Medical Dosimetry

Medical Dosimetry 38 (2013) 366–371

0958-39http://d

RepriOncologChung-S

E-m

journal homepage: www.meddos.org

Superior liver sparing by combined coplanar/noncoplanar volumetric-modulated arc therapy for hepatocellular carcinoma: A planning andfeasibility study

Yi-Chun Tsai, M.S.,* Chiao-Ling Tsai, M.D.,* Feng-Ming Hsu, M.D.,* Jian-Kuen Wu, M.S.,*║

Chien-Jang Wu, Ph.D.,║ and Jason Chia-Hsien Cheng, M.D., Ph.D.*†‡§

*Division of Radiation Oncology, Department of Oncology, National Taiwan University Hospital, Taipei, Taiwan; †Graduate Institute of Oncology, National Taiwan University Collegeof Medicine, Taipei, Taiwan; ‡Graduate Institute of Clinical Medicine, National Taiwan University College of Medicine, Taipei, Taiwan; §Cancer Research Center, National TaiwanUniversity College of Medicine, Taipei, Taiwan; and ║Institute of Electro-Optical Science and Technology, National Taiwan Normal University, Taipei, Taiwan

A R T I C L E I N F O

Article history:Received 16 October 2012Accepted 4 April 2013

Keywords:Volumetric-modulated arc therapyHepatocellular carcinomaLiverNoncoplanar

47/$– see front matter Copyright � 2013 Amx.doi.org/10.1016/j.meddos.2013.04.003

nt requests to: Jason Chia-Hsien Cheng, M.Dy, Department of Oncology, National Taiwahan South Road, Taipei 10002, Taiwan.ail: jasoncheng@ntu.edu.tw

A B S T R A C T

Compared with step-and-shoot intensity-modulated radiotherapy (sIMRT) and tomotherapy, volumetric-modulated arc therapy (VMAT) allows additional arc configurations in treatment planning and non-coplanar (NC) delivery. This study was first to compare VMAT planning with sIMRT planning, and thesecond to evaluate the toxicity of coplanar (C)/NC-VMAT treatment in patients with hepatocellularcarcinoma (HCC). Fifteen patients with HCC (7 with left-lobe and 8 with right-lobe tumors) were plannedwith C-VMAT, C/NC-VMAT, and sIMRT. The median total dose was 49 Gy (range: 40 to 56 Gy), whereasthe median fractional dose was 3.5 Gy (range: 3 to 8 Gy). Different doses/fractionations were convertedto normalized doses of 2 Gy per fraction using an α/β ratio of 2.5. The mean liver dose, volume fractionreceiving more than 10 Gy (V10), 20 Gy (V20), 30 Gy (V30), effective volume (Veff), and equivalentuniform dose (EUD) were compared. C/NC-VMAT in 6 patients was evaluated for delivery accuracy andtreatment-related toxicity. Compared with sIMRT, both C-VMAT (p ¼ 0.001) and C/NC-VMAT (p ¼ 0.03)had significantly improved target conformity index. Compared with C-VMAT and sIMRT, C/NC-VMAT fortreating left-lobe tumors provided significantly better liver sparing as evidenced by differences in meanliver dose (p ¼ 0.03 and p ¼ 0.007), V10 (p ¼ 0.003 and p ¼ 0.009), V20 (p ¼ 0.006 and p ¼ 0.01), V30 (p¼ 0.02 and p ¼ 0.002), Veff (p ¼ 0.006 and p ¼ 0.001), and EUD (p ¼ 0.04 and p ¼ 0.003), respectively. Forright-lobe tumors, there was no difference in liver sparing between C/NC-VMAT, C-VMAT, and sIMRT. Inall patients, dose to more than 95% of target points met the 3%/3 mm criteria. All 6 patients toleratedC/NC-VMAT and none of them had treatment-related ≥ grade 2 toxicity. The C/NC-VMAT can be usedclinically for HCC and provides significantly better liver sparing in patients with left-lobe tumors.

& 2013 American Association of Medical Dosimetrists.

Introduction

Even with the inclusion of radiation therapy (RT) in the multi-modality treatment for hepatocellular carcinoma (HCC), the tumorresponse and disease outcome remain unsatisfactory.1-3 Insuffi-cient dose of radiation delivered to the target hepatic tumor(s) andheightened sensitivity of the diseased liver to high-dose radiationhave been the main barriers to achieving satisfactory results.4-6

erican Association of Medical Dos

., Ph.D., Division of Radiationn University Hospital, No. 7,

Recent efforts to overcome these barriers include development ofhigher-intensity stereotactic body RT (SBRT) and improvement inthe knowledge of radiation-induced liver disease (RILD).7-9 Advan-ces in treatment delivery, including intensity-modulated RT(IMRT) and respiratory control, make dose escalation possibleand liver sparing achievable.

