Overview Imaging for Target Volume Delineation in Rectal Cancer Radiotherapy d A Systematic Review S. Gwynne * , S. Mukherjee y, R. Webster * , E. Spezi z, J. Staffurth x, B. Coles jj, R. Adams x * Department of Clinical Oncology, Velindre Cancer Centre, Cardiff, UK y Gray Institute for Radiation Oncology and Biology, Department of Oncology, Old Road Campus Research Building, University of Oxford, UK z Department of Physics, Velindre Cancer Centre, Cardiff, UK x Division of Cancer, School of Medicine, Cardiff University, Cardiff, UK jj Cancer Research Wales Library, Velindre Cancer Centre, Cardiff, UK Received 28 February 2011; received in revised form 15 September 2011; accepted 19 September 2011 Abstract The global move towards more conformal radiotherapy for rectal cancer requires better imaging modalities that both visualise the disease accurately and are reproducible; to reduce interobserver variation. This review explores the advances in imaging modalities used in target volume delineation, with a view to make recommendations for current clinical practice and to propose future directions for research. A systematic review was conducted using MEDLINE and EMBASE. Articles considered relevant by the authors were included. Planning with orthogonal films is being replaced by computed tomography (CT) simulation. This is now considered the ‘gold standard’ and allows conformal three-dimensional planning. Magnetic resonance imaging (MRI) has been shown to overcome some of the limitations of CT and can be used either as a diagnostic image to visually aid planning, or as a ‘planning’ MRI carried out in the treatment position and co-registered with the planning CT. The latter approach has been shown to change the treated volumes compared with CT and in prostate cancer patients has been shown to reduce interobserver variation. There are remaining issues with four-dimensional motion that are yet to be fully appreciated or overcome. 2-[18F] fluoro- 2-deoxy-D-glucose positron emission tomography/CT co-registered with planning CT results in smaller volumes than CT alone and also reduces interobserver variation, but requires further validation before routine implementation. Experimental work utilising novel positron emission tomography tracers and diffusion- weighted MRI shows promise and requires further evaluation. Rigorous quality assurance is important with processing of newer imaging modalities. Further work needs to be conducted into both interobserver variation and the formal evaluation of the clinical benefits of newer imaging modalities. Developments in image- guided radiotherapy are also required to ensure that improvements in target definition at the planning stage are reproducible throughout treatment. Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. Key words: CT simulation; interobserver variation; MReCT co-registration; MRI planning; PET/CT; radiotherapy planning; rectal cancer Statement of Search Strategies Used and Sources of Information A review of published studies and conference abstracts was conducted using MEDLINE and EMBASE databases based on the terms ‘rectal cancer, radiotherapy planning, CT simulation, CT planning, MRI planning, PET-CT, MR-CT co-registration, interobserver variation’ up to August 2010. The search was limited to human subjects and English language, but no date limits were applied. This was supple- mented by hand searching of conference abstracts from ESTRO and ASTRO 2008 and 2009. These were then assessed for relevance by two authors (SG and SM). Studies were only included if they examined advanced imaging modalities to define radiotherapy target volumes for preoperative radio- therapy of rectal cancer. Introduction Rectal cancer affects over 14,000 people a year in the UK [1]. The last two decades have seen significant advances in the radiological staging [2], pathological staging [3] and surgical management of rectal cancer [4,5], but the 5 year survival is still only 45% [6]. The radiotherapy process has also changed, with an increased use of conformal radiotherapy to reduce toxicity. Imaging modalities that clearly and Author for correspondence: S. Gwynne, Velindre Cancer Centre, Cardiff CF14 2TL, UK. Tel: þ44-29-2061588; Fax: þ44-29-20694181. E-mail address: [email protected](S. Gwynne). Contents lists available at SciVerse ScienceDirect Clinical Oncology journal homepage: www.elsevier.com/locate/clon 0936-6555/$36.00 Ó 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.clon.2011.10.001 Clinical Oncology xxx (2011) 1e 12 Please cite this article in press as: Gwynne S, et al., Imaging for Target Volume Delineation in Rectal Cancer Radiotherapy d A Systematic Review, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
Transcript
at SciVerse ScienceDirect
Clinical Oncology xxx (2011) 1e12
Contents lists available
Clinical Oncology
journal homepage: www.elsevier .com/locate/c lon
Overview
Imaging for Target Volume Delineation in Rectal Cancer Radiotherapy d
A Systematic Review
S. Gwynne *, S. Mukherjee y, R. Webster *, E. Spezi z, J. Staffurth x, B. Coles jj, R. Adams x*Department of Clinical Oncology, Velindre Cancer Centre, Cardiff, UKyGray Institute for Radiation Oncology and Biology, Department of Oncology, Old Road Campus Research Building, University of Oxford, UKzDepartment of Physics, Velindre Cancer Centre, Cardiff, UKxDivision of Cancer, School of Medicine, Cardiff University, Cardiff, UKjjCancer Research Wales Library, Velindre Cancer Centre, Cardiff, UK
Received 28 February 2011; received in revised form 15 September 2011; accepted 19 September 2011
Abstract
The global move towards more conformal radiotherapy for rectal cancer requires better imaging modalities that both visualise the disease accurately and arereproducible; to reduce interobserver variation. This review explores the advances in imaging modalities used in target volume delineation, with a view to makerecommendations for current clinical practice and to propose future directions for research. A systematic review was conducted using MEDLINE and EMBASE.Articles considered relevant by the authors were included. Planning with orthogonal films is being replaced by computed tomography (CT) simulation. This is nowconsidered the ‘gold standard’ and allows conformal three-dimensional planning. Magnetic resonance imaging (MRI) has been shown to overcome some of thelimitations of CT and can be used either as a diagnostic image to visually aid planning, or as a ‘planning’ MRI carried out in the treatment position andco-registered with the planning CT. The latter approach has been shown to change the treated volumes compared with CT and in prostate cancer patients has beenshown to reduce interobserver variation. There are remaining issues with four-dimensional motion that are yet to be fully appreciated or overcome. 2-[18F] fluoro-2-deoxy-D-glucose positron emission tomography/CT co-registered with planning CT results in smaller volumes than CT alone and also reduces interobservervariation, but requires further validation before routine implementation. Experimental work utilising novel positron emission tomography tracers and diffusion-weighted MRI shows promise and requires further evaluation. Rigorous quality assurance is important with processing of newer imaging modalities. Further workneeds to be conducted into both interobserver variation and the formal evaluation of the clinical benefits of newer imaging modalities. Developments in image-guided radiotherapy are also required to ensure that improvements in target definition at the planning stage are reproducible throughout treatment.� 2011 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
Statement of Search Strategies Used andSources of Information
A review of published studies and conference abstractswas conducted using MEDLINE and EMBASE databasesbased on the terms ‘rectal cancer, radiotherapy planning, CTsimulation, CT planning, MRI planning, PET-CT, MR-CTco-registration, interobserver variation’ up to August 2010.The search was limited to human subjects and Englishlanguage, but no date limits were applied. This was supple-mented by hand searching of conference abstracts from
Author for correspondence: S. Gwynne, Velindre Cancer Centre, CardiffCF14 2TL, UK. Tel: þ44-29-2061588; Fax: þ44-29-20694181.
