11–27. 7. Griem ML. Prescribing, Recording, and Reporting Photon Beam Therapy. Radiat. Res. 1994;
138(1):146. 8. Wolthaus JWH, Schneider C, Sonke JJ et al. Mid-ventilation CT scan construction from four-
dimensional respiration-correlated CT scans for radiotherapy planning of lung cancer patients. Int. J. Radiat. Oncol. Biol. Phys. 2006; 65(5):1560–1571.
9. Lens E, Van Der Horst A, Versteijne E et al. Dosimetric advantages of midventilation compared with internal target volume for radiation therapy of pancreatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 2015; 92(3):675–682.
10. Wolthaus JWH, Sonke JJ, van Herk M et al. Comparison of Different Strategies to Use Four-Dimensional Computed Tomography in Treatment Planning for Lung Cancer Patients. Int. J. Radiat. Oncol. Biol. Phys. 2008; 70(4):1229–1238.
11. Britton KR, Starkschall G, Tucker SL et al. Assessment of Gross Tumor Volume Regression and Motion Changes During Radiotherapy for Non–Small-Cell Lung Cancer as Measured by Four-Dimensional Computed Tomography. Int. J. Radiat. Oncol. 2007; 68(4):1036–1046.
12. Ge J, Santanam L, Noel C, Parikh PJ. Planning 4-Dimensional Computed Tomography (4DCT) Cannot Adequately Represent Daily Intrafractional Motion of Abdominal Tumors. Int. J. Radiat. Oncol. 2013; 85(4):999–1005.
13. Lens E, van der Horst A, Kroon PS et al. Differences in respiratory-induced pancreatic tumor motion between 4D treatment planning CT and daily cone beam CT, measured using intratumoral fiducials. Acta Oncol. (Madr). 2014; 53(9):1257–1264.
14. Huijskens SC, van Dijk IWEM, Visser J et al. Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy. Radiother. Oncol. 2017; 123(2):263–269.
15. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
16. Kannan S, Teo BKK, Solberg T, Hill-Kayser C. Organ motion in pediatric high-risk neuroblastoma patients using four-dimensional computed tomography. J. Appl. Clin. Med. Phys. 2017; 18(1):107–114.
17. Guerreiro F, Seravalli E, Janssens GO et al. Intra- and inter-fraction uncertainties during IGRT for Wilms’ tumor. Acta Oncol. (Madr). 2018; 57(7):941–949.
18. Rankine L, Wan H, Parikh P et al. Cone-Beam Computed Tomography Internal Motion Tracking Should Be Used to Validate 4-Dimensional Computed Tomography for Abdominal Radiation Therapy Patients. Int. J. Radiat. Oncol. 2016; 95(2):818–826.
19. Gierga DP, Chen GTY, Kung JH et al. Quantification of respiration-induced abdominal tumor
129
motion and its impact on IMRT dose distributions. Int. J. Radiat. Oncol. Biol. Phys. 2004; 58(5):1584–1595.
20. de Boer HCJ, Heijmen BJM. eNAL: An Extension of the NAL Setup Correction Protocol for Effective Use of Weekly Follow-up Measurements. Int. J. Radiat. Oncol. 2007; 67(5):1586–1595.
21. Rit S, van Herk M, Zijp L, Sonke J-J. Quantification of the Variability of Diaphragm Motion and Implications for Treatment Margin Construction. Int. J. Radiat. Oncol. 2012; 82(3):e399–e407.
22. Marchant TE, Amer AM, Moore CJ. Measurement of inter and intra fraction organ motion in radiotherapy using cone beam CT projection images. Phys. Med. Biol. 2008; 53(4):1087–1098.
23. Wink N, Panknin C, Solberg TD. Phase versus amplitude sorting of 4D-CT data. J. Appl. Clin. Med. Phys. 2006; 7(1):77–85.
24. Jin P, Hulshof MCCM, van Wieringen N et al. Interfractional variability of respiration-induced esophageal tumor motion quantified using fiducial markers and four-dimensional cone-beam computed tomography. Radiother. Oncol. 2017; 124(1):147–154.
25. Yang J, Cai J, Wang H et al. Is Diaphragm Motion a Good Surrogate for Liver Tumor Motion? Int. J. Radiat. Oncol. 2014; 90(4):952–958.
26. Xi M, Liu M-Z, Li Q-Q et al. Analysis of abdominal organ motion using four-dimensional CT. Chin. J. Cancer 2009; 28(9):989–993.
27. Siva S, Pham D, Gill S et al. An analysis of respiratory induced kidney motion on four-dimensional computed tomography and its implications for stereotactic kidney radiotherapy. Radiat. Oncol. 2013; 8(1):248.
28. Brody AS, Frush DP, Huda W, Brent RL. Radiation Risk to Children From Computed Tomography. Pediatrics 2007; 120(3):677–682.
29. Slovis TL. Children, computed tomography radiation dose, and the As Low As Reasonably Achievable (ALARA) concept. Pediatrics 2003; 112(4):971–972.
30. Miglioretti DL, Johnson E, Williams A et al. The Use of Computed Tomography in Pediatrics and the Associated Radiation Exposure and Estimated Cancer Risk. JAMA Pediatr. 2013; 167(8):700.
31. Vedam SS, Kini VR, Keall PJ et al. Quantifying the predictability of diaphragm motion during respiration with a noninvasive external marker. Med. Phys. 2003; 30(4):505–513.
32. Uh J, Krasin MJ, Li Y et al. Quantification of Pediatric Abdominal Organ Motion With a 4-Dimensional Magnetic Resonance Imaging Method. Int. J. Radiat. Oncol. Biol. Phys. 2017; 99(1):227–237.
33. Pai Panandiker AS, Winchell A, Rolen M et al. 4DMRI Provides More Accurate Renal Motion Estimation for IMRT in Young Children. Int. J. Radiat. Oncol. 2013; 87(2):S599.
34. Guckenberger M, Wilbert J, Meyer J et al. Is a Single Respiratory Correlated 4D-CT Study Sufficient for Evaluation of Breathing Motion? Int. J. Radiat. Oncol. 2007; 67(5):1352–1359.
130
Appendix 7
Supplementary Figure 7.1 | ACBCT values (black dots) plotted as function of days (day 0 is the day of 4DCT acquisition). Lines are linear fits to the ACBCT data; slopes (mm/day) are indicated in the legends next to the dotted line symbol. A4DCT values (open dots) were not included in the fit.
131
Chapter 8
General discussion
133
134
8.1 | Pediatric vs. adult radiotherapy
Treatment of pediatric cancer is a challenge for the multimodality treatment team due to the disease rarity, variety of subtypes, and complexity of treatment options that continuously undergo modifications and improvements. Moreover, radiotherapy techniques are evolving rapidly. Image-guided radiotherapy (IGRT), volumetric modulated arc therapy (VMAT) and intensity-modulated radiotherapy (IMRT), with either photons or protons, could lead to improved target conformality and reduce normal tissue dose. While these techniques could be potentially essential in children, research on implementation of these techniques is mainly focused on adult cancer types, and special emphasis on introducing these newer techniques in pediatric radiotherapy is limited. Therefore, achievements in adult radiotherapy are often pragmatically translated to a pediatric setting. An optimal treatment plan in radiotherapy aims to deliver a sufficiently high dose to kill the tumor while at the same time minimizing doses to surrounding healthy tissues, or organs at risk (OARs). The accuracy of the treatment plan and dose delivery is limited by geometrical uncertainties that need to be accounted for by safety margins. While these uncertainties are well studied in adults, pediatric data was lacking. In this thesis (chapter 2, 4, and 5), we have aimed to quantify these uncertainties and present a first estimation of more appropriate margin sizes in abdominal and thoracic areas in children. This thesis also shows that these uncertainties differ between children and adults (chapters 3 and 6), indicating that the size of the margins should potentially be different for children than for adults. This thesis underscored the need for child-specific treatment approaches, both for margin definitions in treatment planning, as well as child-specific imaging procedures and delivery techniques. In this last chapter, the results presented in this thesis are summarized and compared to findings from other pediatric studies. Furthermore, alternative options to account for organ motion are mentioned and future perspectives of pediatric radiotherapy will be discussed. 8.2 | Geometrical uncertainties and safety margins
Although research focusing on organ motion in children during radiotherapy is steadily growing, the use of relative small patient cohorts makes it difficult to derive population-based margins for children. For accurate margin definitions, an extensive understanding of all geometrical uncertainties is essential. The main sources of geometrical uncertainty are considered to be the patient set-up, interfractional position variation, intrafractional motion, and the delineation uncertainty.
135
Interfractional position variation
Daily imaging and patient set-up verification can minimize or eliminate the interfractional set-up error [1, 2]. Since bony anatomy is used as a surrogate of the target, a residual error (i.e., the interfractional position variation) of the tumor or OARs (relative to the bony anatomy) remains after set-up verification. In order to account for it, it needs to be predicted by population-based measurements. This type of pediatric data was lacking and predictions were generally based on clinical experience and available adult data. In chapter 3 we demonstrated significant differences in interfractional position variation of abdominal organs between children and adults. Results from chapter 2, 3 and 4 could, in combination with other pediatric studies focusing on interfractional position variation [1, 3], give an indication for the interfractional component of safety margins for children in abdominal and thoracic radiotherapy (Table 8.1). Since Nazmy et al. solely reported ranges of interfractional position variation [3], we were unable to estimate the systematic and random errors and therefore excluded this study in Table 8.1. We showed large variation between pediatric cancer patients. Although we hypothesized that organ motion might be related to age, height, or weight, we did not find clear correlations between interfractional position variation and such patient-specific factors. Neither did Guerreiro et al. [1], who found slightly smaller values in overall younger patients compared to our cohort. Therefore, we suggested a more individualized margin approach. With the introduction of pre-fraction cone beam computed tomography (CBCT) imaging, the treatment plan can be individualized by re-optimizing the plan based on patient-specific variations that occur over the course of treatment [4]. The success of this adaptive radiotherapy (ART) approach relies on the imaging quality and modality and will be further discussed in paragraph 8.5. On current CBCT images, the tumor is often not directly visible due to poor image quality, and bony anatomy is used as a surrogate for setup verification. In order to mitigate the interfractional position uncertainty, instead of using bony structures as a surrogate, target-based setup verification should be investigated. The use of fiducial markers in the target has been demonstrated to be feasible and accurate in adults with abdominal tumors [5, 6]. However, implementing markers is an invasive procedure, which is clearly unfavorable and not a standard procedure in pediatric radiotherapy. Nevertheless, the tumor is often surgically removed during pre-radiation treatment in children with abdominal tumors and surgical clips are placed at the tumor resection area. These clips, identifying the remaining tumor bed [1], or another organ in close proximity to the primary tumor [7], could be used as a surrogate for the target. In chapter 4, we showed that differences of systematic and random errors between abdominal organs were small, suggesting that for margin definitions, an organ closely located to the tumor could be used as a suitable surrogate. More importantly, image quality of pediatric CBCTs requires sufficient quality to enable rigid registration of surgical clips or abdominal organs. Future studies are needed to investigate the feasibility of using surgical clips or adjacent organs as a suitable surrogate instead of the bony structure and dosimetric consequences need to be compared.
136
Table 8.1| Summary of systematic (Σ) and random errors (σ) of interfractional position variation in the orthogonal directions for the right kidney, left kidney, liver, and spleen and in CC direction for the diaphragm from pediatric cancer patients from this thesis and Guerreiro et al. [1]. Right Kidney Left Kidney Liver Spleen Right
Diaphragm Left
Diaphragm LR CC AP LR CC AP LR CC AP LR CC AP CC CC
Abbreviations: LR = Left–Right; CC = Cranial–Caudal; AP = Anterior–Posterior
a the diaphragm was measured as one complete structure
137
Intrafractional motion Intrafractional motion in the abdomen and thorax is mainly caused by respiration and also needs to be accounted for when population-based safety margins are derived for treatment planning. In chapter 5, we found large variation in amplitude of respiratory-induced diaphragm motion between children (range 4.1-17.4 mm). Others found that intrafractional motion of the tumor bed and OARs was ≤ 5 mm [1, 8–10]. Since we only investigated the intrafractional motion of the diaphragm, which does not necessarily correlate with tumor or organ motion [9], using the diaphragm as a surrogate for abdominal and thoracic organ motion will certainly induce inaccuracies and uncertainties that need to be taken into account for treatment planning purposes. A summary of the currently available data on intrafractional organ motion in pediatric cancer patients is presented in Table 8.2. Since variation between patients was large (chapter 5 and 6), which was also found by others [1, 8–10], we suggested that a more individualized approach with pre-treatment 4DCT was necessary. A pre-treatment 4DCT provides patient-specific 3D data on the tumor position and OARs during several phases of the breathing cycle, which can then be incorporated in an individualized treatment plan [11]. However, in chapter 7, we found that respiratory-induced diaphragm motion in children determined on 4DCT does not accurately predict the daily respiratory-induced diaphragm motion observed on CBCTs. In addition, respiratory-induced diaphragm motion on CBCTs acquired within one treatment session with an interval of minutes already showed significant differences, meaning that the actual respiratory-induced motion during dose delivery can again differ from measurements on the CBCT, questioning the usefulness of ART. In chapter 5, we also showed that respiratory-induced diaphragm motion even varies within consecutive breathing cycles in children. To optimally account for intrafractional motion, real-time imaging might be the solution. Real-time imaging and other options to account for intrafractional motion will be further discussed in this chapter in paragraph 8.3 and 8.5.