The clinically adopted technique known as volumetric-modulated arc therapy (VMAT) improves target conformity andorgan sparing by use of rotational IMRT and more control pointsfor intensity optimization.10-12 Compared with helical tomotherapy(an earlier arc IMRT technique), VMAT can deliver additional non-coplanar arcs of radiation.13 The liver is situated asymmetrically inthe upper abdomen, and is susceptible to damage from even low

imetrists

Table 1Patient characteristics and dose fractionations of 15 patients with hepatocellularcarcinoma normalized to 2-Gy fraction size using the α/β ratio of 2.5

Patientnumber

Tumorlobe

GTV(mL)

Liver volume(mL)

Normalized dose(Gy)

1 Left 704 1753 104.02 Left 266 1223 56.03 Left 144 1812 92.04 Left 65 1182 70.05 Left 221 1653 44.46 Left 114 1148 74.77 Left 5 673 56.08 Right 91 1292 70.09 Right 22 1558 70.0

10 Right 11 850 65.311 Right 75 1054 74.712 Right 14 948 70.013 Right 42 1219 65.314 Right 128 1525 65.315 Right 263 800 65.3

Y-C. Tsai et al. / Medical Dosimetry 38 (2013) 366–371 367

doses of radiation. Left lobe of liver is more ventrally and super-ficially located than right lobe. Noncoplanar VMAT may have theadvantage of taking the shorter path in the liver and further confinethe radiation to a well-defined region of the liver. For the certainranges of noncoplanar beam paths, the freedom is limited owing tothe possible mechanical collision of gantry head and couch base.Therefore, it is technically not possible to have the full-arc non-coplanar VMAT. The target conformity may be compromised withonly the asymmetric partial-arc noncoplanar VMAT design. Thecomplementary beam design with the coplanar VMAT compensatesthis defect. In this study, we used the combined coplanar/non-coplanar VMAT, rather than sole noncoplanar VMAT, to demonstrateits dosimetric advantage in sparing liver and maintaining targetconformity. The purpose of this study was first to compare the step-and-shoot IMRT (sIMRT), coplanar VMAT (C-VMAT), the combinedcoplanar/noncoplanar VMAT (C/NC-VMAT) treatment plans of 15patients with HCC. In the second part of the study, we used theVMAT treatment plan that provided comparable or better targetconformity (conformity index [CI]) and liver sparing (proportion ofliver receiving certain doses, effective volume [Veff], and equivalentuniform dose [EUD]) to treat 6 of the 15 patients. The dose deliveryaccuracy and treatment-related toxicity were evaluated in thesepatients.

Methods and Materials

Patient characteristics

Fifteen patients with HCC (13 males and 2 females; median age, 61 years[range: 40 to 82]) who were not candidates for conventional therapies (surgery,transcatheter arterial chemoembolization, percutaneous ethanol injection therapy,or radiofrequency ablation) and not eligible for or unresponsive to sorafenib,underwent RT for localized liver tumors from January 2008 to March 2011, and thedata from these patients were retrospectively analyzed. HCC was Barcelona ClinicLiver Cancer Stage A in 1 patient, Stage B in 6, and Stage C in 8. Only 1 patient withStage A HCC, had a recurrence near the diaphragm after chemoembolization anddeclined the recommended surgery owing to his old age (82 years). Twelve patientswere diagnosed serologically with type B chronic viral infection, whereas 2 patientswere with type C. All 15 patients were classified as having Child-Pugh class Acirrhosis of the liver, and also underwent transcatheter arterial chemoembolizationor radiofrequency ablation before RT or both.

The median gross tumor volume (GTV) and liver volume for 15 patients were91 mL (range: 5 to 705 mL) and 1219 mL (range: 673 to 1812 mL), respectively. Themedian ratio (GTV divided by liver volume) was 0.07 (range: 0.01 to 0.40).Radiation dose was dependent on the volume of the liver irradiated as determinedby the effective volume method and the estimated risk of liver toxicity.14

Patient setup and immobilization

Each patient was placed in the supine position and immobilized with aspecially designed evacuated vacuum bag. To reduce respiratory movement, anactive breathing coordinator device, body-fix device, and body-frame device wereused in 7 patients, 5 patients, and 1 patient, respectively. The goal of respiratorycontrol is to reduce the craniocaudal diaphragmatic motion amplitude to be 5 mmor less on fluoroscopy. Besides the original plan for the actual treatment, 3-dimensional computed tomography datasets were used to design sIMRT, C-VMAT,and C/NC-VMAT treatment plans with the same goals and constraints for eachpatient.