0936-6555/$36.00 � 2011 The Royal College of Radiologists. Published by Elsevidoi:10.1016/j.clon.2011.10.001
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
ESTRO and ASTRO 2008 and 2009. These were then assessedfor relevance by two authors (SG and SM). Studies were onlyincluded if they examined advanced imaging modalities todefine radiotherapy target volumes for preoperative radio-therapy of rectal cancer.
Introduction
Rectal cancer affects over 14,000 people a year in the UK[1]. The last two decades have seen significant advances inthe radiological staging [2], pathological staging [3] andsurgical management of rectal cancer [4,5], but the 5 yearsurvival is still only 45% [6]. The radiotherapy process has alsochanged, with an increased use of conformal radiotherapy toreduce toxicity. Imaging modalities that clearly and
er Ltd. All rights reserved.
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e122
accurately visualise the disease, improved understanding oforgan motion and a better process of quality assuranceshouldminimise the risk of geographicalmiss. The aimof thisreview is to look at the advanced imaging modalities thathave been used for target volume delineation (TVD) in rectalcancer tomake recommendations forcurrent clinical practiceand to propose future directions for research.
Role of Radiotherapy in Rectal Cancer
Three radiotherapy approaches are commonly in use forrectal cancer in the UK: short-course preoperative radio-therapy (SCPRT) (25 Gy in five fractions over 5 days fol-lowed by surgery within 1 week) has been shown to reducepelvic recurrences in resectable rectal cancer [7,8]; long-course preoperative chemoradiotherapy (45e50.4 Gy in25e28 fractions) has been shown to downstage tumourswhen the circumferential resection margin is threatened,improve margin negative (R0) resection rates and reducelocal recurrence rates [9,10]. Preoperative treatment isrecommended where possible [11] and postoperativechemoradiotherapy after R1 resection, which has becomeless frequent as preoperative staging has improved [9].
What should be the Optimal Target Volume?
There is an increasing consensus on the structures thatshould be included in the target volume, mainly guided byrecurrence data [12e14] and recent Radiation TherapyOncology Group guidelines have been published [15].Current practice is to include the same structures for bothSCPRT and LCPRT. Roels et al. [12] conducted a systematicreview to determine what should be included in the clinicaltarget volume (CTV) based on published recurrence data andconcluded that the tumour, mesorectum, presacral regionand internal iliac lymph nodes should be included, anapproach broadly adopted by the National Cancer ResearchInstitute-approved ARISTOTLE trial for rectal cancer. WithinARISTOTLE, the gross tumour volume (GTV) is defined as allgross sites of disease, and would include the whole rectalwall at the level of the tumour with a 1 cm margin to createCTV A. This would then be added to CTV B, which includesthe whole mesorectum and nodes to create a final CTV,which is then grown by 1 cm to create the planning targetvolume (PTV). Although the recurrence data used to definethese volumes is based on patients planned conventionally,locoregional control seems favourable with more conformalmethods, albeit with limited follow-up [15]. This evolutiontowards conformal radiotherapy has also demandedimproved methods to accurately assess the spatial relation-ship between target volumes and surrounding tissues [16].
Orthogonal Films (Conventional Planning) and ComputedTomography Planning
Traditionally, radiotherapy planning for rectal cancer hasbeen based on two-dimensional radiological anatomy. Rectalcontrast (with or without oral contrast) and bony landmarks
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
were used to delineate the treatment volumes [17e19],supplemented by clinical examination to aid definition ofthe inferior extent of the tumour. This should be consideredoutdated, having been replaced by computed tomography(CT) planning with all available diagnostic imaging asa minimum, as there are major limitations affecting theability to accurately define the tumour itself [17,20] and anylocal extension into the mesorectum or surrounding struc-tures [21]. Studies have shown that CT planning hasadvantages over orthogonal films in terms of better defini-tion of anterior and superior borders and reduced toxicitycompared with historical controls [17e19,22e24]. However,CT simulation has its limitations because of poor contrastbetween faeces and tumour, partial volume effects due tothe curves/valves of Houston and imaging of the horizontalsigmoid [21]. Some improvement in the CT image for con-touring can be achieved by changing the grey scale tomaximise the contrast between the soft tissue infiltrationand normal fat. The routine window for abdominal CT is notoptimal for this purpose and Myerson and Drzymala [25]recommend a level of about e60 Hounsfield units anda somewhat larger than usual window of about 600Hounsfield units to help better identify both loops of boweland perirectal soft tissue densities.
Advances in Radiotherapy Planning
Magnetic Resonance Imaging
Magnetic resonance imaging (MRI) is generally consid-ered the gold standard for staging rectal cancer [2] and inthe era of conformal radiotherapy most UK radiationoncologists would have MRI images available at the time ofplanning to aid the delineation of the target volume, a factconfirmed by recent pan-UK audit data (Katharine Parker,personal communication). MRI addresses many of thelimitations of CT, such as definition of depth of invasionthrough the rectal wall into local structures [21] andextension into presacral space and mesorectal circumfer-ence [12,21], which are high-risk areas for recurrence [12].The visual transfer of data from MRI to CT is susceptible toerrors in interpretation and transfer [26]. One method toovercome this is co-registration of the images where MRimages are used for optimal outlining while retaining the CTdata for dose calculations. This is now considered the goldstandard in prostate cancer radiotherapy planning [27] andis recommended in a 2004 Royal College of Radiologists’document [28]. The images can be co-registered eithermanually or automatically using the planning software. Theformer can be inefficient and error prone [29], whereasautomated approaches reduce interoperator variability andallow more accurate registration between multimodalityimaging systems [29,30].
Diffusion-weighted Magnetic Resonance Imaging
Diffusion-weighted MRI evaluates the diffusion capacityof water molecules and obtains information about
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e12 3
microscopic structures such as cell density or necrotic cellclusters, which indirectly assesses tumour aggressiveness[31]. In quantitative diffusion-weighted imaging, themagnetic resonance signal arises from both intracellular andextracellular compartments, and the result is given in termsof the apparent diffusion coefficient (ADC), which isa weighted sum of these contributions. An early increase inmean tumour ADC and a low pretherapy mean ADC in rectalcarcinoma have been shown to correlate with a goodresponse to chemoradiotherapy [32].
Positron Emission Tomography
CT has a relatively high spatial resolution, but is limited inrespect of specificity. In contrast, 2-[18F] fluoro-2-deoxy-D-glucose (FDG) positronemission tomography (PET) canbetterdefine macroscopic colorectal cancer tissue and improvesstaging accuracy before surgery or at restaging [33e35]. PETTVD can be carried out by hand orwith the help of automaticcontouring technologies. The former approach is limited bypotential interobserver variability induced by factors such aslimited resolution of PET images, display settings (windowlevel and width, display colour map and saturation) andexperience of the physician [36,37]. Although this approach isused commonly in clinical practice, no standardised tech-nique has been published [38]. The latter can be achieved by‘thresholding’, which classifies the pixels of an image asa function of their value. These values can be absolute, whenthe PET image is calibrated in terms of standardised updatevalues (SUV) [39e42] or relative, when the image is normal-ised to the uptake value in a specific region of the body (e.g.target). It has been shown that PET-based tumour volumesare strongly affected by the choice of the threshold level [43].A value of about 40% of the maximum value in the target isoften reported for methods using a fixed threshold level[44e46]. Other approaches, such as region growing [47,48] orgradient based [49,50], have been successfully used fortumour volume segmentation on PET images.