138
Table 8.2| Summary of mean peak-to-peak amplitude (A) and standard deviation (random error σ) in mm in the orthogonal directions for the right kidney, left kidney, liver, and spleen and in CC direction for the diaphragm from this thesis and [1, 8-10]. Right Kidney Left Kidney Liver Spleen Right
Diaphragm Left
Diaphragm LR CC AP LR CC AP LR CC AP LR CC AP CC CC
Kannan et al. [9] 0.4 1.8 0.8 1.1 2.0 0.6 1.1 2.4 0.8 1.0 2.0 1.3 2.0 2.3 Jinsoo et al. [10] 2.1 2.7 3.0 3.9 2.6 Abbreviations: LR = Left–Right; CC = Cranial–Caudal; AP = Anterior–Posterior
a the diaphragm was measured as one complete structure
139
Delineation uncertainty Target delineation is performed by a radiation oncologist at the start of the radiotherapy process and the delineation uncertainty is another component of the margin recipe, albeit not investigated in this thesis. Inter-observer variability in delineation has been presented in numerous adult studies [12], but pediatric delineation studies are scarce, due to the large variety of tumor locations and site-specific protocols. A few studies, with a varying focus on medulloblastoma [13, 14], nephroblastoma [15], Hodgkin’s lymphoma [16–18], and rhabdomyosarcoma [17, 18] patients, demonstrated considerable inter-observer variability in delineation of the target volumes (conformity index ranging from 0.3 to 0.7). Differences were caused by ambiguities in the guidelines, interpretation differences of both guidelines and images, and varying levels of experience and individual performance. Inter-observer variability in delineation not only affected target volume delineation, but also the dose distribution in the surrounding healthy OARs [15, 17, 18]. These studies emphasize the need for highly specific protocol definitions and training to improve uniformity in target delineation and sparing of OARs. Improvements in diagnostic imaging (further discussed in paragraph 8.4) might also lead to higher uniformity in target definition. Nevertheless, although we did not study the delineation uncertainty in this thesis, it is clear that this component adds another significant contribution to the planning target volume (PTV) and planning risk volume (PRV) margins. It might be even a bigger contribution than the inter- and intrafractional uncertainties, as we know from adult studies [19–22]. However, current available pediatric literature only reported the conformity index, and from this metric no definitive statement regarding the delineation error (Σdelineation) can be made. Resulting safety margins and clinical application
The margin used greatly varies depending on availability of imaging and the planning strategy used in the clinic [23–25]. Current protocols for pediatric abdominal and thoracic tumors recommend an isotropic margin around the clinical tumor volume (CTV), defining the PTV, ranging from 10-20 mm [26–29], but no details are given on the origin of these margins. In general, they mention that institutional and individual experience should be a determining factor for choosing the appropriate margin size. According to van Herk’s CTV-to-PTV (2.5Σ+0.7σ) and McKenzies PRV (1.3Σ+0.5σ) recipes, these margins should be derived from population-based systematic (Σ) and random errors (σ) [30, 31]. However, small pediatric patient cohorts and large variation in childhood cancer types, subtypes, and treatment sites complicate this process. Next, the variety in age, height, and weight and the lack of correlation of organ position variation and respiratory motion with these patient-specific factors, make margins based on age or height not (yet) feasible. Nevertheless, we have summarized our results with the available literature on interfractional position variation and intrafractional motion (Table 8.1 and 8.2), which form the basis of a first estimation of population-based child-specific margins. The delineation and setup uncertainty had not been measured for our cohort. We used a set-up error from [2], and considered a delineation error of 3 or 5 mm, which was estimated based on values available from adult literature [19–21]. Systematic and random errors should be added quadratically, accounting for the geometric uncertainties by:
140
When a conventional free-breathing 3D planning CT is used without daily pre-fraction imaging, the required abdominal CTV-to-PTV margins in the left-right (LR) and anterior-posterior (AP) direction should be approximately 10-12 mm (PRV 6 mm) and 14-16 mm in the cranial-caudal (CC) direction (PRV 8-9 mm). If 3DCT imaging is followed by daily imaging and online set-up verification using bony anatomy, the interfractional setup uncertainty (Σsetup) is negligible [1, 2], resulting in slightly smaller margins; 9-10 mm in the LR and AP directions, and 12-15 mm in CC direction. In case 4DCT imaging is available, the systematic intrafractional component (Σintra) will be accounted for by an individualized internal target volume (ITV) margin, leading to slightly smaller ITV-to-PTV margins, but will result in overall larger PTV margins. If the delineation error would be considered 5 mm, it will greatly impact the margin size, adding approximately 4-5 mm up to the PTV and 2-3 mm to PRV margins. These estimated margins, as also recommended in chapter 2, 3 and 4, suggest that safety margins should be applied anistropically rather than isotropically. Additionally, current pediatric protocols do not report on specific PRV margins. Presented PRV margins could support the radiation oncologist in optimizing treatment plans with better sparing of dose to the OARs. In conclusion, the expected delineation error and the interfractional position variation seem to show the largest contribution to the size of the margin. Therefore, pediatric delineation studies are needed to accurately quantify the Σdelineation and the feasibility of target-based setup verification should be investigated. It should be noted that the currently available literature describes varying methodologies for the quantification of geometrical uncertainties, and consistency in reporting outcomes is lacking, varying in means, medians, ranges or confidence intervals. This makes it difficult to make a definitive statement on systematic and random errors. A pooled analysis of all (raw) available pediatric data would allow for suggestions based on a larger scale, or quantitative studies including more patients of different age, height, and weight are needed. The establishment of such an international database has large potential as a source of information and knowledge exchange, and would provide a solution to reach consensus, which is also suggested by several others [23, 32–35]. In children, treatments involving radiotherapy directed at abdominal and thoracic areas show good local control rates [7, 36–39]. This could imply that the used margins were sufficient for target coverage. However, radiation-induced adverse events later in life remain of great concern in children. Investigation of the dose-response relationship and the treatment-related risk factors for the occurrence and severity of these adverse events is ongoing research [40–46], providing fundamental information on which organs are more sensitive and likely to develop adverse events. Knowledge on organ position variation provides a lower bound on the achievable precision of patient selection, when (automating) selecting similar patients from a database of patients’ CT scans is applied for historical dose reconstruction [42, 43]. With this valuable information on organ dose-effect relationships in children, sparing of OARs in abdominal and thoracic radiotherapy will play an important role in future decision making.
141
8.3 | Other options to account for motion
Besides the application of a safety margin to account for organ motion, other approaches are available to either manage or reduce organ motion. Compared to radiotherapy in adults, different approaches and strategies are used in pediatric radiotherapy. Often, the young age of the patient plays a role. Compliance, cooperation and understanding of the technique can be challenging in (young) children, and make it difficult to implement the technique in a pediatric setting.
Immobilization
The importance of patient immobilization has increased as treatment delivery techniques have become more complex and precise. Especially for children, lying still is challenging when being separated from their parents when left alone in the treatment room [47–50]. Therefore, to ensure a stable and reproducible patient position, children are often fixated in a customized vacuum matrass [23]. Interestingly, most of the published studies report the use of a customized vacuum matrass [1–3, 8–10], also for adolescents, while in our department a vacuum matrass is rarely used. In our multicenter study in chapter 2, institution-based protocols differed in determining whether or not to use a vacuum matrass. However, interfractional position variation with respect to the bony anatomy, as quantified in chapter 2, is not affected by the immobilization system used and we therefore did not analyze patients treated with or without a vacuum matrass separately. It is therefore questionable if, when image-guided daily position verification is acquired, the use of a vacuum matrass would contribute to more accurate treatment delivery. The effect of being positioned in a vacuum matrass during treatment delivery on the intrafractional motion has not been studied, since the included patients in chapter 5, 6, and 7 were all treated at our department, and thus without a vacuum matrass.
General anesthesia
For some children, usually younger patients (< approximately 5 years [51, 52]), who cannot lie still during treatment, general anesthesia (GA) is used. In most studies, and similar to our anesthetic protocol, propofol was used as the drug of choice for sedation, meaning that no airway device was needed and patients maintained spontaneous breathing. There are no clear indications that GA has a consistent effect on the magnitude of organ motion [1, 8–10]. Similar to our results in chapter 3, Guerreiro et al. found that interfractional organ position variation in children treated under GA did not significantly differ from children who did not receive GA [1]. Regarding the effect of GA on respiratory-induced diaphragm motion, we found in chapter 5 that the variability of the amplitude during a fraction was significantly smaller in children treated under GA than in children of similar ages treated without GA. Kannan et al. also found very regular breathing in both amplitude and cycle time in children treated under GA, while those who did not receive GA were not as uniformly consistent [8]. In accordance to our findings in chapter 5, this did not translate to a difference in magnitude of the motion between the two groups. Guerreiro et al. also did not find a difference in intrafractional organ motion magnitude between both groups [1]. On the other hand, Uh et al. showed that for children with slow respiration rates (i.e., children receiving GA (< 8 years)), intrafractional organ motion tended to be larger [10]. Unlike Uh et al., Panandiker et al. found that intrafractional kidney motion was smaller for children treated under GA (< 8 years) than for children
142
treated without (> 8 years) [9]. Although GA will assure patient immobilization during treatment delivery, it seems that the influence of GA on respiratory-induced motion is inconsistent, assuming that breathing remains patient-specific. Since it also requires adequate consumption restrictions, increases health risks, is labor intensive, increases preparation and treatment time, and could have negative effects after treatment [52, 53], the use of GA on a daily basis for prolonged periods is not preferred and other options should be considered. Play preparation by a specialized radiation therapist can increase comfort and familiarity with radiation treatment in order to minimize or even eliminate the use of sedatives and GA [47, 49, 54]. Additionally, children can benefit from having audiovisual interventions as a distraction from the treatment delivery, and this can also reduce the need of daily anesthesia for radiotherapy in children [55, 56]. However, the direct impact of these video or musical interventions on interfractional position variation or respiratory motion has not been studied so far.
Breath holding
In the second part of this thesis we showed large variation in respiratory-induced diaphragm motion in children. Others found substantial motion of surrounding OARs in the three orthogonal directions [1, 8–10]. In adults, respiratory motion management systems have been introduced to control or reduce the respiratory-induced motion of organs and tumors in the upper-abdomen and thoracic area [57]. Breath-holding techniques are feasible and widely used in adult patients, achieving a more stable tumor position and thereby improving treatment delivery [58]. However, it is of concern that tumors do not stay completely still during breath holding. During inhalation breath-holding, the lung volume gradually decreases because oxygen is extracted from alveolar gas [59]. This leads to motion of the diaphragm [59–61], and possibly contributes to drifting of the tumor as well [62]. Nevertheless, inhalation breath-holding is widely used in adult radiotherapy. In particular for breast radiotherapy, in order to minimize high doses to the heart without compromising target coverage [63]. For a similar purpose in pediatric patients with Hodgkin’s disease with mediastinal involvement, Claude et al. showed that with an active breathing control (ABC) device children older than 13 years were capable of holding their breath for 15-20 s without any problem, stress or discomfort [64]. Lung and heart dose was decreased using ABC, while for other OARs no significant change was found. Since the procedure of ABC required two to three times more time for explanation and training, the benefit of dose reduction should be well balanced with increasing time and thereby costs [64]. Demoor et al. showed that children (age range 9-20 years) with a deep inspiration breath hold (DIBH) technique were able to hold their breath for 10-18 seconds [65]. However, although children treated with DIBH showed a significantly sparing of normal liver tissue compared to children treated under free breathing, no differences in toxicity between the two groups were found. To further establish the clinical long-term benefit, a prospective study is needed. In particular to know how much reduction of irradiated volume of the OARs, and more precisely in which parts of the OARs, leads to clinical benefits. Therefore, recently, Lundgaard et al. introduced a study protocol TEDDI (radiotherapy delivery in deep-inspiration) to assess the potential benefit of DIBH in children [66].
143
Other
Other options, which are technically more challenging techniques, including gating or tracking can also be employed to account for respiratory motion. While these techniques are successfully implemented in radiotherapy for some adult cancer types, implementing these techniques in a pediatric setting is more complicated. It is often a prerequisite for an ethical approval that the technique is proven beneficial in adults. Also, the benefits of the techniques should certainly outweigh the additional patient load, treatment time and costs, compared to current standards. Explanation and training of the techniques make that the full procedure requires more time. Prospective studies are needed to carefully evaluate the feasibility and potential clinical (and long-term) benefit of these techniques in children.
8.4 | Imaging modalities
Although 3DCT is the standard imaging technique for treatment planning purposes, there is a growing demand for better and more imaging. By incorporating complementary information from multimodality imaging (e.g., 4DCT, MRI, positron emission tomography (PET)), improved target delineation and more accurate treatment planning might be achieved. Especially, in the abdominal and thoracic area, where moving structures can cause motion artefacts and high soft tissue contrast is crucial, achievements in imaging techniques and thereby improving image quality are continuously investigated. However, concerning additional dose from CT imaging and the ALARA principle (keeping doses As Low As Reasonably Achievable), there is always an ongoing discussion and clinical implementations move slowly forward in pediatric radiotherapy.