Although PET/CTwith FDG has a high sensitivity for rectalcancer, there are some limitations in specificity, mainly dueto the uptake of FDG in macrophages. Efforts have thereforebeen made to develop other more tumour-specific tracers.Fluorothymidine (18F-FLT) and fluoromisonidazole(18F-FMISO) are two such agents, the former a measure ofcell proliferation, the latter of hypoxia [51].
Interobserver Variation
TVD is an inherently observer-biased procedure with theextent of tumour defined differently by different practi-tioners, even within the same department, leading tosignificant inter- and intraobserver variation in multipletumour sites if CT alone is used [35], and this has beenshown to negatively affect trial outcomes in pancreatic andhead and neck cancer [52,53]. There is a paucity of pub-lished studies quantifying interobserver variation in rectalcancer radiotherapy outlining in the conformal era, but therecently published consensus atlas for outlining anorectalcancer was initiated because of perceived inadequate
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
contouring within a chemoradiotherapy trial for anal cancer[15]. There is some evidence from other cancer sites thatadditional imaging with MRI and PET can reduce interob-server variation [38,54].
Objective
The aim of this review was to examine the benefits ofadvanced imaging techniques over the use of CT withdiagnostic MRI alone in terms of TVD for rectal cancer andreducing interobserver variation and to make recommen-dations for clinical practice.
Materials and Methods
Our initial search used MEDLINE and EMBASE databasesusing the terms ‘rectal cancer, radiotherapy planning, CTsimulation, CT planning, MRI planning, PET-CT, MR-CT co-registration’. We followed the same procedure as a Cochranereview. Both articles and conference abstracts wereincluded. We searched MEDLINE from 1948 up to August2010. This yielded 545 titles. One author (SG) independentlyassessed the abstracts or, in the event of uncertainty, the full-length articles to determine whether they met the inclusioncriteria and discussed with another author (SM). Articles onwhich both authors concurred were included. Initial selec-tion involved the exclusion of any studies that did not reporton imaging modalities to define the target volume or reduceinterobserver variation in the TVD for preoperative radio-therapy of rectal cancer. We excluded papers looking atimaging for staging and treatment response. In view of thesmall number of studies for eachmodality, we did not set theminimum number of patients as an inclusion criteria. Onlypapers published in English were included. This left 15 titles,two of which were conference abstracts [55,56]. SG alsohand-searched abstracts from ESTRO 27, ASTRO 2008 andASTRO 2009 and identified two further studies [31,57].Overall, 17 titles [13 full-length articles and four conferenceabstracts] were available for analysis. No studies wereexcluded for methodological weakness, as there wasa limited number of studies, but they were classifiedaccording to the system of levels of evidence published bythe Oxford Centre for Evidence-based Medicine e Levels ofEvidence in 2009 [58].
Analyses
Details regarding the number of patients, stage oftumour, imaging modality investigated and main outcomesof the studies were recorded. For interobserver variationstudies, the number of observers, intervention used andmain outcomes were recorded.
Results
Seventeen studies were identified. Fifteen were obser-vational studies (one multicentre [59] and 14 single centre
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e124
[31,33,34,51,55e57,60e66]), one a double-blind rando-mised control trial [67] and one a caseecontrol study [68].One study looked at clips/fiducial markers [68], three at MRI[60e62], eight at PET [33,34,51,56,57,64e66] and one atdiffusion-weighted MRI [31]. One study looked at both MRIand PET/CT [55] and three studies looked at interobservervariation in outlining [59,63,67].
Clips/Fiducial Markers
In a single-centre caseecontrol study, Vorwerk et al. [68]placed gold markers in the mesorectum at the caudal end ofthe tumour using rigid rectoscopy to aid identification ofthe caudal extent of the GTV. In all patients, the insertionwas simple and they found no increased toxicity, pain orbleeding in this group compared with patients withoutmarkers, based on data from historical controls. Movementof markers during the course of treatment was low, witha mean of <0.25 cm in all directions. There was no detectedmarker loss during the treatment period [68].
Table 1Magnetic resonance imaging (MRI) studies
Reference No. patients Stage Methods
[55] 19 Short-courseradiotherapy
Volume of GTV on Mand PET/CT correlatedwith pathologicalmeasurements
[60] 10 T3/4low rectalcancer
CT planning scan2e3 weeks later MRIscan in same positionMeasured tumourlength, width, distancto anal verge and bonlandmarks (sacralpromontory, coccyxand upper and lowermargins of pubis) forboth imagingmodalities
[61] 15 T3/N1LCPCRT
MRIeCT fusionGTVs drawn on bothCT and fused MRIeCT
[62] 10 T3/T4 Boost volume drawnon MRI and then fusewith planning CT scan
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
In summary, only one study has addressed this issue andalthough it may be beneficial identifying the lower borderof the tumour, further work is required to define theoptimal role for clips/fiducial markers in rectal cancerradiotherapy through a larger number of studies. Theadditional information derived from advanced imagingmayreplace the value of clips.
Magnetic Resonance Imaging-based Planning Techniques
Magnetic resonance imagingFour studies using MRI for TVD were identified
[55,60e62]. These are detailed in Table 1. O’Neill et al. [60],in a retrospective single-centre study of 10 patients,compared GTV delineation on CT and MRI and found thatthe CT consistently overestimated tumour dimensions (CTvolume mean of 18 cm3 greater than MRI) and there wasa statistically significant difference in the height of theproximal end of the tumour from the anal verge, tumourwidth and tumour volume (P� 0.05). In an observational
MRIsequence
Comments
RI NS MRI-based measurements correlatedwith Pearson coefficient of 0.75(P< 0.001)
ey
T2Slicethickness2.5 mm
CT volume mean of 18 cm3 greaterthan MRIStatistically significant increase in meantumour length, maximum width andheight of proximal tumour from theanal verge on CT compared with MRI(3.23 cm, 0.48 cm and 2.93 cm greater,respectively)
Pre- andpostcontrastaxial,coronalsagittal T1and T2imagesSagittalSTIR imagesSlicethickness4 mm
The mean CT-GTV/MR-GTV ratiowas 1.2 (range 0.5e2.9)Major discrepancies were noted inthree cases.Overestimation of GTV by CT usuallya result of faeces in close proximityto the tumourUnderestimation of the GTV by CTdue to difficulty visualising extent oftumour within anus and sigmoid colon
dSagittal andaxial T2Slice thickness5 mm
Radiologist uninvolved in the care ofthe patient drew boost volumes
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e12 5
study of nine cases, Buijsen and Bogaard [55] looked at thevolume of GTV on MRI and correlated this with pathology.They found that there was reasonable correlation witha Pearson correlation of 0.75. They did not compare CT-based GTV with pathology, but found very poor correla-tion between endoscopic measurements and pathology(Pearson correlation 0.64).
Magnetic resonance imagingecomputed tomography fusionTwo studies have looked at MRIeCT fusion, one for
delineating the CTV [61], the other for delineating a boostvolume [62] (Table 1). In a prospective single-centre obser-vational study, Tan et al. [61] delineated GTV in 15 cases onboth simulation CT and fused MRIeCT and analysed thevolume and spatial relationship of the CT-basedGTVwith theMRI-basedGTV.Although inmost cases theCToverestimatedthe GTV, usually as a result of faeces in close proximity to thetumour, underestimation of the GTV by CT was seen in twopatients due to difficulty visualising the extent of tumourwithin the anus and sigmoid colon. In another prospectivesingle-centre observational study, Seierstaad et al. [62] usedCTeMRI co-registration to plan a simultaneous integratedboost for rectal cancer radiotherapy. The focus of the paperwas dose escalation of the GTV rather than TVD and did notassess any comparisons with CT.