4DCT
This thesis showed the need for a more individualized treatment approach in pediatric radiotherapy. A pre-treatment 4DCT is an effective tool to determine intrafractional motion from respiration. There are different strategies for using 4DCT in treatment planning [11]. Solely accounting for the internal motion leads to the ITV, which includes the CTV plus an internal margin, covering the entire respiratory-induced motion. However, to account for the remaining geometrical uncertainties an additional margin is added to the ITV (as earlier discussed), leading to large PTV margins and increased dose to healthy surrounding tissues. An alternative approach, the mid-ventilation based PTV planning, leads to smaller margins, simultaneously accounting for respiratory motion and the other geometrical uncertainties [11, 67, 68]. Previously, the use of 4DCT was only reported by Panandiker et al. [9, 69]. Recently, others also reported the use of 4DCT in children and described the use of ITV-to-PTV margins [1, 8, 69–71]. However, the optimal strategy how to use 4DCT in pediatric treatment planning is yet unknown and needs to be investigated. More importantly, although we suggested in chapter 6 that the 4DCT should be effective for treatment planning purposes in children, we found in chapter 7 that respiratory-induced motion as measured on 4DCT is not always representative for respiratory-induced motion during treatment. Using such a single measurement could possibly lead to insufficient target coverage. Therefore, our results suggest to monitor respiratory motion on a more regular basis, and adapt treatment plans when necessary, which will both be discussed further in this and the following paragraph.
144
(4D-) CBCT
To take full advantage of IGRT, acquisition of daily CBCT imaging is essential. More importantly, the process of ART relies on daily imaging. However, the additional imaging dose from CBCT imaging and increased risk for radiation induced adverse events is of great concern in children, with their sensitivity to radiation and long life expectancy. Efforts to reduce imaging dose have resulted in low dose CBCT protocols [72–74]. The design allows for considerable dose reduction, thereby avoiding unnecessary toxicity without compromising image quality for the registration accuracy of bony structures (Figure 8.1). However, poor soft tissue contrasts and lower image quality could hamper the registration of surgical clips or abdominal organs, which might be interesting for both research and clinical purposes. Lower doses also result in fewer 2D frames, which hamper the use of automatic methods to track intrafractional diaphragm motion as used in [75]. For potential online adaptive procedures including breathing motion, as suggested in chapter 7, sufficient image quality and number of projection images is crucial. Such a respiratory-correlated 4D-CBCT technique [76] requires additional imaging time and dose, but would allow for online verification of moving structures prior to treatment and adjustment of planning when needed. To overcome these limitations, different reconstruction methods are actively studied [77–79]. However, in chapter 7 we also showed that respiratory-induced diaphragm motion can vary within minutes and thus, motion as measured on a 4D-CBCT can again differ from actual motion during dose delivery.
Figure 8.1 | Example of optimizing imaging dose in pediatric CBCT imaging in the lumbar region. Exposure parameters used are 40ms, 32mA with a normal gantry rotation speed of 0.5 rpm (a) and 10ms, 10mA and a mimicked high gantry rotation speed (1.0 rpm) by reconstruction of the CBCT image with half of the projection frames (360 vs. 180 frames), resulting in imaging doses of approximately 2.0 cGy (left) and 0.2 cGy (right). Figure adapted from [72].
145
(4D-) MRI
With MR imaging, additional ionizing radiation doses can be avoided. Also, MR imaging provides superior soft tissue contrast and a higher resolution in the longitudinal plane compared to (CB)CT imaging. For delineation purposes, this additional imaging information might improve the identification of the tumor and its boundaries and reduce inter-observer variation [80, 81]. Also, 4D-MRI provides a non-ionizing option to measure and characterize respiratory-induced abdominal organ motion and has been successfully used in children [10, 82]. However, Panandiker et al. showed that the serial acquisition of simulation CT, 4DCT, and MRI required approximately 65 minutes for each child [69]. Additionally, incorporating supplementary MR imaging next to CT imaging, provides two imaging modalities for decision-making, which is at the same time more time-consuming and difficult, affecting treatment efficiency and clinical workload. Therefore, efforts have been made to eliminate CT imaging to create an MR-only workflow, generating synthetic CT-imaging from MR imaging [83], and has also been shown feasible for children for both photon and proton treatment planning [84–86]. Additionally, with the introduction of MR imaging integrated into the treatment machines [87, 88], online visualization of abdominal and thoracic motion during irradiation becomes possible. These MR-workflows might become increasingly important in future pediatric radiotherapy.
8.5 | Treatment delivery options
The development of modern treatment technologies, such as MR-guided radiotherapy and particle therapy, is at the same time putting high demands on experienced cooperating multimodality teams to create and deliver accurate and precise treatment plans.
IGRT
Novel delivery techniques, like VMAT and IMRT, are slowly integrated in pediatric radiotherapy and could improve target conformality and reduce normal tissue dose [89–93]. For Wilms’ tumor patients, the VMAT technique is associated with less radiation dose to the remaining kidney and better PTV coverage than conformal 3D radiotherapy [89, 90]. For abdominal neuroblastoma patients, the replacement of conformal radiotherapy by IMRT is controversial [91–94], since IMRT also slightly increases integral doses compared to less sophisticated techniques [93, 94] (Figure 8.1). One must balance if this increased dose is considered as an acceptable trade-off with regard to improved target coverage. Moreover, these delivery techniques demand the addition of pre-fraction imaging for precise patient positioning. Additional imaging dose is often not included in total dose calculations. Each CBCT increases the total dose by a finite percentage and causes scattered radiation to surrounding organs [95, 96]. Hess et al. proposed two pediatric IGRT decision trees, in which radiation oncologists are assisted to choose the most appropriate image modality in children and when CT is selected as the image modality, to assist in choosing optimal acquisition parameters [97]. For areas, with less motion expected, the individualized IGRT decision-making trees might help to define which patients would not necessarily benefit from daily IGRT, and will eliminate unnecessary dose to those patients. However, in the abdominal and thoracic areas, substantial interfractional position variation and respiratory motion is present and pre-fraction imaging is essential for higher accuracy.
146
ART
To anticipate on day-to-day anatomical and respiratory variations, ART might be the solution. Pre-fraction imaging enables to evaluate patient-specific variations and customize treatment plans according to the observed anatomical changes [4]. ART based on CBCT imaging can be performed offline or online. Most CBCT-based ART approaches are performed offline. A re-planning strategy requires a new CT to create a new treatment plan when the anatomical changes cause changes in the extent that the planning criteria are no longer met. For a plan selection strategy multiple treatment plans are created, where for each fraction, the plan that best fits the anatomy of that day is selected. An online ART workflow based on CBCT imaging is a more complex time-demanding process [98], requiring daily re-contouring and re-planning, while the patient is waiting. For a number of treatment sites in adults, ART has demonstrated to offer clinically relevant improvements [98–102]. The need for an adaptive approach in children with tumors in the head and neck region was not consistent [103–105]. Daily monitoring was suggested to assess which patients would benefit from an adaptive strategy. Guerreiro et al. investigated the dosimetric impact of daily anatomical changes in the abdominal region in children [106]. They found target under dosage and over dosage of the assessed kidneys, although differences were small. Re-planning should be considered when normal tissue tolerance of OARs is reached, however, due to poor soft tissue contracts of CBCT imaging assessment of other abdominal OARs was unfeasible. The effectiveness of an adaptive strategy depends on the quality of daily imaging. Present quality in pediatric CBCT imaging and the absence of real-time imaging cause restrictions and limit the possibility of ART in abdominal and thoracic areas in children. Daily imaging comes with the drawback of increased treatment times and additional doses, but improved CBCT reconstruction methods could minimize these limitations. Future studies are needed to investigate the feasibility of ART based on pediatric CBCT imaging and whether the increased clinical workload, due to evaluating pre-fraction imaging, will outweigh the potential gain in accuracy and dosimetric benefits.
MR-Guidance
The introduction of MR-guided treatment machines allow for daily and online visualization of the tumor and OARs, real-time verification of plan delivery, and accurate treatment adaptations. Recently, the first clinical adult patients were treated [107]. MR-guided approaches introduce superior imaging for ART compared to CBCT imaging, without any additional imaging dose. In children, MR-guided IMRT planning versus CBCT-guided VMAT planning showed promising results; smaller PTV margins and a reduction of normal tissue exposure [24]. The MR linac has the potential to provide a high-precision, adaptive, image-guided approach, with less toxicity in the surrounding healthy tissues, and will play an important role in the future of pediatric radiotherapy.
147
Figure 8.2 | Comparison of dose distributions using conformal radiotherapy (CRT), intensity modulated radiotherapy (IMRT), and proton beam radiotherapy (PBT) for a neuroblastoma patient. Area irradiated with 95% (red), 70% (yellow), 40% (green), and 10% (blue) of prescribed dose are shown superimposed on the CT image. Figure adapted from [82].
Particle therapy
Irradiation of cancer with heavy particles (e.g., protons or carbon ions) has an important benefit compared to conventional photons, even when advanced photon techniques are used (i.e., VMAT, IMRT) (Figure 8.2). The volume in which an ion beam deposits most of its dose is much sharper defined than with a photon beam, and for carbon ions even sharper than with a proton beam. This leads to reduction of the integral dose and better saving of the surrounding healthy tissues. Clinical data supporting the benefits of particle therapy are promising, however still scarce [32, 108, 109]. Recently, the Pediatric Proton Consortium Registry, a collaboration of 13 major pediatric cancer centers with proton therapy, was established to facilitate long-awaited clinical outcomes of children irradiated with proton therapy, but results are not yet available [35, 110]. The number of children treated with particle therapy is growing since proton facilities are rapidly expanding worldwide. Therefore, treatment with protons is expected to become more easily accessible and expertise will continue to grow. In the Netherlands, childhood cancer is a standard indication for treatment with protons. Two proton facilities have opened, of which one center expects to treat 80-100 children per year. Carbon ion treatment facilities are still limited, and mainly situated in Japan and Germany, also treating children [111, 112]. With its physical and biological benefits, treatment with carbon ions is expected to play an emerging role in radiotherapy future [113]. Several dosimetric studies have clearly demonstrated that particle therapy limit the irradiated volume compared to photon techniques. Most commonly, childhood cancer types in the central nervous system and head and neck, with static and solid tumors and vital tissues in close proximity, have clear indications for treatment with particles [32, 110, 114, 115]. Also, tumors in the abdomen and thorax show promising results [25, 93, 116–120]. However, as shown in this thesis, substantial organ position variation is present in the abdominal and thoracic area. These uncertainties, in combination with volumetric changes in gas content in the abdomen, can affect the particle beam range, causing excessive dose irregularities [121]. Robustly optimizing treatment plans can account for these range and setup uncertainties [122, 123], by including the expected anatomical variations
148
during treatment in the dose optimization. Respiratory motion can cause the interplay effect, and is more difficult to account for. However, Boria et al. showed that in children with mobile tumors receiving proton therapy, the interplay effect was largely negligible for younger children (2-6 years) with minimal CC target motion (< 5 mm) [85]. For mobile tumors treated with particle therapy, a 4D-based approach is paramount to quantify respiratory-induced motion and determine which strategy should be used [124]. As suggested by Boria et al. with respiratory motion > 10 mm, techniques to minimize the motion such as breath holding or respiratory gating should be considered [85, 125, 126]. Ultimately, to optimally benefit from particle therapy in those cases subject to tumor and organ motion, real-time imaging and the capability to adjust for anatomical changes is paramount. Pediatric patients could also benefit from ion radiography [112, 127], where the ion source is also used as imaging technique and the patient is imaged in the actual treatment position. Bony structures were clearly visible, allowing for accurate patient set-up, while reducing the imaging dose compared to X-ray imaging [127]. It could also be used during treatment delivery for online tumor tracking and dose adjustments, potentially leading to an online adaptive ion approach, but due to technical challenges it is still in an experimental phase. This also applies for achievements in combining MRI and particle therapy, which might ultimately provide the optimal treatment modality including an online non-ionizing imaging technique [128–131]. More realistic solutions are MR-guided photon therapy and image-guided proton therapy. They will become easier available in the Netherlands in the near future and are expected to play an important role in future pediatric radiotherapy. Until then, based on the available imaging techniques per institute (e.g., in-room CBCT, 4DCT), the optimal imaging and treatment approach will have to be chosen for each individual child, and the outcomes presented in this thesis will be applicable in treatment planning for defining appropriate margins.
149
References 1. Guerreiro F, Seravalli E, Janssens GO et al. Intra- and inter-fraction uncertainties during IGRT
for Wilms’ tumor. Acta Oncol. (Madr). 2018; 57(7):941–949. 2. Beltran C, Pai Panandiker AS, Krasin MJ, Merchant TE. Daily image-guided localization for
neuroblastoma. J. Appl. Clin. Med. Phys. 2010; 11(4):162–169. 3. Nazmy MS, Khafaga Y, Mousa A, Khalil E. Cone beam CT for organs motion evaluation in
pediatric abdominal neuroblastoma. Radiother. Oncol. 2012; 102(3):388–392. 4. Yan D, Vicini F, Wong J, Martinez A. Adaptive radiation therapy. Phys. Med. Biol. 1997;
42(1):123–132. 5. van der Horst A, Houweling AC, van Tienhoven G et al. Dosimetric effects of anatomical
changes during fractionated photon radiation therapy in pancreatic cancer patients. J. Appl. Clin. Med. Phys. 2017; 18(6):142–151.