Diffusion-weighted magnetic resonance imagingBrussel et al. [31], in a single-centre observational study
published only in abstract form, used diffusion-weightedMRI in five rectal cancer patients undergoing intensity-modulated radiotherapy for dose painting. The informa-tion gained was used to boost areas with the maximumdiffusion restriction on high b value (1000 s/mm2) images.Dose volume constraints for small bowel were met in allfive patients with a combination of six, seven or eightbeams. No outcome data were presented.
In summary, incorporating MRI in TVD leads to a smallerGTV compared with CTalone, but it has been investigated ina small number of patients (44 in total). Only one study hascorrelated the MRI volume with pathology [55]. There aretechnical issues that need to be overcome before thewidespread implementation of CTeMRI co-registration forradiotherapy planning and further work is required tooptimise the use of MRI in TVD for rectal cancer. Until suchtime, MRI should be available at the time of planning toassist in TVD, accepting the inherent errors that can beassociated with this approach. With only one study, withfive patients, published in abstract form, diffusion-weighted MRI is an exciting area for future work, butcannot currently be recommended for routine use on thecurrent evidence.
Nine studies looked at PET for rectal cancer planning[33,34,51,55e57,64e66]. All were single-centre observa-tional studies with a maximum of 36 cases. Details of thesestudies are given in Table 2. In all studies, the CT and PETscans were carried out in the prone position. These studies
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
showed that PET can affect radiotherapy planning bydetecting new areas of locoregional or metastatic diseasealtering either PTV extent or treatment intent [33,57,64,65].Studies have also compared the GTV volumes between CTand PET. Two studies showed the CT volume to be greater[33,65], whereas two studies showed the PET-definedvolume to be greater [56,64]. Anderson et al. [33] showedthat the correlation between the two volumes increased asthe size of the tumour increased.
Four studies [34,51,57,66] used an automated delineationtechnique. Day et al. [66] studied 18 rectal cancer patientswho had received preoperative radiotherapy with volumesbased on CT. They retrospectively segmented the PET scansand estimated the tumour volume using four methods d
PET avid areas, SUV2.5, SUV43% and an auto-contouringmethod using the insight toolkit confidence connectedregion growing technique (CCRG). The latter is based ona region growing method using the pixel intensity data ofthe tumour region in an iterative statistical manner in orderto detect the edge of the tumour mass. PET avid volumeswere delineated by a radiation oncologist, and werereviewed by a nuclear medicine physician and a secondradiation oncologist. Volumes of tumour from the PET avidcontours ranged from 10.5 to 129.5 cm3 with a median of24 cm3. CCRG volumes were very similar with a range of10e110.4 cm3 and a median of 23 cm3. SUV43% volumeswere smaller than PET avid volumes, whereas SUV2.5 wasboth greater and smaller than the PET avid volumes [66].
One method to validate PET volume delineation wouldbe correlation with pathology and this was carried out byCiernik et al. [34] and Buijsen and Bogaard [55]. The formershowed that the pretreatment PET volumes were smallerthan pathological specimens after chemoradiotherapydespite the fact that the neoadjuvant therapy should havereduced the size of the tumour [34]. The authors pointedout that this may not have been an appropriate comparisonbecause of the expected change in size of the tumour overtime. On the other hand, Buijsen and Bogaard [55] looked atGTV volume; PET CT versus pathological measurementshowed good correlation, with a Pearson correlation of 0.91.Their patient group had SCPRT and surgery within 3 daysand so no significant change in tumour size would beanticipated. This may be a better patient group in which tovalidate CT PET by pathology, although this generally relatesto less advanced tumours, in which microscopic extensionis less likely.
Roels et al. [51] conducted a study of 15 patients withbiopsy-provenT2/3-N1/2MO adenocarcinoma of the rectumcomparing the use of these new tracers for planning radio-therapy. Ten patients underwent PET CT with FDG and FLTand the other five with FDG and FMISO. Both PET CT scanswere taken in the prone positionwith the use of a belly boardand therewasat least 24hbetween the scans to avoidoverlapof tracers. All the images obtainedwith all three tracers wereautomatically processed and outlined by use of a gradient-based segmentation method, a technique validated in headand neck cancer [51]. Scans were taken before therapy andafter 10 days of chemoradiotherapy. The mean FDG-PETtarget volume before and during chemoradiotherapy was
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
Reference No. patients Stage Tracer Method of defining PET volumes Volume of GTVon CT (cm3)
Volume of GTVon CT PET (cm3)
Comments
[33] 23(20 rectal,3 anal)
Long-coursechemoradiotherapy
FDG Radiation oncologist delineationSUV 2.5
Mean 99.6(17e570)
Mean 91.7(2.9e859)
17% had change in PTV26% had some change in planningprocess
[64] 25 T3/T4Short- and long-course radiotherapy
Consensus delineation of tworadiation oncologistsSUV 40%
77.2 96.8 Fixed threshold value of 40% ofmaximum uptakePET/CT increased GTV by 25%Stage variation in 12%Treatment intent change in 4%
[57] (abstractonly)
24 T1-T4,N0-3, M0 FDG Automated region growingalgorithm
NS NS Mean length of tumour smaller onPET/CT than colonoscopy, but biggerthan on endorectal ultrasound
[55] (abstractonly)
19 Short-courseradiotherapy
FDG Automatic tumour delineationSUV 40%
NS NS Strong correlation between PET GTVand pathologyPearson correlation of 0.91(P� 0.001)
[34] 11 T2/3 N0/1 FDG Automatic delineation NS See text CT GTV correlated better withpathology than PET GTVSee text for further discussion
[66] 18 NS FDG Four methods used:PET avidSUV2.5
SUV43%
CCRGPET avid volumes drawn byradiation oncologistOther three automatic
NS NS PET avid and CCRG volumes very similarSUV 43% volumes smaller than PET avidSUV 2.5 volumes both greater andsmaller than PET avidSee text for further discussion
Mean 163 Mean 62 PET GTV only areas of SUV >2.5PET volumes smaller than CT(P< 0.05)46% had need for modification ofusual target volumes because ofdetection of geographical miss
[56]* 23 FDG One radiation oncologistSUV 40%
Median 25.7 cm3
(3.17e135.6)Median 40 cm3
(10.6e177.7)Median of 35% of volumeidentified by PET but not CT
[51] 15 T2/3-N1/2 M0 FDGFMISOFLT
Automatically processed by useof gradient-based segmentationmethod
NS See text FDG-PET volumes significantlyhigher than the FMISO-PETtarget volumeSee text for further details
CT, computed tomography; SUV, standardised uptake value; CCRG, confidence connected region growing technique; FDG, 2-[18F] fluoro-2-deoxy-D-glucose; GTV, gross tumourvolume; PTV, planning target volume; FLT, fluorothymidine; FMISO, fluoromisonidazole.* At the time of the literature search, only the abstract was available. This has now been published in full as [76].
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e12 7
significantly larger than the FMISO-PET target volume(P¼ 0.001), but no significant difference was seen betweenthe mean FDG and FLT volumes. They recommended thatFDG- and FLT-PET are used for radiotherapy planning as FLTshows better spatial correspondence and the uptake is morestable over time in comparison with FMISO. They did notcarry out any pathological correlation and this would behelpful in validating this technique further [51].
In summary, there have been nine studies with a total of198 patients looking at the use of PET with various radio-isotopes in addition to CT for TVD. All of these studies arelevel 4 evidence. PET seems to be promising, but there areseveral methodological issues that need to be addressed.Two important issues are pathological validation of PET andcomparison with MRI.
Interobserver Variation in Outlining
Three studies [59,63,67] looked at interobserver varia-tion in outlining rectal cancer volumes. These are shown inTable 3. Fuller et al. [67], in a double-blind, randomisedcontrolled pilot study showed that access to a contouringatlas can improve interobserver variation in rectal cancerdelineation, particularly for the nodal areas. The atlas is anintervention that can be implemented widely, as it is nowwidely available on the internet [69]. Krengli et al. [59] ina multicentre observational study with two cases and 10observers and Patel et al. [63] in a single-centre study withsix cases and four observers, showed that the addition ofPET can reduce outlining variation, which is in keeping withstudies in non-small cell lung cancer [38].
Interobserver variation has not been well studied inrectal cancer and only one published study has looked at anintervention to improve the variation on CT scanning alone.This was a well-designed randomised controlled trial andconstituted level 1 evidence. Although incorporating PETinto radiotherapy planning is promising, it is subject to thelimitations discussed under PET above and cannot currentlybe recommended. The focus should be on training and theuse of atlases.
Table 3Studies on interobserver variation
Reference No. participants Intervention
[67] 15 radiation oncologists1 rectal cancer case
Contouring atlas
[59] 10 radiation oncologists2 rectal cancer cases
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
Discussion
We have shown that the evidence for the use ofadvanced imaging in TVD for rectal cancer is level 4 or less.MRI and PET used for radiotherapy planning can reduce thesize of the GTV (and therefore potentially spare normalstructures). In addition, FDG-PET can aid identification ofparts of the GTV that are not detected by CT or MRI and canreduce interobserver variability in GTV delineation. It mayalso allow identification of areas of high metabolic activitywithin the GTV that could benefit from dose escalation.However, it does not help to delineate all structures withinthe CTV as these are often not PET positive (Figure 1). Theavailable evidence on PET for TVD in rectal cancer has beenreviewed by the IAEA [38]. Their current guidance suggestsPET CT is only used routinely in the radiotherapy planningof non-small cell lung cancer because of its accuracy instaging and the demonstration of a powerful effect ontreatment volumes in all the published radiotherapy plan-ning studies [38]. There are some data on rectal cancer, butthe volume and quality of the studies cannot currentlycompare with non-small cell lung cancer. They also advisecaution when using automated approaches for PET out-lining because of the inability to distinguish FDG uptakefrom benign and malignant processes [38] and Ciernicket al. [34] found it was difficult to define the border in thedirection of neighbouring signals such as the bladder.Further work is required to define the optimal methods ofdelineation before its widespread implementation. TheNational Cancer Research Institute framework for PETresearch also highlights this as an area requiring furtherresearch [70]. In particular, clinical research protocols mustbe carefully defined and tumour delineation methodsstandardised.
There are technical issues associated with theseadvanced imaging modalities that need to be overcome.This includes imaging and scan acquisition protocols. Theformer for MRI requires the optimal sequence for rectalradiotherapy planning to be defined. Work by Khoo et al.[71] on prostate cancer concluded the best sequence was
Method Comments
2 groups asked to contourvolumes on T3NO rectalcancer before and afteraccess to RTOG contouringatlas
Access to the atlas increasedvolumetric agreement onCTV A (internal iliac,presacral and perirectalnodes) between observers
5 participants outlined onCT alone, 5 on PET/CT
PET/CT increasedreproducibility of GTVdelineation
Hypothetical tumour boost Reduced interobservervariability, especially fornodal disease
Fig 1. Computed tomography simulation (left) and 2-[18F] fluoro-2-deoxy-D-glucose (FDG) positron emission tomography/computed tomog-raphy image (right) showing rectal cancer. These images show the benefit of positron emission tomography in determining the gross tumourvolume but show, in this case, how it does not help delineate the non-FDG avid areas that would be included in the clinical target volume.
Fig 2. Co-registered planning computed tomography and magnetic resonance imaging scan showing discordant rectal and bladder filling on thetwo modalities. The magnetic resonance imaging scan was carried out 45 min after the planning computed tomography scan.
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e128
Please cite this article in press as: Gwynne S, et al., Imaging for Target Volume Delineation in Rectal Cancer Radiotherapy d A SystematicReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e12 9
F3D [T1-weighted spoiled gradient echo (fast low angleshot FLASH) sequence]. This may provide a good base withwhich tomove forward with developingMRI sequences forrectal cancer. The studies included in this review usedmostly T2 images [60,62], with one also using T1 and STIRimages [26]. For PET, the equivalent issue will be thethreshold levels and clear protocols for TVD. Recently, theEuropean Association of Nuclear Medicine published a setof guidelines for the use of FDG-PET and PET/CT for tumourimaging, including patient preparation, scan acquisitionand reconstruction parameters and reporting [72],whereas the IAEA have published an expert report on theuse of PET and PET/CT specifically for radiotherapyplanning [38].
For MRI scan acquisition, a flat table top must be usedand the scanning table must reproduce patient set-up onthe simulator and radiotherapy treatment couch. Anyimmobilisation device needs to be assessed for itsmagnetic safety as well as being able to fit into the patienttunnel and not interfere with the MRI examination coils[29]. For PET, the IAEA have published an expert report onthe use of PET and PET/CT specifically for radiotherapyplanning, emphasising the need for rigorous qualityassurance as would be applied at every other step in theplanning process. This would include the use of a firm flatcouch top, appropriate immobilisation devices, laserbeams for patient alignment and a wide bore scanner; thelatter to allow for fixation systems and/or positioningtechniques [38].
Co-registration is the best way to utilise the newimaging modalities, but despite being recommended bythe Royal College of Radiologists in 2004, only two studieshave examined its use and it has not gained widespreadfavour among the rectal radiotherapy community in theUK. One issue has been that the MRI scanning for rectalcancer has traditionally been conducted in the supineposition, whereas many centres prefer the prone positionfor radiotherapy planning and delivery. The reason forusing the prone position is that the volume of small bowelwithin the field is less, but there are issues with positionalreproducibility and discomfort for patients, particularlythose with a stoma. More recently it has been shown thatMRI images can be obtained successfully in the proneposition [30]. A further potential issue with co-registrationis differences in bladder and rectal filling between the twomodalities if these are not carried out within a short timeof one another and can lead to substantial discordance[61,73] (Figure 2). Tan et al. [61] attempted to overcome thevariation in rectal filling by asking patients to empty theirbowels before imaging and during treatment delivery. Arecently published study that co-registered MRI and PETimages for cervical cancer radiotherapy planning useda urinary catheter to overcome the issues of bladder filling[74]. The differences seen may be beneficial as it gives anidea of potential organ motion during the 5 day or 5 weekcourse of radiotherapy. However, the issues surroundingorgan motion in rectal cancer are complex and are not thesubject of this paper; they are dealt with more fullyelsewhere [75].
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
Limitations of this Systematic Review
Although there is a wide range of modalities available,none has been studied in randomised clinical trials and asa result the studies included in this review were oftensingle-centre series with a small number of patients.Because of the small number of studies available and theaim to include as much current data as possible, we haveincluded some studies that have only been published inabstract form. We have not been able to excludepublication bias, but have attempted to reduce this byincluding conference abstracts. We have not been ableto comment on the quality of the imaging used in thestudies.