6. Nederveen AJ, Dehnad H, van der Heide UA et al. Comparison of megavoltage position verification for prostate irradiation based on bony anatomy and implanted fiducials. Radiother. Oncol. 2003; 68(1):81–88.
7. Pai Panandiker AS, Beltran C, Billups CA et al. Intensity modulated radiation therapy provides excellent local control in high-risk abdominal neuroblastoma. Pediatr. Blood Cancer 2013; 60(5):761–765.
8. Kannan S, Teo BKK, Solberg T, Hill-Kayser C. Organ motion in pediatric high-risk neuroblastoma patients using four-dimensional computed tomography. J. Appl. Clin. Med. Phys. 2017; 18(1):107–114.
9. Pai Panandiker AS, Sharma S, Naik MH et al. Novel assessment of renal motion in children as measured via four-dimensional computed tomography. Int. J. Radiat. Oncol. Biol. Phys. 2012; 82(5):1771–1776.
10. Uh J, Krasin MJ, Li Y et al. Quantification of Pediatric Abdominal Organ Motion With a 4-Dimensional Magnetic Resonance Imaging Method. Int. J. Radiat. Oncol. Biol. Phys. 2017; 99(1):227–237.
11. Wolthaus JWH, Sonke JJ, van Herk M et al. Comparison of Different Strategies to Use Four-Dimensional Computed Tomography in Treatment Planning for Lung Cancer Patients. Int. J. Radiat. Oncol. Biol. Phys. 2008; 70(4):1229–1238.
12. Vinod SK, Jameson MG, Min M, Holloway LC. Uncertainties in volume delineation in radiation oncology: A systematic review and recommendations for future studies. Radiother. Oncol. 2016; 121(2):169–179.
13. X. Muracciole, J. Cuilliere, S. Hoffstetter, C. Alapetite, P. Quetin, M. Baro, Z. Gaci, J. Maire SC, Carrie C. Quality assurance of a french multicentric conformal radiatherapy protocol for low-stage medulloblastoma : Variability in Target Volume Delineation. Int. J. Radiat. Oncol. Biol. Phys. 2002; 54(2):S148-149.
14. Coles CE, Hoole ACF, Harden S V. et al. Quantitative assessment of inter-clinician variability of target volume delineation for medulloblastoma: quality assurance for the SIOP PNET 4 trial protocol. Radiother. Oncol. 2003; 69(2):189–194.
15. Padovani L, Huchet A, Claude L et al. Inter-clinician variability in making dosimetric decisions in pediatric treatment: A balance between efficacy and late effects. Radiother. Oncol. 2009; 93(2):372–376.
16. Lütgendorf-Caucig C, Fotina I, Gallop-Evans E et al. Multicenterevaluierung unterschiedlicher Zielvolumenkonzepte für pädiatrische Hodgkin-Erkrankungen: Eine Fallstudie. Strahlentherapie und Onkol. 2012; 188(11):1025–1030.
17. Kristensen I, Agrup M, Bergström P et al. Assessment of volume segmentation in radiotherapy of adolescents; A treatment planning study by the Swedish Workgroup for Paediatric Radiotherapy. Acta Oncol. (Madr). 2014; 53(1):126–133.
150
18. Kristensen I, Nilsson K, Agrup M et al. A dose based approach for evaluation of inter-observer variations in target delineation. Tech. Innov. Patient Support Radiat. Oncol. 2017; 3–4:41–47.
19. Versteijne E, Gurney-Champion OJ, van der Horst A et al. Considerable interobserver variation in delineation of pancreatic cancer on 3DCT and 4DCT: a multi-institutional study. Radiat. Oncol. 2017; 12(1):58.
20. van Mourik AM, Elkhuizen PHM, Minkema D et al. Multiinstitutional study on target volume delineation variation in breast radiotherapy in the presence of guidelines. Radiother. Oncol. 2010. doi:10.1016/j.radonc.2010.01.009.
21. Nijkamp J, De Haas-Kock DFM, Beukema JC et al. Target volume delineation variation in radiotherapy for early stage rectal cancer in the Netherlands. Radiother. Oncol. 2012. doi:10.1016/j.radonc.2011.08.011.
22. Deurloo KEI, Steenbakkers RJHM, Zijp LJ et al. Quantification of shape variation of prostate and seminal vesicles during external beam radiotherapy. Int. J. Radiat. Oncol. 2005; 61(1):228–238.
23. Alcorn SR, Chen MJ, Claude L et al. Practice patterns of photon and proton pediatric image guided radiation treatment: Results from an International Pediatric Research Consortium. Pract. Radiat. Oncol. 2014; 4(5):336–341.
24. Guerreiro F, Seravalli E, Janssens GO et al. Potential gain of MRI-guided IMRT planning versus current clinical CBCT-guided VMAT for Wilms’ tumor. Radiother. Oncol. 2018; 127(1):S884–S885.
25. Guerreiro F, Seravalli E, Janssens GO et al. Dosimetric comparison of VMAT and pencil beam scanning proton therapy for abdominal pediatric tumors. Radiother. Oncol. 2018; 127(1):S885-886.
26. De Kraker J, Graf N, Pritchard-Jones K. Nephroblastoma clinical trial and study SIOP 2001, Protocol. SIOP RTSG 2001.
27. SKION. DCOG NBL 2009 Treatment Protocol for Risk Adapted Treatment of Children with Neuroblastoma. 2015.
28. European paediatric Soft Tissue Sarcoma Study Group. Protocol for Localized Non-Rhabdomyosarcoma Soft Tissue Sarcoma. 2005.
29. EuroNet-Paediatric Hodgkin’s Lymphoma Group. First international Inter-Group Study for classical Hodgkin’s Lymphoma in Children and Adolescents. 2006.
30. van Herk M. Errors and margins in radiotherapy. Semin. Radiat. Oncol. 2004; 14(1):52–64. 31. McKenzie A, van Herk M, Mijnheer B. Margins for geometric uncertainty around organs at risk
in radiotherapy. Radiother. Oncol. 2002; 62(3):299–307. 32. Leroy R, Benahmed N, Hulstaert F et al. Proton Therapy in Children: A Systematic Review of
Clinical Effectiveness in 15 Pediatric Cancers. Int. J. Radiat. Oncol. Biol. Phys. 2016; 95(1):267–278.
33. Kortmann R-D, Freeman C, Marcus K et al. Paediatric radiation oncology in the care of childhood cancer: A position paper by the International Paediatric Radiation Oncology Society (PROS). Radiother. Oncol. 2016; 119(2):357–360.
34. Bhatia S, Armenian SH, Armstrong GT et al. Collaborative Research in Childhood Cancer Survivorship: The Current Landscape. J. Clin. Oncol. 2015; 33(27):3055–3064.
35. Kasper HB, Raeke L, Indelicato DJ et al. The Pediatric Proton Consortium Registry: A Multi-institutional Collaboration in U.S. Proton Centers. Int. J. Part. Ther. 2014; 1(2):323–333.
36. Bouzid N, Valteau Couanet D, Bolle S. Impact of radiation therapy on outcome in high-risk neuroblastoma. Radiother. Oncol. 2018; 127:S883.
37. Jo JH, Ahn SD, Choi EK et al. Patterns of Recurrence after Radiation Therapy for High-risk Neuroblastoma. Radiother. Oncol. 2018; 127:S883.
38. Skamene S, Abish S, Mitchell D, Freeman C. Radiotherapy is Important for Local Control at Primary and Metastatic Sites in Pediatric Rhabdomyosarcoma. Cureus 2015; 7(11):e388.
39. Gatcombe HG, Marcus RB, Katzenstein HM et al. Excellent Local Control From Radiation
151
Therapy for High-Risk Neuroblastoma. Int. J. Radiat. Oncol. 2009; 74(5):1549–1554. 40. Stovall M, Weathers R, Kasper C et al. Dose Reconstruction for Therapeutic and Diagnostic
Radiation Exposures: Use in Epidemiological Studies. Radiat. Res. 2006; 166(1):141–157. 41. Ng A, Brock KK, Sharpe MB et al. Individualized 3D Reconstruction of Normal Tissue Dose for
Patients With Long-term Follow-up: A Step Toward Understanding Dose Risk for Late Toxicity. Int. J. Radiat. Oncol. 2012; 84(4):e557–e563.
42. Wang Z, van Dijk IWEM, Wiersma J et al. Are age and gender suitable matching criteria in organ dose reconstruction using surrogate childhood cancer patients’ CT scans? Med. Phys. 2018; 45(6):2628–2638.
43. Virgolin M, van Dijk IWEM, Wiersma J et al. On the feasibility of automatically selecting similar patients in highly individualized radiotherapy dose reconstruction for historic data of pediatric cancer survivors. Med. Phys. 2018; 45(4):1504–1517.
44. Inkoom S, Raissaki M, Perisinakis K et al. Location of radiosensitive organs inside pediatric anthropomorphic phantoms: Data required for dosimetry. Phys. Medica 2015; 31(8):882–888.
45. Bölling T, Könemann S, Ernst I, Willich N. Late effects of thoracic irradiation in children. Strahlentherapie und Onkol. 2008; 184(6):289–295.
46. Bölling T, Willich N, Ernst I. Late effects of abdominal irradiation in children: A review of the literature. Anticancer Res. 2010; 30(1):227–232.
47. Scott MT, Todd KE, Oakley H et al. Reducing Anesthesia and Health Care Cost Through Utilization of Child Life Specialists in Pediatric Radiation Oncology. Int. J. Radiat. Oncol. Biol. Phys. 2016; 96(2):401–405.
48. Filin A, Treisman S, Peles Bortz A. Radiation Therapy Preparation by a Multidisciplinary Team for Childhood Cancer Patients Aged 3½ to 6 Years. J. Pediatr. Oncol. Nurs. 2009; 26(2):81–85.
49. Woodman H. Put yourself in their shoes the vulnerability of children and their families when attending for radiotherapy treatment: The role of the specialist paediatric radiotherapy radiographer. Radiography 2013; 19(4):311–314.
50. Engvall G, Ångström-Brännström C, Mullaney T et al. It is tough and tiring but it works- children’s experiences of undergoing radiotherapy. PLoS One 2016; 11(4):1–13.
51. Fortney JT, Halperin EC, Hertz CM, Schulman SR. Anesthesia for pediatric external beam radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 1999; 44(3):587–91.
52. Evans P, Chisholm D. Anaesthesia and paediatric oncology. Curr. Anaesth. Crit. Care 2008; 19(2):50–58.
53. Vigneron C, Schwartz É, Trojé C et al. Anesthésie générale en radiothérapie pédiatrique. Cancer/Radiothérapie 2013; 17(5–6):534–537.
54. Grissom S, Boles J, Bailey K et al. Play-based procedural preparation and support intervention for cranial radiation. Support. Care Cancer 2016; 24(6):2421–2427.
55. Willis D, Barry P. Audiovisual Interventions to Reduce the Use of General Anaesthesia with Paediatric Patients during Radiation Therapy. J. Med. Imaging Radiat. Oncol. 2010; 54(3):249–255.
56. O?Callaghan C, Sexton M, Wheeler G. Music therapy as a non-pharmacological anxiolytic for paediatric radiotherapy patients. Australas. Radiol. 2007; 51(2):159–162.
57. Keall PJ, Mageras GS, Balter JM et al. The management of respiratory motion in radiation oncology report of AAPM Task Group 76a). Med. Phys. 2006; 33(10):3874–3900.
58. Boda-Heggemann J, Knopf AC, Simeonova-Chergou A et al. Deep Inspiration Breath Hold - Based Radiation Therapy: A Clinical Review. Int. J. Radiat. Oncol. Biol. Phys. 2016; 94(3):478–492.
59. Parkes MJ. Breath-holding and its breakpoint. Exp. Physiol. 2006; 91(1):1–15. 60. Hong SK, Lin YC, Lally DA et al. Alveolar gas exchanges and cardiovascular functions during
breath holding with air. J. Appl. Physiol. 1971; 30(4):540–7. 61. Barnes E a, Murray BR, Robinson DM et al. Dosimetric evaluation of lung tumor
immobilization using breath hold at deep inspiration. Int. J. Radiat. Oncol. 2001; 50(4):1091–
152
1098. 62. Lens E, van der Horst A, Versteijne E et al. Considerable pancreatic tumor motion during
breath-holding. Acta Oncol. (Madr). 2016; 55(11):1360–1368. 63. Latty D, Stuart KE, Wang W, Ahern V. Review of deep inspiration breath-hold techniques for
the treatment of breast cancer. J. Med. Radiat. Sci. 2015; 62(1):74–81. 64. Claude L, Malet C, Pommier P et al. Active Breathing Control for Hodgkin’s Disease in
Childhood and Adolescence: Feasibility, Advantages, and Limits. Int. J. Radiat. Oncol. Biol. Phys. 2007; 67(5):1470–1475.
65. Demoor-Goldschmidt C, Chiavassa S, Josset S et al. Asservissement respiratoire lors d’une radiothérapie pulmonaire bilatérale pour le sarcome d’Ewing ou le néphroblastome chez des enfants et jeunes adultes : études dosimétrique et clinique de faisabilité. Cancer/Radiotherapie 2017; 21(2):124–129.
66. Lundgaard AY, Hjalgrim LL, Rechner LA et al. TEDDI: Radiotherapy delivery in deep inspiration for pediatric patients - A NOPHO feasibility study. Radiat. Oncol. 2018; 13(1):11–13.
67. Lens E, Van Der Horst A, Versteijne E et al. Dosimetric advantages of midventilation compared with internal target volume for radiation therapy of pancreatic cancer. Int. J. Radiat. Oncol. Biol. Phys. 2015; 92(3):675–682.
68. Wolthaus JWH, Schneider C, Sonke JJ et al. Mid-ventilation CT scan construction from four-dimensional respiration-correlated CT scans for radiotherapy planning of lung cancer patients. Int. J. Radiat. Oncol. Biol. Phys. 2006; 65(5):1560–1571.
69. Pai Panandiker AS, Beltran C, Gray J, Hua C. Methods for image guided and intensity modulated radiation therapy in high-risk abdominal neuroblastoma. Pract. Radiat. Oncol. 2013; 3(2):107–114.
70. Lavan N CE, Burland H, Ockwell C, Saran FH MH. Abstracts From the 48th Congress of the International Society of Paediatric Oncology (SIOP) Dublin, Ireland. Pediatr. Blood Cancer 2016; 63(Suppl 3):S5-321.
71. Lavan N, Smyth G, McQuaid D et al. Four-dimensional computed tomography-derived internal target volumes reduce volume of norml tissue in the high region in intensity modulated radiotherapy planning for high risk neuroblastoma. Pediatr. Blood Cancer 2016; 63:S91.
72. De Jong R, Lens E, Van Herk M et al. Optimizing cone-beam CT presets for children to reduce imaging dose illustrated with craniospinal axis. Radiother. Oncol. 2014; 111:S109–S110.
73. Atkinson B, Whitfield G, van Herk M. Design of an optimised bow-tie filter for cone-beam CT image guidance for children. Phys. Medica 2018; 52(1):11.
74. Alaei P, Spezi E. Imaging dose from cone beam computed tomography in radiation therapy. Phys. Medica 2015; 31(7):647–658.
75. Rit S, van Herk M, Zijp L, Sonke J-J. Quantification of the Variability of Diaphragm Motion and Implications for Treatment Margin Construction. Int. J. Radiat. Oncol. 2012; 82(3):e399–e407.
76. Sonke J-J, Zijp L, Remeijer P, van Herk M. Respiratory correlated cone beam CT. Med. Phys. 2005; 32(4):1176–1186.
79. Hansen DC, Sørensen TS. Fast 4D cone-beam CT from 60 s acquisitions. Phys. Imaging Radiat. Oncol. 2018; 5:69–75.
80. Gurney-Champion OJ, Versteijne E, van der Horst A et al. Addition of MRI for CT-based pancreatic tumor delineation: a feasibility study. Acta Oncol. (Madr). 2017; 56(7):923–930.
81. van de Schoot AJAJ, de Boer P, Buist MR et al. Quantification of delineation errors of the gross tumor volume on magnetic resonance imaging in uterine cervical cancer using pathology data and deformation correction. Acta Oncol. (Madr). 2015; 54(2):224–231.
153
82. Pai Panandiker AS, Winchell A, Rolen M et al. 4DMRI Provides More Accurate Renal Motion Estimation for IMRT in Young Children. Int. J. Radiat. Oncol. 2013; 87(2):S599.
83. Johnstone E, Wyatt JJ, Henry AM et al. Systematic Review of Synthetic Computed Tomography Generation Methodologies for Use in Magnetic Resonance Imaging–Only Radiation Therapy. Int. J. Radiat. Oncol. 2018; 100(1):199–217.
84. Guerreiro F. Abstracts from MR in RT 2018: Feasibility of MRI-only photon and proton dose calculations for abdominal pediatric patients. 2018.
85. Boria AJ, Uh J, Pirlepesov F et al. Interplay Effect of Target Motion and Pencil-Beam Scanning in Proton Therapy for Pediatric Patients. Int. J. Part. Ther. 2018; In press:IJPT-17-00030.1.
86. Uh J, Krasin MJ, Hua C. Technical Note: Feasibility of MRI-based estimation of water-equivalent path length to detect changes in proton range during treatment courses. Med. Phys. 2018; 45(4):1677–1683.
87. Lagendijk JJW, Raaymakers BW, van Vulpen M. The Magnetic Resonance Imaging–Linac System. Semin. Radiat. Oncol. 2014; 24(3):207–209.
88. Mutic S, Dempsey JF. The ViewRay System: Magnetic Resonance–Guided and Controlled Radiotherapy. Semin. Radiat. Oncol. 2014; 24(3):196–199.
89. Chen MJ, Leao CR, Carioca AL et al. Kidney-sparing whole abdominal irradiation in Wilms Tumor: potential advantages of VMAT technique. Radiother. Oncol. 2018; 217:S439–S440.
90. Davila-Fajardo R, Seravalli E, Ruebe C et al. Improved Target Coverage While Simultaneous-Integrated-Sparing of Kidney and Liver by VMAT in Children Undergoing Whole-Abdomen Irradiation for Wilms Tumour. Pediatr. Blood Cancer 2016; 63:S89.
91. Nazmy MS, Khafaga Y. Clinical experience in pediatric neuroblastoma intensity modulated radiotherapy. J. Egypt. Natl. Canc. Inst. 2012; 24(4):185–189.
92. Beneyton V, Niederst C, Vigneron C et al. Comparison of the dosimetries of 3-dimensions Radiotherapy (3D-RT) with linear accelerator and intensity modulated radiotherapy (IMRT) with helical tomotherapy in children irradiated for neuroblastoma. BMC Med. Phys. 2012; 12(1):2.
93. Fuji H, Schneider U, Ishida Y et al. Assessment of organ dose reduction and secondary cancer risk associated with the use of proton beam therapy and intensity modulated radiation therapy in treatment of neuroblastomas. Radiat. Oncol. 2013; 8(1):255.
94. Paulino AC, Ferenci MS, Chiang K-Y et al. Comparison of conventional to intensity modulated radiation therapy for abdominal neuroblastoma. Pediatr. Blood Cancer 2006; 46(7):739–744.
95. Ding GX, Coffey CW. Radiation Dose From Kilovoltage Cone Beam Computed Tomography in an Image-Guided Radiotherapy Procedure. Int. J. Radiat. Oncol. 2009; 73(2):610–617.
96. Deng J, Chen Z, Roberts KB, Nath R. Kilovoltage imaging doses in the radiotherapy of pediatric cancer patients. Int. J. Radiat. Oncol. Biol. Phys. 2012. doi:10.1016/j.ijrobp.2011.01.062.
97. Hess CB, Thompson HM, Benedict SH et al. Exposure Risks among Children Undergoing Radiation Therapy: Considerations in the Era of Image Guided Radiation Therapy. Int. J. Radiat. Oncol. Biol. Phys. 2016; 94(5):978–992.
98. Lim-Reinders S, Keller BM, Al-Ward S et al. Online Adaptive Radiation Therapy. Int. J. Radiat. Oncol. 2017; 99(4):994–1003.
99. Nijkamp J, Pos FJ, Nuver TT et al. Adaptive Radiotherapy for Prostate Cancer Using Kilovoltage Cone-Beam Computed Tomography: First Clinical Results. Int. J. Radiat. Oncol. Biol. Phys. 2008. doi:10.1016/j.ijrobp.2007.05.046.
100. Lutkenhaus LJ, Visser J, de Jong R et al. Evaluation of delivered dose for a clinical daily adaptive plan selection strategy for bladder cancer radiotherapy. Radiother. Oncol. 2015; 116(1):51–56.
101. Lutkenhaus LJ, de Jong R, Geijsen ED et al. Potential dosimetric benefit of an adaptive plan selection strategy for short-course radiotherapy in rectal cancer patients. Radiother. Oncol. 2016; 119(3):525–530.
102. Sonke J-J, Belderbos J. Adaptive Radiotherapy for Lung Cancer. Semin. Radiat. Oncol. 2010;
154
20(2):94–106. 103. Laskar S, Pandit P, Mallik S et al. Adaptive radiation therapy for pediatric head and neck
malignancies: Dosimetric implications. Pract. Radiat. Oncol. 2015; 5(2):e87–e94. 104. Beltran C, Sharma S, Merchant TE. Role of adaptive radiation therapy for pediatric patients
with diffuse pontine glioma. J. Appl. Clin. Med. Phys. 2011; 12(2):3421. 105. Beltran C, Naik M, Merchant TE. Dosimetric effect of target expansion and setup uncertainty
during radiation therapy in pediatric craniopharyngioma. Radiother. Oncol. 2010; 97(3):399–403.
106. Guerreiro F, Janssens GORJ, Seravalli E, Raaymakers BW. Abdominal Diameter Changes in Children During Volumetric Modulated Arc Therapy (VMAT): Is Re‐Planning Needed? J. Appl. Clin. Med. Phys., 2016; 43(6):3420–3420.
107. Raaymakers BW, Jürgenliemk-Schulz IM, Bol GH et al. First patients treated with a 1.5 T MRI-Linac: clinical proof of concept of a high-precision, high-field MRI guided radiotherapy treatment. Phys. Med. Biol. 2017; 62(23):L41–L50.
109. Mizumoto M, Murayama S, Akimoto T et al. Long-term follow-up after proton beam therapy for pediatric tumors: a Japanese national survey. Cancer Sci. 2017; 108(3):444–447.
110. Hess CB, Indelicato DJ, Paulino AC et al. An Update From the Pediatric Proton Consortium Registry. Front. Oncol. 2018; 8(May):1–10.
111. Combs SE, Kessel KA, Herfarth K et al. Treatment of pediatric patients and young adults with particle therapy at the Heidelberg Ion Therapy Center (HIT): establishment of workflow and initial clinical data. Radiat. Oncol. 2012; 7(1):170.
112. Combs SE, Ellerbrock M, Haberer T et al. Heidelberg Ion Therapy Center (HIT): Initial clinical experience in the first 80 patients. Acta Oncol. (Madr). 2010; 49(7):1132–1140.
113. Ebner DK, Kamada T. The Emerging Role of Carbon-Ion Radiotherapy. Front. Oncol. 2016. doi:10.3389/fonc.2016.00140.
114. Timmermann B. Proton Beam Therapy for Childhood Malignancies: Status Report. Klin. Pädiatrie 2010; 222(03):127–133.
115. Buchsbaum JC. Pediatric proton therapy in 2015: Indications , applications and considerations. Appl Rad Onc 2015.
116. Hattangadi JA, Rombi B, Yock TI et al. Proton Radiotherapy for High-Risk Pediatric Neuroblastoma: Early Outcomes and Dose Comparison. Int. J. Radiat. Oncol. 2012; 83(3):1015–1022.
117. Ladra MM, Edgington SK, Mahajan A et al. A dosimetric comparison of proton and intensity modulated radiation therapy in pediatric rhabdomyosarcoma patients enrolled on a prospective phase II proton study. Radiother. Oncol. 2014; 113(1):77–83.
118. Vogel J, Lin H, Both S et al. Pencil beam scanning proton therapy for treatment of the retroperitoneum after nephrectomy for Wilms tumor: A dosimetric comparison study. Pediatr. Blood Cancer 2017; 64(1):39–45.
119. Hoppe BS, Flampouri S, Zaiden R et al. Involved-node proton therapy in combined modality therapy for hodgkin lymphoma: Results of a phase 2 study. Int. J. Radiat. Oncol. Biol. Phys. 2014; 89(5):1053–1059.
120. Weber DC, Murray FR, Correia D et al. Pencil beam scanned protons for the treatment of patients with Ewing sarcoma. Pediatr. Blood Cancer 2017; 64(12):1–8.
121. Houweling AC, Fukata K, Kubota Y et al. The impact of interfractional anatomical changes on the accumulated dose in carbon ion therapy of pancreatic cancer patients. Radiother. Oncol. 2016; 119(2):319–325.
122. Liu W, Zhang X, Li Y, Mohan R. Robust optimization of intensity modulated proton therapy. Med. Phys. 2012; 39(2):1079–1091.
123. Unkelbach J, Paganetti H. Robust Proton Treatment Planning: Physical and Biological
155
Optimization. Semin. Radiat. Oncol. 2018; 28(2):88–96. 124. Chang JY, Zhang X, Knopf A et al. Consensus Guidelines for Implementing Pencil‐Beam
Scanning Proton Therapy for Thoracic Malignancies on Behalf of the PTCOG Thoracic and Lymphoma Subcommittee. Int. J. Radiat. Oncol. Biol. Phys. 2017; 99(1):41–50.
125. Tashiro M, Ishii T, Koya JI et al. Technical approach to individualized respiratory‐gated carbon‐ion therapy for mobile organs. Radiol. Phys. Technol. 2013; 6(2):356–366.
126. Minohara S, Kanai T, Endo M et al. Respiratory gated irradiation system for heavy‐ion radiotherapy. Int. J. Radiat. Oncol. 2000; 47(4):1097–1103.