Conclusions
Progress has clearly been made since the early days ofradiotherapy planning when orthogonal films and rectalcontrast were used to define target volumes. Conformalplanning with CT has become the new standard, but thereare limitations with this technique. In addition to theimaging modalities discussed, it is important not to over-look the importance of integrating information gainedfrom endoscopy and rectal examination. MRI and PET areboth being used increasingly for TVD. Based on theavailable evidence, the authors recommend that thediagnostic MRI scans be available at the time of planningand that co-registration is undertaken with caution, withattention given to the changes in organ motion andfilling. PET/CT for radiotherapy planning remains experi-mental and should be used in a research setting untilfurther data are available and recommendations for its usemade from an official body, akin to the situation in lungcancer [38].
Intra- and interobserver variation will probably becomean increasingly important issue in the era of radiotherapyclinical trials for rectal cancer and although there are somedata to suggest the use of MRI and PET may reduce thisvariation, training in TVD, such as the workshops forARISTOTLE, may reduce this variation without the need foradditional imaging. This hypothesis needs to be explored ina prospective clinical trial.
Recommendations for current practice are detailed inBox 1 and priorities for further research and developmentin the field are detailed in Box 2. As newer modalities areintroduced into radiotherapy planning, there should bea parallel effort to establish appropriate quality assurancefor these newer modalities. Newer modalities may pick upsmall volume disease that would have previously goneuntreated leading to increased PTVs corresponding to anincrease in doses to organs at risk. Parallel development ofimage-guided radiotherapy techniques will becomecrucial to limit this damage as will be collection ofoutcome data to ensure that moving to these morecomplex radiotherapy techniques translate into true clin-ical benefits.
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
Conformal radiotherapy is now the gold standard in rectal cancer, but progress is being made towards intensity-modulatedradiotherapy and dose escalation. These will require more advanced imaging than computed tomography (CT) alone.
Magnetic resonance imaging (MRI).
The ability to delineate the target volumes on MRI co-registered with CT will aid target volume delineation (TVD) and potentiallyreduce interobserver variation. The technology for co-registration exists and can be used in both the prone and supine positions.
Issues that need addressing are defining the optimal MRI sequence for radiotherapy planning and the issues of organmotion.With theincreasing availability of MRI scanning in oncology centres this should be a priority for further work.
The IAEA report recommends the use of PET in lung cancer radiotherapy planning because of the large number of studies that havevalidated its use. Further studies are required in rectal cancer to determine whether this is an approach that should be widely adopted.
Issues that need addressing are the optimal method of defining tumour and validation of autosegmentation methods that wouldreduce interobserver variation.
PET/MRI fusion
In one study, PET showed better correlation with pathology than MRI and a future study should combine PET and MRI to combine theanatomical and functional information to provide the optimal method of TVD, as has been done in cervical cancer.
Diffusion-weighted MRI/novel PET tracers.
For dose escalation it will be necessary to identify areas that may benefit from a boost and diffusion-weighted MRI and novel PETtracers possibly hold promise in identifying areas of tumour that could be boosted.
Image-guided radiotherapy
All of these approaches, while improving TVD, will not improve radiotherapy delivery unless it is accompanied by a paralleldevelopment in image-guided radiotherapy techniques.
Box 1. Recommendations for clinical practice
Orthogonal films are outdated and should be replaced with computed tomography (CT) planned conformal techniques. (Level 3evidence e Grade B recommendation)
Diagnostic magnetic resonance imaging (MRI) scans should be available at the time of planning.
(Level 5 evidence e Grade D recommendation).
Interobserver variation is an important issue in the era of conformal radiotherapy and an increased emphasis on training is needed.(Level 1 evidence e Grade A recommendation)
There are insufficient data to support the use of clips and gold markers for target volume delineation (Level 4 evidence)
Fusion of diagnostic MRI with planning CT should be carried out with caution, taking into consideration the possible issues with organmotion. (Level 4 evidence e Grade C recommendation)
Fusion of positron emission tomography scans with the planning CT should only be undertaken in a research setting. (Level 5evidence e Grade D recommendation)
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e1210
Acknowledgement
SM is part-funded by NIHR Biomedical Research, Oxford.
References
[1] Cancer Research UK. Latest UK cancer incidence and mortalitysummary. Available at: http://info.cancerresearchuk.org/prod_consump/groups/cr_common/@nre/@sta/documents/generalcontent/crukmig_1000ast-2735.pdf. [accessed21.02.11].
[2] MERCURY Study Group. Diagnostic accuracy of preoperativemagnetic resonance imaging in predicting curative resection
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
of rectal cancer: prospective observational study. Br Med J2006;333:779.
[3] Quirke P, Durdey P, Dixon MF, Williams NS. Local recurrenceof rectal adenocarcinoma due to inadequate surgical resec-tion. Histopathological study of lateral tumour spread andsurgical excision. Lancet 1986;2:996e999.
[5] Carlsen E, Schlichting E, Guldvog I, Johnson E, Heald R. Effectof the introduction of total mesorectal excision for the treat-ment of rectal cancer. Br J Surg 1998;85:526e529.
[6] Cancer Research UK. Relative five-year survival estimatesbased on survival probabilities observed during 2000e2001,
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e12 11
by sex and site, England and Wales. Available at: http://info.cancerresearchuk.org/prod_consump/groups/cr_common/@nre/@sta/documents/generalcontent/crukmig_1000ast-2774.xls. [accessed 21.02.11].
[7] Kapiteijn E, Marijnen CA, Nagtegaal ID, et al. Preoperativeradiotherapy combined with total mesorectal excision forresectable rectal cancer. N Engl J Med 2001;345:638e646.
[8] Sebag-Montefiore D, Stephens RJ, Steele R, et al. Preoperativeradiotherapy versus selective postoperative chemo-radiotherapy in patients with rectal cancer (MRC CR07 andNCIC-CTG C016): a multicentre, randomised trial. Lancet2009;373:811e820.
[9] Sauer R, Becker H, Hohenberger W, et al. Preoperative versuspostoperative chemoradiotherapy for rectal cancer.N Engl J Med2004;351:1731e1740.
[10] Bosset J, Collette L, Calais G, et al. Chemotherapy withpreoperative radiotherapy in rectal cancer. N Engl J Med2006;355:1114e1123.
[11] Wong RK, Berry S, Spithoff K, et al. Preoperative or post-operative therapy for stage II or III rectal cancer: an updatedpractice guideline. Clin Oncol (R Coll Radiol) 2010;22:265e271.
[12] Roels S, Duthoy W, Haustermans K, et al. Definition anddelineation of the clinical target volume for rectal cancer. Int JRadiat Oncol Biol Phys 2006;65:1129e1142.
[13] Syk E, Torkzad MR, Blomqvist L, Nilsson PJ, Glimelius B. Localrecurrence in rectal cancer: anatomic localization and effecton radiation target. Int J Radiat Oncol Biol Phys 2008;72:658e664.
[14] Nijkamp J, Kusters M, Beets-Tan RG, et al. Three-dimensionalanalysis of recurrence patterns in rectal cancer: the cranialborder in hypofractionated preoperative radiotherapy can belowered. Int J Radiat Oncol Biol Phys 2011;80:103e110.