127. Depauw N, Dias MF, Rosenfeld A, Seco J. Ion radiography as a tool for patient set‐up and image guided particle therapy: a Monte Carlo study. Technol. Cancer Res. Treat. 2014; 13(1):69–76.
128. Crijns SPM, Raaymakers BW, Lagendijk JJW. Real‐time correction of magnetic field inhomogeneity‐induced image distortions for MRI‐guided conventional and proton radiotherapy. Phys. Med. Biol. 2011; 56(1):289–297.
129. Hartman J, Kontaxis C, Bol GH et al. Dosimetric feasibility of intensity modulated proton therapy in a transverse magnetic field of 1.5 T. Phys. Med. Biol. 2015; 60(15):5955–5969.
130. Raaymakers BW, Raaijmakers AJE, Lagendijk JJW. Feasibility of MRI guided proton therapy: magnetic field dose effects. Phys. Med. Biol. 2008; 53(20):5615–22.
131. Moteabbed M, Schuemann J, Paganetti H. Dosimetric feasibility of real‐time MRI‐guided proton therapy. Med. Phys. 2014; 41(11):111713.
Summary
157
Continuous developments of effective multimodality treatment strategies for pediatric cancer have led to a steep increase in the number of childhood cancer survivors during the last decades. Inextricably, with the enhanced cancer survival, the incidence of treatment-related adverse events has become evident. Particularly, treatment with radiotherapy significantly contributes to the risk of developing adverse events. Radiotherapy uses ionizing radiation to kill the tumor cells. Unavoidably, radiation will pass through the patient’s body and healthy tissues will receive dose as well, resulting in acute and/or late toxicities and complications. Therefore, solely focusing on cure of cancer is not enough and reduction of adverse events deserves more attention. Although delivering adequate tumor dose to kill the tumor cells is the primary goal in radiotherapy, sparing the vital and long-term functions of adjacent organs at risk (OARs) is also paramount. Especially in children, who have a relative long life expectancy when surviving cancer, organs are still growing and have low tolerance to radiation. Therefore, extremely high precision in irradiation and avoiding collateral damage is pertinent. However, geometrical uncertainties are present and limit the accuracy. To account for these uncertainties, the clinical tumor volume (CTV) is extended with an isotopic safety margin, thereby defining the planning target volume (PTV). Similar margin definitions are also taken into consideration for OARs to define adequate planning risk volumes (PRVs). In adults, many studies focus on quantification of the geometrical uncertainties in order to define accurate population-based driven margins for specific tumor sites. Since childhood cancer is a rare disease, the small population makes it difficult to derive population-based margins for children. More importantly, childhood cancer patients are a highly diverse group, varying from infants to adolescents with varying heights and weights. Therefore, margins for children should be defined with more distinction. Moreover, achievements in radiotherapy are mainly focused on adult patients and pragmatically translated and implemented into a pediatric setting. Data on the geometrical uncertainties in pediatric radiotherapy is lacking, and clinically used margins are mainly based on experience of the radiotherapist or knowledge based on available adult data. Since the introduction of image guidance in pediatric radiotherapy, imaging data has become available to quantify the geometrical uncertainties in children, which can be described by interfractional position variation (i.e., day-to-day variations of the anatomy) or intrafractional motion (mainly caused by respiration). These geometrical uncertainties are a superposition of a systematic and a random error that form the basis for the population-based CTV-to-PTV or PRV margins. Systematic errors originate in the treatment preparation phase and therefore affect all treatment fractions. Random errors occur arbitrarily and could have a different effect each single fraction. Quantification and an extensive understanding of these uncertainties in children must lead to appropriate child-specific based margins, which is essential for high-accuracy image-guided radiotherapy in children. Children are treated with abdominal and thoracic radiotherapy for a wide range of primary cancer diagnosis. Especially in the abdominal and thoracic area, tumors and organs are prone to motion, and the tumor can be in very close proximity to radiosensitive OARs. In the first part of this thesis we focused on interfractional abdominal organ position variation. In chapter 2 we described our retrospective multicenter study in which we quantified interfractional position variation of the kidneys and the diaphragm in a cohort of 39 childhood cancer patients in order to estimate the standard deviation (SD) of systematic errors (Σ) and the root mean square of random errors (σ). We also analyzed the possible correlations between the interfractional position variation and patient-specific factors, such as age, height, and weight. Σ in left-right (LR), cranial-caudal (CC), and anterior-posterior (AP) directions was on average 1.2, 4.0, 1.8 mm and σ in these three directions was on average 1.2,
158
3.3, 1.9 mm, respectively. Our results showed that kidney position variation was the largest in the CC direction, suggesting that margins should be applied anisotropic ally rather than isotropically. Overall, interfractional position variation did not correlate with patient-specific factors and variation between patients was found to be large. This suggested that individualized margin approaches might be required. As margins for pediatric patients are usually pragmatically translated from adult data, comparison of our results with those reported on adults in literature showed that interfractional position variation was seemingly smaller in our pediatric cohort than in adults. However, caution in interpretation of this difference was warranted concerning differences in the methodology. Therefore, to yield a straightforward comparison we pooled pediatric and adult data of patients treated at our institute in chapter 3. All patients, including 35 children and 35 adults, were analyzed with identical methodology performed by the same observer as in chapter 2. Outcomes of both groups were compared, and patient-specific factors were observed as continuous values to investigate possible correlations between age, height, weight and organ position variation. Generally, Σ and σ values in the three main directions for the children were significantly smaller than those for the adults, especially in the CC direction (mean difference 3.1 mm). However, organ position variation and patient-specific factors were only negligibly correlated. Nevertheless, these results indicate that the size of the margins should potentially be different for children than for adults. If localization or quantification of tumor position variation is unfeasible, an anatomical structure close to the target could function as a surrogate. It is therefore crucial to have a clear understanding of the correlation of organ position variation between the tumor or organ and a particular surrogate. Therefore, to increase insight in abdominal and thoracic areas, in chapter 4, we quantified interfractional position variation of the diaphragm, liver, spleen and both kidneys in 20 children. We evaluated possible correlations between abdominal organs and determined whether organ position variation is location-dependent, and if this had possible consequences for margin values. No (strong) correlations between interfractional position variations of abdominal organs in children were observed. We concluded that diaphragm dome position variations could be more representative for superiorly located abdominal (liver, spleen) organ position variations than for inferiorly located (kidneys) organ position variations. Differences of Σ and σ between abdominal organs were small (<2 mm), suggesting that for margin definitions, there was insufficient evidence of a dependence of organ position variation on anatomical location. Besides interfractional position variation, intrafractional motion is an important component of the CTV-to-PTV margin and is the focus of the second part of this thesis. Especially, since in children, motion management techniques to minimize breathing motion, such as breath holding, are not widely used. Additionally, respiratory-induced organ motion is often not known prior to treatment in children and therefore needs to be accounted for in the CTV-to-PTV margin. In chapter 5 we described a retrospective study in which we analyzed respiratory-induced diaphragm motion in 45 children using daily or weekly acquired cone-beam CT (CBCT) scans. Variabilities within and between fractions were analyzed and possible correlations between respiratory-induced diaphragm motion and age, height, and weight, were investigated. Over all patients, interfractional variability was smaller than intrafractional variability. We found large ranges of respiratory-induced diaphragm amplitude motion (range 4.1-17.4 mm) and cycle time (2.1-3.9 s), indicating substantial differences between patients.
159
Moreover, no clinically significant correlations were found between amplitude and cycle time, and patient-specific factors, confirming that respiratory-induced diaphragm motion is patient-specific and requires an individualized approach to account for. Since interfractional variability was small, we suggested that a pre-treatment 4DCT in children could be sufficiently predictive to quantify respiratory-induced organ motion. However, only few studies on 4DCT in children are available in literature, implying that 4DCT is not widely used in pediatric radiotherapy. In adults, 4DCT is extensively used. However, day-to-day (interfractional) variability and irregular respiration (intrafractional variability) have shown to be limiting factors of 4DCT effectiveness. In order to evaluate the applicability of a 4DCT in children, in chapter 6, we retrospectively included 90 patients (45 children and 45 adults) and quantified inter- and intrafractional variability of respiratory motion. Using daily/weekly acquired CBCT imaging, identical methodology, as used in chapter 5, was used to extract respiratory-induced diaphragm motion and calculate the inter- and intrafractional variabilities. This pooled analysis enabled a solid comparison to reveal the effectiveness of 4DCT application for planning purposes in pediatric radiotherapy. The mean amplitude was slightly smaller in children than in adults (10.7 mm vs. 12.3 mm). Overall variability was smaller in children than in adults, indicating that respiratory motion is more regular in children than in adults. This implied that a single pre-treatment 4DCT could be a good representation of daily respiratory-induced organ motion in children and could be at least equally beneficial for planning purposes as it is in adults. However, studies on adult patients have indicated that respiratory motion, as measured on 4DCT, is not always representative for respiratory motion during the subsequent treatment course. Therefore, a single pre-treatment measurement for planning purposes might be a misrepresentation and could lead to under- or overestimating respiratory motion, yielding insufficient target coverage or undesired dose to OARs. Therefore, in chapter 7, we investigated the predictive value of 4DCT in 12 children. We analyzed whether respiratory-induced diaphragm motion in children on a single pre-treatment 4DCT can accurately predict respiratory-induced diaphragm motion as observed on daily CBCTs. We also analyzed possible time trends in respiratory-induced diaphragm motion over the complete treatment course. In 9 out of 12 patients, we observed a significant difference between the amplitude measured on 4DCT and the amplitude obtained from the daily/weekly CBCTs. The overall pre-treatment amplitude was smaller than the amplitude on CBCTs (10.4 mm vs. 11.6 mm). No obvious time trends in respiratory-induced diaphragm motion over the course of treatment were found, but significant baseline shifts of the diaphragm position were present in 7 out of 12 patients. These findings suggest that respiratory-induced diaphragm motion as measured on 4DCT was not representative for respiratory motion during the treatment course in children. Our results show the limitations of using a single pre-treatment 4DCT to take the patient-specific respiratory motion for treatment planning purposes into account. Therefore, regular monitoring of respiratory motion with CBCTs could yield a higher accuracy when a treatment plan is adapted to the actual breathing amplitude when necessary. Small pediatric patient cohorts and large variation in childhood cancer types, subtypes, and treatment sites make it difficult to derive population-based margins for children. Next, the variety in age, height, and weight and the lack of correlation of organ position variation and respiratory motion with these patient-specific factors, make margins based on age or height not (yet) feasible. Nevertheless, we have summarized our results with the available literature (chapter 8), which form the basis of a first
160
estimation of population-based child-specific margins. The large individual variations underscore the need for a more individualized approach. For newer treatment delivery techniques daily image guidance in children is essential, which might in the future lead to adaptive approaches in pediatric radiotherapy to account for anatomical variations that occur over the course of treatment. Besides the application of an accurate safety margin to account for organ motion, other approaches and techniques are available and evolving rapidly to either manage or reduce organ motion (chapter 8). Although research on implementation of these techniques is mainly focused on adult cancer types, they could be potentially beneficial in children as well. Future prospective studies are needed to assess the potential clinical (and long-term) benefits of these techniques in children. Moreover, for techniques like proton and carbon ion therapy, where, compared to photon therapy, the very sharp dose fall-off of protons and carbon ions are even less forgiving for anatomical variations during treatment, accounting for organ motion is paramount. It is clear that radiotherapy is nowadays a rapidly evolving field in cancer care. New techniques and approaches are continuously improving and also need to be integrated in the pediatric clinical setting. This will eventually lead to high-precision radiotherapy and minimizing dose to surrounding healthy tissues. This will contribute to the ambition for higher survival rates and maximizing quality of life in childhood cancer survivors.
161
Nederlandse samenvatting
163
In de afgelopen decennia is er grote vooruitgang geboekt in de multidisciplinaire behandeling van kinderkanker. Dit heeft ertoe geleid dat het aantal overlevenden van kinderkanker enorm is toegenomen. Deze kinderen krijgen later in hun leven vaak te maken met behandelingsgerelateerde late effecten, waarbij de behandeling met radiotherapie significant bijdraagt aan het risico op het ontwikkelen van deze late effecten. Radiotherapie maakt gebruik van ioniserende straling om de tumorcellen te doden. Onvermijdelijk zal het omliggende gezonde weefsel ook stralingsdosis ontvangen, en dit zal acute en/of late toxiciteit en complicaties veroorzaken. Hoewel een adequate dosis om de tumorcellen te doden het primaire doel is bij radiotherapie, is het sparen van de vitale en lange termijn functies van de aangrenzende gezonde organen ook van groot belang. Vooral bij kinderen, die een relatief lange levensverwachting hebben bij het overleven van kinderkanker, zijn de organen nog steeds in ontwikkeling en hebben ze een lagere tolerantie voor straling dan volwassenen. Daarom is uiterst hoge precisie nodig bij bestraling van de tumor. Er zijn echter geometrische onzekerheden aanwezig die de nauwkeurigheid van de behandeling beperken. Een grote onzekerheid die optreedt is de beweging van de tumor en de omliggende organen, met name in het abdominale en thoracale gedeelte. Om rekening te houden met deze onzekerheden, wordt het “clinical tumor volume” (CTV) uitgebreid met een isotropische veiligheidsmarge, waardoor het “plannings target volume” (PTV) ontstaat. Vergelijkbare marge-definities worden ook gebruikt om adequate ‘planning risk volumes” (PRVs) te definiëren voor de omliggende risico organen. Bij volwassenen zijn veel studies gericht op het kwantificeren van de geometrische onzekerheden om zo nauwkeurig mogelijk de marges voor specifieke tumors te definiëren. Kinderkanker komt echter relatief maar weinig voor, en er bestaan veel verschillende types tumoren op allerlei locaties in het lichaam. Door de kleine populaties per type tumor is het moeilijk om voor kinderen nauwkeurig marges te definiëren. Daarnaast zijn kinderkanker patiënten een zeer diverse groep, variërend van baby's tot tieners met verschillende lengtes en gewichten. Misschien kunnen de marges voor kinderen wel met meer onderscheid worden gedefinieerd. Echter, de ontwikkelingen in de radiotherapie zijn voornamelijk gericht op volwassen patiënten en meestal worden de behaalde resultaten pragmatisch vertaald naar een pediatrische setting. Gegevens over de geometrische onzekerheden bij radiotherapie in kinderen ontbraken en de marges die worden gebruikt zijn voornamelijk bepaald op grond van de ervaring van de radiotherapeut en/of op basis van de beschikbare gegevens van volwassenen. Sinds de introductie van beeldgestuurde radiotherapie bij kinderen zijn er inmiddels beelden beschikbaar gekomen waarbij het mogelijk is om de geometrische onzekerheden ook bij kinderen te kwantificeren. Deze kunnen worden beschreven als interfractionele positie variatie (de dag-tot-dag variaties van de anatomie van de patiënt) of intrafractionele beweging (voornamelijk veroorzaakt door ademhaling). Deze geometrische onzekerheden zijn een superpositie van een systematische (Σ) en een random fout (σ) die de basis vormen voor de CTV-PTV of PRV marges. Systematische fouten treden op tijdens de voorbereiding van de behandeling en zijn daarom van invloed op alle behandelingsfracties. Random fouten komen willekeurig voor en zouden elke afzonderlijke fractie een ander effect kunnen hebben. Het doel van dit proefschrift is het kwantificeren van de geometrische onzekerheden en het uitrekenen van de fouten bij bestralingen in kinderen in het abdomen en thoracale gedeelte. Dit moet ertoe leiden dat we nauwkeuriger kunnen bepalen wat de juiste marges zijn voor kinderen met tumoren en risico organen in het abdomen en thoracale gedeelte. Kinderen worden behandeld met abdominale en thoracale radiotherapie voor verschillende primaire kanker types. Vooral in het abdominale en thoracale gebied liggen tumoren en risico organen dicht bij elkaar en kunnen ze veel bewegen. In het eerste deel van dit proefschrift hebben we ons gericht op
164
interfractionele positie variatie. In hoofdstuk 2 beschrijven we onze retrospectieve multicenter studie waarin we de interfractionele positie variatie van de nieren en het diafragma hebben gekwantificeerd in een cohort van 39 kinderen met kinderkanker. We hebben de systematische (Σ) en random (σ) fouten bepaald en analyseerden ook de mogelijke correlaties tussen de interfractionele positie variatie en patiënt specifieke factoren, zoals leeftijd, lengte en gewicht. Σ in de links-rechts (LR), craniaal-caudaal (CC) en anterieur-posterieur (AP) richting was gemiddeld 1.2, 4.0, 1.8 mm en σ in deze drie richtingen was gemiddeld 1.2, 3.3, 1.9 mm, respectievelijk. Dit laat zien dat de positie variatie van de nier het grootste was in de CC-richting, wat suggereert dat marges eerder anisotropisch dan isotropisch moeten worden toegepast. Over het algemeen correleerde de interfractionele positie variatie niet met patiënt-specifieke factoren en de onderlinge variatie tussen patiënten bleek groot te zijn. Dit suggereerde dat geïndividualiseerde marges nodig zijn. Om de mogelijke verschillen tussen kinderen en volwassen in kaart te brengen, vergeleken we onze resultaten uit hoofdstuk 2 met die van volwassenen zoals bekend in de literatuur. We zagen dat de interfractionele positie variatie in ons pediatrische cohort kleiner was dan bij volwassenen. Echter, door de verschillen in methodologie en verschillende type tumoren was voorzichtigheid geboden bij de interpretatie van dit verschil. We hebben daarom voor een eenduidige vergelijking de gegevens van pediatrische en volwassen patiënten verzameld die allen in onze instelling waren behandeld (hoofdstuk 3). Alle patiënten, 35 kinderen en 35 volwassenen, werden met identieke methodologie geanalyseerd, uitgevoerd door dezelfde onderzoeker als in hoofdstuk 2. Resultaten van beide groepen werden vergeleken en patiënt-specifieke factoren werden geanalyseerd als continue factoren om mogelijke correlaties tussen interfractionele positie variatie en leeftijd, lengte, en gewicht te onderzoeken. Σ en σ waren in de drie richtingen voor de kinderen significant kleiner dan voor volwassenen, vooral in de CC-richting (gemiddeld verschil van 3.1 mm). Interfractionele positie variatie en patiënt specifieke factoren waren verwaarloosbaar gecorreleerd. Niettemin geven deze resultaten aan dat de grootte van de marges voor kinderen anders zou moeten zijn dan voor volwassenen. In het geval dat de tumor of een risico orgaan niet goed zichtbaar is tijdens beeldgestuurde radiotherapie, zou een andere anatomische structuur die dichtbij het doelgebied ligt kunnen functioneren als een surrogaat. Het is daarom van cruciaal belang om een duidelijk beeld te hebben van de correlatie tussen de positie variatie van de tumor of het orgaan en het desbetreffende surrogaat. Daarom hebben we, om het inzicht in beweging in het abdomen en thoracale gebied te vergroten, in hoofdstuk 4 de interfractionele positie variatie van het diafragma, de lever, de milt en beide nieren gekwantificeerd bij 20 kinderen. We evalueerden mogelijke correlaties tussen de organen en bepaalden of de positie variatie van de organen afhankelijk was van de locatie en of dit mogelijke gevolgen had voor de grootte van de marges. Er werden geen (sterke) correlaties waargenomen tussen interfractionele positie variaties van organen in het abdomen bij kinderen. We concludeerden dat de interfractionele positie variatie van het diafragma representatiever zou kunnen zijn voor positie variaties van superieur gelegen abdominale (lever, milt) organen dan voor positie variaties in de nieren. Verschillen tussen Σ en σ tussen de abdominale organen waren echter klein (<2 mm), wat suggereert dat er voor het definiëren van marges onvoldoende bewijs was voor een afhankelijkheid van anatomische locatie.
165
In het tweede deel van dit proefschrift focusten we op de intrafractionele beweging in het abdomen en thoracale gebied. In het algemeen worden bij kinderen technieken waarbij ademhalingsbewegingen geminimaliseerd worden, zoals het vasthouden van de adem, nog nauwelijks toegepast. Bovendien is ademhalingsgerelateerde beweging vaak niet bekend voorafgaand aan de behandeling bij kinderen en moet daarom worden meegenomen in de CTV-PTV marge. In hoofdstuk 5 beschrijven we een retrospectieve studie waarin we de ademhalingsbeweging van het diafragma analyseerden bij 45 kinderen met behulp van cone beam CT (CBCT) scans. Variaties binnen en tussen fracties werden geanalyseerd en mogelijke correlaties tussen de ademhalingsbeweging van het diafragma en leeftijd, lengte, en gewicht werden onderzocht. Over alle patiënten was de interfractionele variabiliteit kleiner dan de intrafractionele variabiliteit. We vonden grote verschillen tussen patiënten in de amplitude van de ademhalingsbeweging van het diafragma (variërend van 4.1-17.4 mm) en de ademhalingsperiode (2.1-3.9 s). Bovendien vonden we geen significante correlaties tussen ademhalingsamplitude en -periode en patiënt specifieke factoren. Dit bevestigt dat de ademhalingsbeweging van het diafragma patiënt specifiek is en een geïndividualiseerde benadering nodig is om ervoor te corrigeren. Omdat de interfractionele variabiliteit klein was, suggereerden we dat een 4DCT voorafgaand aan de behandeling bij kinderen voorspellend zou kunnen zijn om de ademhalingsbeweging van het diafragma te kwantificeren. In de literatuur zijn echter maar weinig studies over 4DCT bij kinderen beschikbaar, wat suggereert dat 4DCT nog nauwelijks wordt toegepast in pediatrische radiotherapie. Bij volwassen wordt 4DCT wel vaak gebruikt om de ademhalingsbeweging te meten voorafgaand aan de behandeling om dan mee te kunnen nemen in het behandelplan. Echter, de effectiviteit van een 4DCT wordt beperkt door de dagelijkse (interfractionele) variabiliteit en een onregelmatige ademhaling (intrafractionele variabiliteit). Om de toepasbaarheid van een 4DCT bij kinderen te evalueren, hebben we retrospectief van 90 patiënten (45 kinderen en 45 volwassenen) de inter- en intrafractionele variabiliteit van de ademhalingsbeweging van het diafragma gekwantificeerd (hoofdstuk 6). Daarbij maakten we gebruik van de dagelijkse/wekelijkse CBCT beelden en voerden we identieke methodologie, zoals ook gebruikt in hoofdstuk 5, uit. Deze gebundelde dataset maakte een robuuste vergelijking mogelijk waarbij we de effectiviteit van de 4DCT-toepassing voor planningsdoeleinden bij pediatrische radiotherapie onderzochten. De gemiddelde amplitude was iets kleiner bij kinderen dan bij volwassenen (10.7 mm vs. 12.3 mm). De algemene variabiliteit was kleiner bij kinderen dan bij volwassenen, wat aangeeft dat de ademhaling bij kinderen regelmatiger is dan bij volwassenen. Dit impliceerde dat een enkele 4DCT voorafgaand aan de behandeling, de dagelijkse ademhalingsbeweging bij kinderen goed zou kunnen voorspellen en minstens net zo effectief zou zijn voor planningsdoeleinden als bij volwassenen. Studies bij volwassen patiënten hebben echter aangetoond dat ademhalingsbewegingen, gemeten met een 4DCT, niet altijd representatief zijn voor de ademhalingsbeweging tijdens de daaropvolgende behandeling. Daarom kan een enkele meting een onjuiste voorspelling zijn en leiden tot onder- of overschatting van de ademhalingsbeweging, wat weer kan leiden tot onvoldoende dosis in het doelgebied en/of ongewenste dosis in risico organen. Daarom hebben we in hoofdstuk 7 de voorspellende waarde van de 4DCT bij 12 kinderen onderzocht. We analyseerden of de ademhalingsbeweging van het diafragma bij kinderen, gemeten op een enkele 4DCT voor de behandeling, representatief is voor de ademhalingsbeweging van het diafragma, zoals waargenomen op dagelijkse CBCTs. We analyseerden ook mogelijke tijdtrends over het gehele verloop van de
166
behandeling. Bij 9 van de 12 patiënten, zagen we een significant verschil tussen de amplitude gemeten in de 4DCT en de amplitude verkregen uit de dagelijkse/wekelijkse CBCTs. De amplitude van de 4DCT was kleiner dan de amplitude op CBCTs (10.4 mm vs. 11.6 mm). Er werden tijdens de behandeling geen duidelijke trends in de tijd gevonden, maar bij 7 van de 12 patiënten waren er significante verschuivingen van het diafragma ten opzichte van de diafragmapositie tijdens het plannen, ook na de setup verificatie. Deze bevindingen suggereren dat de ademhalingsbewegingen van het diafragma zoals gemeten op de 4DCT niet representatief zijn voor ademhalingsbewegingen tijdens de behandeling. Onze resultaten demonstreren de beperkingen van het gebruik van een enkele 4DCT voor de behandeling om te corrigeren voor patiënt-specifieke ademhalingsbewegingen in het behandelplan. Daarom zou regelmatige monitoring van ademhalingsbeweging met bijvoorbeeld dagelijkse CBCTs een hogere nauwkeurigheid kunnen bieden wanneer een behandelplan wordt aangepast aan de werkelijke ademhalingsamplitude. Ondanks dat de kleine aantallen in pediatrische radiotherapie het moeilijk maken om op populatie gebaseerde marges voor kinderen te bepalen, presenteren we in dit proefschrift een eerste resultaat om nauwkeurige marges voor abdominale en thoracale locaties in kinderen te definiëren. We hebben aangetoond dat zowel interfractionele positie variatie als intrafractionele beweging in het abdomen en thoracale gebied verschilt tussen kinderen en volwassenen. Dit geeft aan dat de marges voor kinderen anders zouden moeten zijn dan voor volwassenen. Grote individuele variaties en zwakke correlaties met leeftijd, lengte en gewicht, benadrukken echter de noodzaak voor een meer geïndividualiseerde benadering voor het bepalen van de marges. Dit proefschrift toont ook aan dat de inter- en intrafractionele variabiliteit ertoe leidt dat beeldgestuurde radiotherapie voor abdominale en thoracale bestralingen noodzakelijk is. Daarmee kan mogelijk in de toekomst ook voor kinderen een adaptieve benadering worden toegepast, waarbij het behandelplan ten gevolge van dagelijkse anatomische veranderingen kan worden herzien en worden geoptimaliseerd. Naast het toepassen van nauwkeurige marges om te corrigeren voor de beweging van de tumor en omliggende organen, zijn er ook andere benaderingen en technieken beschikbaar en in ontwikkeling om beweging te controleren of te verminderen. Hoewel onderzoek naar de implementatie van deze technieken vooral gericht is op volwassenen, kunnen ze mogelijk ook gunstig zijn voor kinderen. Prospectieve studies zijn nodig om de mogelijke klinische (en lange termijn) voordelen van deze technieken bij kinderen te beoordelen. Bestraling met protonen en ionen is een andere ontwikkeling die steeds vaker wordt toegepast. In vergelijking met fotonen, zijn protonen en ionen door de zeer scherpe dosisafname zelfs nog minder vergevingsgezind voor anatomische variaties en beweging tijdens de behandeling en daarom is beeldgestuurde therapie van nog groter belang. Radiotherapie technieken ontwikkelen zich voortdurend, waarbij de nieuwe technieken en benaderingen ook geïntegreerd (moeten) worden in de pediatrische klinische setting. Uiteindelijk zal dit ertoe moeten leiden dat we een zeer nauwkeurige dosis afgifte in het doelgebied krijgen en een zo laag mogelijke blootstelling aan straling in omliggend, gezond weefsel. Dit zal dan bijdragen aan de ambitie voor nog hogere overlevingspercentages en het voorkomen en beperken van late effecten in kinderen die kanker overleven.