[15] Myerson RJ, Garofalo MC, El Naqa I, et al. Elective clinicaltarget volumes for conformal therapy in anorectal cancer:a radiation therapy oncology group consensus panel con-touring atlas. Int J Radiat Oncol Biol Phys 2009;74:824e830.
[16] Tait D. Advances in chemoradiation therapy in rectal cancer:the impact of imaging. Br J Radiol 2005;78(Spec no 2):S131eS137.
[17] Cella L, Ciscognetti N, Martin G, et al. Preoperative radiationtreatment for rectal cancer: comparison of target coverageand small bowel NTCP in conventional vs. 3D-conformalplanning. Med Dosim 2009;34:75e81.
[18] Borger J, Bogaard JVD, Haas DD, et al. Evaluation of threedifferent CT simulation and planning procedures for thepreoperative irradiation of operable rectal cancer. RadiotherOncol 2008;87:350e356.
[19] Martin N, Nawaz A, Schuller B, Kozak K, Hong T, Kachnic L.Dosimetric comparison of radiation techniques to the pronepelvis for rectal cancer: 3 field based on bony landmarks (2D)vs 3-dimensional conformal (3D) vs intensity modulatedradiotherapy. Int J Radiat Oncol Biol Phys 2008;72:S548.
[20] Brown G, Davies S, Williams G, et al. Effectiveness of preop-erative staging in rectal cancer: digital rectal examination,endoluminal ultrasound or magnetic resonance imaging? BrJ Cancer 2004;91:23e29.
[21] Khoo VS, Joon DL. New developments in MRI for targetvolume delineation in radiotherapy. Br J Radiol 2006;79(Spec.no 1):S2eS15.
[22] Chao M, Gibbs P, Tjandra J, Darben P, Lim-Joon D, Jones I.Evaluation of the use of computed tomography versusconventional orthogonal X-ray simulation in the treatment ofrectal cancer. Australas Radiol 2005;49:122e126.
[23] Corner C, Khimii F, Tsang YM, Harrison M, Jones RG, Hughes R.Comparison of conventional and 3D conformal CT planning
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
techniques for preoperative chemoradiotherapy for locallyadvanced rectal cancer. Int J Radiat Oncol Biol Phys 2008;72:S547.
[24] Yu S, Farmakidis D, Raouf S. Evaluation in acute toxicity in CTconformal versus simulator planned preoperative chemo-radiotherapy in rectal carcinoma. 12th World Congress onGastrointestinal Cancer. Barcelona: ESMO; 2010.
[25] Myerson R, Drzymala R. Technical aspects of image-basedtreatment planning of rectal carcinoma. Semin Radiat Oncol2003;13:433e440.
[26] Khoo V. MRI-"Magic radiotherapy imaging" for treatmentplanning? Br J Radiol 2000;73:229e233.
[27] Chen L, Price Jr RA, Nguyen TB, et al. Dosimetric evaluation ofMRI-based treatment planning for prostate cancer. Phys MedBiol 2004;49:5157e5170.
[28] The Royal College of Radiologists. Imaging for oncology. Avail-able at: http://www.rcr.ac.uk/docs/oncology/pdf/BFCO(04)2_Imaging_Oncology.pdf. [accessed 09.05.11].
[29] Khoo VS, Dearnaley DP, Finnigan DJ, Padhani A, Tanner SF,Leach MO. Magnetic resonance imaging (MRI): considerationsand applications in radiotherapy treatment planning. Radio-ther Oncol 1997;42:1e15.
[30] Dean C, Sykes J, Cooper R, et al. An evaluation of four CT-MRIco-registration techniques for radiotherapy treatment plan-ning of prone rectal cancer patients. Br J Radiol, in press.
[31] Brussel SV, Deckers F, Laere SV, et al. IMRT for rectal cancerpatients based on diffusion weighted MRI (DWMRI). Int JRadiat Oncol Biol Phys 2008;72:S251.
[32] Sun YS, Zhang XP, Tang L, et al. Locally advanced rectalcarcinoma treated with preoperative chemotherapy andradiation therapy: preliminary analysis of diffusion-weightedMR imaging for early detection of tumor histopathologicdownstaging. Radiology 2010;254:170e178.
[33] Anderson C, Koshy M, Staley C, et al. PET-CT fusion in radia-tion management of patients with anorectal tumors. IntJ Radiat Oncol Biol Phys 2007;69:155e162.
[34] Ciernik IF, Huser M, Burger C, Davis JB, Szekely G. Auto-mated functional image-guided radiation treatment plan-ning for rectal cancer. Int J Radiat Oncol Biol Phys 2005;62:893e900.
[35] Gregoire V, Haustermans K, Geets X, Roels S, Lonneux M. PET-based treatment planning in radiotherapy: a new standard?J Nucl Med 2007;48(Suppl. 1):68Se77S.
[36] Nestle U, Kremp S, Schaefer-Schuler A, et al. Comparison ofdifferent methods for delineation of 18F-FDG PET-positivetissue for target volume definition in radiotherapy ofpatients with non-small cell lung cancer. J Nucl Med 2005;46:1342e1348.
[37] Brunetti J, Caggiano A, Rosenbluth B, Vialotti C. Technicalaspects of positron emission tomography/computed tomog-raphy fusion planning. Semin Nucl Med 2008;38:129e136.
[38] MacManus M, Nestle U, Rosenzweig KE, et al. Use of PET andPET/CT for radiation therapy planning: IAEA expert report2006e2007. Radiother Oncol 2009;91:85e94.
[39] Schinagl DA, Kaanders JH, Oyen WJ. From anatomical to bio-logical target volumes: the role of PET in radiation treatmentplanning. Cancer Imaging 2006;6:S107eS116.
[40] Steenbakkers RJ, Duppen JC, Fitton I, et al. Reduction ofobserver variation using matched CT-PET for lung cancerdelineation: a three-dimensional analysis. Int J Radiat OncolBiol Phys 2006;64:435e448.
[41] Vees H, Senthamizhchelvan S, Miralbell R, Weber DC, Ratib O,Zaidi H. Assessment of various strategies for 18F-FET PET-guided delineation of target volumes in high-grade gliomapatients. Eur J Nucl Med Mol Imaging 2009;36:182e193.
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
S. Gwynne et al. / Clinical Oncology xxx (2011) 1e1212
[42] Black QC, Grills IS, Kestin LL, et al. Defining a radiotherapytarget with positron emission tomography. Int J Radiat OncolBiol Phys 2004;60:1272e1282.
[43] Ford EC, Kinahan PE, Hanlon L, et al. Tumor delineation usingPET in head and neck cancers: threshold contouring andlesion volumes. Med Phys 2006;33:4280e4288.
[44] Erdi YE, Mawlawi O, Larson SM, et al. Segmentation of lunglesion volume by adaptive positron emission tomographyimage thresholding. Cancer 1997;80:2505e2509.
[45] Davis JB, Reiner B, Huser M, Burger C, Szekely G, Ciernik IF.Assessment of 18F PET signals for automatic target volumedefinition in radiotherapy treatment planning. RadiotherOncol 2006;80:43e50.
[46] Biehl KJ, Kong FM, Dehdashti F, et al. 18F-FDG PET definitionof gross tumor volume for radiotherapy of non-small cell lungcancer: is a single standardized uptake value thresholdapproach appropriate? J Nucl Med 2006;47:1808e1812.