167
Addendum List of publications
PhD portfolio
Curriculum vitae
Dankwoord
169
170
List of publications Peer-reviewed (this thesis)
Quantification of renal and diaphragmatic interfractional motion in pediatric image-guided radiation therapy: a multicenter study Sophie C. Huijskens, Irma W.E.M. van Dijk, Rianne de Jong, Jorrit Visser, Raquel Dávila Fajardo, Cécile M. Ronckers, Geert O.R.J. Janssens, John H. Maduro, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel Radiotherapy and Oncology 2015; Volume 117 (3): 425-431 Interfractional renal and diaphragmatic position variation during radiotherapy in children and adults: is there a difference? Irma W. E. M. van Dijk, Sophie C. Huijskens, Rianne de Jong, Jorrit Visser, Raquel Dávila Fajardo, Coen R. N. Rasch, Tanja Alderliesten, Arjan Bel Acta Oncologica 2017; Volume 56(8): 1065-1071 Abdominal organ position variation in children during image-guided radiotherapy Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Dirk te Lindert, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel Radiation Oncology 2018;13:173 Magnitude and variability of respiratory-induced diaphragm motion in children during image-guided radiotherapy Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel Radiotherapy and Oncology 2017; Volume 123 (2): 263-269 The effectiveness of 4DCT in children and adults: a pooled analysis Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel Journal of Applied Clinical Medical Physics 2019; Volume 20(1): 276-283 Predictive value of pediatric respiratory-induced diaphragm motion quantified using pre-treatment 4DCT and CBCTs Sophie C. Huijskens, Irma W.E.M. van Dijk, Jorrit Visser, Brian V. Balgobind, Coen R.N. Rasch, Tanja Alderliesten, Arjan Bel Radiation Oncology 2018;13: 198
Other publications Heart volume reduction during radiotherapy involving the thoracic region in children: An unexplained phenomenon Irma W.E.M. van Dijk, Jorrit Visser, Jan Wiersma, Jessica R. van Boggelen, Brain V. Balgobind, Elizabeth A.M. Feijen, Sophie C. Huijskens, Wouter E.M. Kok, Leontien C.M. Kremer, Coen R.N. Rasch, Arjan Bel Radiotherapy and Oncology 2018; Volume 128 (2): 214-220
171
PhD Portfolio
Name PhD student: Sophie Huijskens
Institute: Amsterdam UMC, location AMC
Department: AMC Graduate School
PhD period: October 2014 – September 2018
Promotor: Prof. Dr. C.R.N. Rasch
Copromotores: Dr. A. Bel
Dr. T. Alderliesten
1ECTS = 28 hours
PhD training Year Workload (ECTS)
General courses
AMC World of Science
Scientific Writing for publication (English)
Project Management
Computing in R
Citation Analysis and Impact Factors
Entrepreneurship in Health and Life Sciences
2014
2015
2015
2016
2017
2018
0.7
1.5
0.6
0.4
0.3
0.6
Specific courses
Pediatric radiation oncology course (ESTRO)
Winter School on Quantitative Analysis of Medical Images
2015
2016
1.0
1.0
Seminars, workshops and master classes Radiation oncology departmental meetings Physics staff meetings Multidisciplinary staff meetings Research meetings Regional radiotherapy lectures (AMC, VUmc, NKI-AvL) Scientific RKF project days (AMC, UMCU, EUMC) Masterclass Amsterdam Kinder Symposium
2014-2018 2014-2018 2014-2018 2016
172
International conferences
3rd ESTRO Forum, Barcelona, Spain 48th SIOP, Dublin, Ireland ESTRO 36, Vienna, Austria PROS 2017, New York, USA 50th SIOP, Kyoto, Japan
2015 2016 2017 2017 2018
Symposia & meetings Late effecten na Kankerbehandeling, NVvO Milestonedag Annual graduate student retreat (OOA) Amsterdam Kinder Symposium (AKS) 7th NCS Lustrum Symposium Institute QuantiVision Conference (iQC) Symposium ART
2014 2014 2015-2017 2017 2017 2018
Presentations
The quantification of renal and diaphragmatic interfraction motion in children - poster presentation, OOA retreat, Rennesse
2014 0.2
The quantification of renal and diaphragmatic interfraction motion in children - poster presentation, 3rd ESTRO forum, Barcelona, Spain
2015 0.2
Radiotherapy in children, is there a difference with adults? - oral presentation, AKS 2016, Amsterdam
2016 0.3
Respiration-induced organ motion in children during image-guided radiation therapy - poster presentation, ESTRO 35, Turin, Italy
2016 0.2
Magnitude of respiration-induced diaphragm motion during radiotherapy: Is it smaller in children than in adults? – oral presentation, 48th SIOP, Dublin, Ireland
2016 0.4
Respiratory-induced diaphragm motion during radiotherapy – oral presentation, AKS 2017, Amsterdam
2017 0.4
Comparison of respiratory-induced diaphragm motion during radiotherapy between children and adults – poster presentation, iQC 2017, Amsterdam
2017 0.2
173
Comparison of respiratory-induced diaphragm motion during radiotherapy between children and adults – poster presentation, ESTRO 36, Vienna, Austria
2017 0.3
Is interfractional organ position variation during image-guided radiotherapy in children different for organs located in the left side versus the right side of the abdomen? – poster presentation, PROS 2017, New York, USA
2017 0.3
Respiratory-induced abdominal motion during image-guided radiotherapy in children can be reliably estimated from a single pre-treatment 4D-CT – poster presentation, PROS 2017, New York, USA
2017 0.3
Can respiratory-induced diaphragm motion in children be reliably estimated from a single pre-treatment 4D-CT? – oral presentation, AKS 2018, Amsterdam
2018 0.2
Can respiratory-induced diaphragm motion in children be reliably estimated from a single 4D-CT? – poster presentation, ESTRO 37, Barcelona, Spain
2018 0.3
Is respiration of children constant during a radiotherapy session? – poster presentation, 50th SIOP, Kyoto, Japan
2018 0.2
174
Supervising and teaching K. van Oort, master student technical medicine at Twente University, Enschede M. van Erp, master student technical medicine at Twente University, Enschede S. Boonstra, master student technical medicine at Twente University, Enschede N. Morsink, master student technical medicine at Twente University, Enschede C. Jeltes, master student technical medicine at Twente University, Enschede
2015 2016 2016 2017 2017
3 3 3 3 3
Parameters of esteem
Grants
Ter Meulen Grant, The Royal Netherlands Academy of Arts and Sciences
2017
Awards and Prizes
2nd Best poster award, category Translational, OOA retreat
Best speaker award, AKS 2017
2nd Best poster award, iQC 2017
2014
2017
2017
175
Curriculum Vitae Sophie Huijskens was born on July 12, 1988 in Leidschendam (The Netherlands) and was raised in Voorschoten. After graduating from the Stedelijk Gymnasium Leiden in 2006, she started the bachelor program Life, Science and Technology at the University of Groningen, with a Major in Biomedical Engineering. In 2011, she started the master program Biomedical Engineering, with a specialization in Clinical Physics, and a close collaboration with the University Medical Center Groningen. As a part of her master education, she carried out her graduation research project at the MRI Research Centre at the University of British Columbia in Vancouver (Canada). She performed her master’s internship project at the department of Radiation Oncology at the Amsterdam UMC, location AMC. Sophie also followed the Master’s Honours program at the University of Groningen. After obtaining her master’s degree, she started in October 2014 her PhD research at the department of Radiation Oncology at the Amsterdam UMC, location AMC. This thesis is the result of the four years of research during her PhD project. Recently, she had the opportunity to collaborate with the Gunma University Heavy Ion Medical Center in Japan, which was funded by the Ter Meulen Grant from The Royal Netherlands Academy of Arts and Sciences. Currently Sophie is continuing her research as a postdoctoral researcher at the Amsterdam UMC, location AMC.
176
Dankwoord Met heel veel plezier heb ik 4 jaar aan dit proefschrift gewerkt. Op deze plek wil ik graag iedereen bedanken die op zijn of haar manier hieraan heeft bijgedragen! Coen, bedankt voor je fijne, directe en kritische begeleiding en het vertrouwen in mij om dit project (binnen 4 jaar) tot een succesvol einde te brengen! Het is jammer dat we jou als verbindende factor tussen de fysica en kliniek nu moeten missen, maar ik hoop dat we elkaar in deze deeltjeswereld nog eens tegenkomen. Arjan en Tanja, in 2014 kwam ik aankloppen bij jullie voor een stage plek. Al snel voelde ik me op mijn plek hier en ontstond deze succesvolle en fijne samenwerking. Heel veel dank voor jullie inspiratie en motivatie en ook het vertrouwen in mij om het KiKa (en daarna KWF) project uit te voeren. Het voelde als een vliegende start! Tanja, ik heb veel mogen leren van jouw kennis en toewijding als onderzoeker. Altijd zeer kritisch en in voor nog een revisie ronde. Jouw correcties op de manuscripten waren altijd scherp en gedetailleerd. Bijzonder leerzaam en waardevol voor de kwaliteit van dit proefschrift. Arjan, ons donderdag-eindevande-middag overleg was een onmisbaar deel van mijn promotie. Het bracht altijd nieuwe ideeën, kostbare leermomenten, nieuwe motivatie en energie. We opperde vaak het overleg naar een serieuzer tijdstip te verplaatsen maar ik ben blij dat we dat nooit hebben doorgezet. Op 1 tissue na, viel er ook altijd wel wat te vieren. Bijna altijd waren we het eens, met als resultaat dit mooie boekje! Ik hoop dat ‘ie een mooi plekje op de plank krijgt! Overige leden uit deze promotiecommissie, ik wil jullie bedanken voor het lezen en beoordelen van mijn manuscript en het zitting nemen in de oppositie. Ik kijk ernaar uit om met jullie van gedachten te wisselen! Co-auteurs en TG studenten, dank voor jullie inzet die de studies tot een succes hebben gemaakt! Alle collega’s van de afdeling, oud en nieuw, en in het speciaal alle mede-phders! Het voelde vanaf het eerste moment dat ik als stagiair begon meteen goed. Zo’n diverse afdeling met technische, fysische en klinische aspecten en toch een enorme samenhorigheid. Iedereen altijd bereid om mee te denken en te helpen! En natuurlijk op talloze borrels, afdelingsuitjes, congressen en natuurlijk de beroemde AMC sportpoules altijd genoeg enthousiaste collega’s! Door deze fijne sfeer vlogen de jaren voorbij en ik ben blij nog steeds bij dit team te horen! Zonder wie ik dit traject zeker niet had kunnen volbrengen is natuurlijk Irma. Jouw promotie was mijn startsein, dat wilde ik ook! Er is geen idee, abstract, poster, figuur, tabel of manuscript wat jij niet hebt gezien en altijd van waardevol en nuttig feedback voorzag. Ook liet je duidelijk zien als je trots op me was, dat gaf veel vertrouwen en het was een feest om onder jouw hoede zo te groeien als onderzoeker. Maar niet alleen inhoudelijk nam je altijd de tijd, ook op alle andere vlakken stond je altijd voor me klaar. En natuurlijk alle leuke uitjes en congressen, van het AKS tot New York en mij wegbrengen naar Japan! Ik bewonder hoe dierbaar jij al je contacten onderhoudt. Daarom weet ik zeker dat we elkaar altijd zullen blijven zien en spreken!
177
Jetta, dank voor de prachtige cover! Lieve paranimfen, met heel veel koffietjes en een pepernootje, alle PhD perikelen kunnen delen samen en vandaag het laatste loodje! Lieve vrienden, lieve vriendinnen, en liefste familie, ik geloof niet dat een dankwoord in dit proefschrift nodig is om duidelijk te maken dat ik alle uren en dagen dat ik niet aan dit proefschrift heb gewerkt, (en dat waren er gelukkig meer dan genoeg!!) het liefste met jullie om mij heen was.. want dat was ik! Arigato gozaimasu!