[47] Li H, Thorstad WL, Biehl KJ, et al. A novel PET tumor delin-eation method based on adaptive region-growing and dual-front active contours. Med Phys 2008;35:3711e3721.
[48] Geets X, Lee JA, Bol A, LonneuxM, Gregoire V. A gradient-basedmethod for segmenting FDG-PET images: methodology andvalidation. Eur J Nucl Med Mol Imaging 2007;34:1427e1438.
[49] Wanet M, Lee JA, Weynand B, et al. Gradient-based delinea-tion of the primary GTV on FDG-PET in non-small cell lungcancer: a comparison with threshold-based approaches, CTand surgical specimens. Radiother Oncol 2011;98:117e125.
[50] Yaremko B, Riauka T, Robinson D, et al. Thresholding in PETimages of static and moving targets. Phys Med Biol2005;50:5969e5982.
[51] Roels S, Slagmolen P, Nuyts J, et al. Biological image-guidedradiotherapy in rectal cancer: is there a role for FMISO orFLT, next to FDG? Acta Oncol 2008;47:1237e1248.
[52] Abrams R, Winter K, Regine W, et al. RTOG 9704-radiotherapyquality assurance (QA) review and survival. Int J Radiat OncolBiol Phys 2006;66:S22.
[53] Peters L, O’Sullivan B, Giralt J, et al. Critical impact of radio-therapy protocol compliance and quality in the treatment ofadvanced head and neck cancer: results from TROG 02.02. JClin Oncol 2010;28:2996e3001.
[54] Villeirs GM, Van Vaerenbergh K, Vakaet L, et al. Interobserverdelineation variation using CT versus combined CT þ MRI inintensity-modulated radiotherapy for prostate cancer. Strah-lenther Onkol 2005;181:424e430.
[55] Buijsen J, Bogaard JVD. PET-CT can reliably define the tumourdimensions of rectal cancer. Joint ECCO 15 e 34th ESMOMultidisciplinary Congress, 2009. Eur J Cancer 2009:168e169.Suppl. 7.
[56] YavuzM, PehlivanB, AydinM, et al. Comparison of CT-guided andPET-CT guided radiotherapy planning in patients with rectumcancer treated preoperatively. Joint ECCO 15 e 34th ESMOMultidisciplinary Congress, 2009. Conference Publication 7: 329.
[57] Betler J, Day E, Kirichenko A, et al. Evaluation of FDG-PET/CTimaging in treatment planning for distal rectal (DR) and analcanal (AC) carcinomas. Int J Radiat Oncol Biol Phys 2008;72:S258.
[58] Oxford Centre of Evidence Based Medicine. Levels ofevidence. Available at: http://www.cebm.net/index.aspx?o¼1025. [accessed 09.05.11].
[59] Krengli M, Cannillo B, Turri L, et al. Target volume delineationfor preoperative radiotherapy or rectal cancer: inter-observervariability and potential impact of FDG-PET/CT imaging.Technol Cancer Res Treat 2010;9:393e398.
Please cite this article in press as: Gwynne S, et al., Imaging for TargetReview, Clinical Oncology (2011), doi:10.1016/j.clon.2011.10.001
[60] O’Neill BD, Salerno G, Thomas K, Tait DM, Brown G. MR vs CTimaging: low rectal cancer tumour delineation forthree-dimensional conformal radiotherapy. Br J Radiol 2009;82:509e513.
[61] Tan J, Lim Joon D, Fitt G, et al. The utility of multimodalityimaging with CT and MRI in defining rectal tumour volumesfor radiotherapy treatment planning: a pilot study. J MedImaging Radiat Oncol 2010;54:562e568.
[62] Seierstad T, Hole KH, Saelen E, Ree AH, Flatmark K, Malinen E.MR-guided simultaneous integrated boost in preoperativeradiotherapy of locally advanced rectal cancer following neo-adjuvant chemotherapy. Radiother Oncol 2009;93:279e284.
[63] Patel D, Chang S, Goodman K, et al. Impact of integrated PET/CT on variability of target volume delineation in rectal cancer.Technol Cancer Res Treat 2007;6:31e35.
[64] Bassi M, Turri L, Sacchetti G, et al. FDG-PET/CT imaging forstaging and target volume delineation in preoperativeconformal radiotherapy of rectal cancer. Int J Radiat Oncol BiolPhys 2008;70:1423e1426.
[65] Paskeviciute B, Bolling T, Brinkmann M, et al. Impact of FDG-PET/CT on staging and irradiation of patients with locallyadvanced rectal cancer. Strahlenther Onkol 2009;185:260e265.
[66] Day E, Betler J, Parda D, Mohammadi S, Miften M. A regiongrowingmethod for tumor volume segmentation on PET imagesforrectalandanalcancerpatients.MedPhys2009;36:4349e4358.
[67] Fuller CD, Nijkamp J, Duppen JC, et al. Prospective randomizeddouble-blind pilot study of site-specific consensus atlasimplementation for rectal cancer target volume delineation inthe cooperative group setting. Int J Radiat Oncol Biol Phys2011;79:481e489.
[68] Vorwerk H, Liersch T, Rothe H, et al. Gold markers for tumorlocalization and target volume delineation in radiotherapy forrectal cancer. Strahlenther Onkol 2009;185:127e133.
[69] Myerson R, Garofalo M, Naqa I, et al. Elective clinical targetvolumes in anorectal cancer: an RTOG consensus panel con-touring atlas. Available at: http://www.rtog.org/pdf_file2.html?pdf_document¼AnorectalContouringGuidelines.pdf. [accessed24.02.11].
[70] NCRI. A framework for PET research in theUK: Report of theNCRIPET strategic planning group. Available at: http://www.ncri.org.uk/includes/What_We_Do/researchnetworks/documents/NCRIFrameworkforPETResearch2007_000.pdf. [accessed 24.02.11].
[71] Khoo VS, Padhani AR, Tanner SF, Finnigan DJ, Leach MO,Dearnaley DP. Comparison of MRI with CT for the radio-therapy planning of prostate cancer: a feasibility study. Br JRadiol 1999;72:590e597.
[72] Boellaard R, O’Doherty MJ, Weber WA, et al. FDG PET and PET/CT: EANM procedure guidelines for tumour PET imaging:version 1.0. Eur J Nucl Med Mol Imaging 2010;37:181e200.
[73] Chen L, Paskalev K, Xu X, et al. Rectal dose variation duringthe course of image-guided radiation therapy of prostatecancer. Radiother Oncol 2010;95:198e202.
[74] Ma DJ, Zhu JM, Grigsby PW. Tumor volume discrepanciesbetween FDG-PET and MRI for cervical cancer. Radiother Oncol2011;98:139e142.
[75] Gwynne S, Webster R, Mukherjee S, Staffurth J, Spezi E,Adams R. Image-guided radiotherapy for rectal cancer e
a systematic review. Clin Oncol (R Coll Radiol) 2011. doi:10.1016/j.clon.2011.07.012.
[76] YavuzMN,TopkanE, YavuzAA, et al. FDG-PET/CT imaging-basedtarget volume delineation for preoperative conformal radio-therapy of rectal carcinoma. Int J Haem Oncol 2010;20:67e74.
Volume Delineation in Rectal Cancer Radiotherapy d A Systematic
![[From image-guided radiotherapy to dose-guided radiotherapy]](https://static.documents.pub/doc/80x56/63500d830ca35926a709599d/from-image-guided-radiotherapy-to-dose-guided-radiotherapy.jpg)