Dosimetric plans

GTV was defined as the gross tumor volume, visualized by 3-dimensionalcomputation of contrast-enhanced computed tomography–defined gross tumorcontours. Clinical target volume was defined as the GTV plus a 0.5-cm margin.Planning target volume (PTV) was defined as the clinical target volume plus a 0.5-cm margin on the medial/lateral/ventral/dorsal sides, as well as a 0.5 to 1.0 cmmargin on the cranial/caudal sides to account for daily setup error and organmotion due to respiration. The median original prescribed dose was 49 Gy (range:40 to 56 Gy in 6 to 16 fractions), whereas the median dose per fraction was 3.5 Gy(range: 3 to 8 Gy). All the prescribed doses normalized to the fraction size of 2 Gyusing an α/β ratio of 2.5 are shown in Table 1. The normalized doses formed thebasis of comparisons between techniques and patients. All 3 plans had comparabletarget coverage for each patient (4 95% PTV covered by 4 93% of the prescribed

dose in 3 patients and by 4 95% in 12 patients). The volume of normal liver wascalculated by subtracting the GTV from the volume of whole liver.

The beam setups of each individual patient in this study were to find thebest achievable designs of the coplanar or noncoplanar beam combination orboth for the 3 techniques. sIMRT plans were designed using the Pinnacle3

treatment planning system version 9.0 (ADAC Laboratories, Philips MedicalSystems, Milpitas, CA) by the Direct Machine Parameter Optimization algo-rithm. The minimum segment area was set to 5 cm2, and the minimum segmentmonitor unit (MU) was 5 MU. The VMAT plans were designed by the SmartArcmodule of Pinnacle3 system for an Elekta Synergy linear accelerator (ElektaOncology System Ltd., Crawley, West Sussex, UK), allowing the use of a binnedvariable dose rate. Continuous gantry motion, dose-rate variation, and multi-leaf collimator motion were approximated by optimizing individual beams at 41gantry angle increments with multileaf collimator leaf positions varying by upto 4.6 mm for every 11 of gantry rotation. Elekta VMAT delivery was basically byMU-based servo control. The accelerator used automatic dose-rate selectionthat ensures that the maximal possible dose rate was chosen for eachindividual segment of the arc. The possible dose rates were 440, 222, 112,and 57 MU/min. Dose grid resolution was 0.4 cm for the inverse planning.Pinnacle plans were transferred in Radiotherapy Treatment Plans exportprotocols through MOSAIQ version 1.6 (IMPAC Medical Systems, Inc., Sunny-vale, CA) to the linear accelerator. Helical Tomotherapy was not used in thiscomparison study, for its limited couch rotation and the small-scale non-coplanar VMAT.

All patients were treated with 10-MV photon beams in supine position. Thebeams were directed toward the tumor along paths through the smallest livervolume to minimize the amount of normal liver exposed to even low doses ofradiation. The quantitative estimation was conducted in the process of selecting thebest noncoplanar beam combinations. The maximum dimensions of PTV were usedto design an open field for the volume estimation of irradiated liver by variousnoncoplanar beam combinations, both for sIMRT and for VMAT plans. In general,the commonly used beams were from right, cranial, and ventral to left, caudal, anddorsal directions. For sIMRT plans, the median number of beams used was 6 (range:5 to 8), which included both coplanar and noncoplanar beams. The median numberof segments per beam was 7 (range: 2 to 9). For C-VMAT plans, planning wasperformed using 2 clockwise-counterclockwise coplanar partial-arc beams, sur-rounding the involved lobe of liver, in all patients. For C/NC-VMAT plans, 2clockwise-counterclockwise coplanar partial arcs and 2 ventral clockwise-counterclockwise noncoplanar arcs surrounding the involved lobe of liver wereused in 12 patients, only ventral clockwise-counterclockwise noncoplanar partialarcs were used in 2 patients, and 1 coplanar partial-arc plus 4 ventral clockwise-counterclockwise noncoplanar arcs were used in 1 patient.

Doses were prescribed to a peripheral covering isodose covering the PTV.Assuming dose was normalized to this isodose at 100%, the maximal dose can be120% and the minimum PTV dose 90%. Any dose 4 110% must be within the PTV.The normal liver was defined as the normal liver volume minus GTV. In allpatients, it was required that there is at least 700 cc of normal liver, and thisvolume received less than 15 Gy. No more than 30% of the normal liver receivedmore than 27 Gy, and no more than 50% of normal liver received over 24 Gy. Forkidneys, no more than 50% of the combined renal volume received 20 Gy or more.Maximal permitted dose to spinal cord was 37 Gy. Maximal permitted dose tosmall bowel, duodenum, and stomach was 42 Gy for any 3 cc volume. Whenplanning IMRT or VMAT, the iterations were based on both the target goals anddose constraints of critical structures, with the liver constraints as the priority.Therefore, the iterations would continue for the better target conformity if liverdose-volume data allow.

Y-C. Tsai et al. / Medical Dosimetry 38 (2013) 366–371368

Conformity index of target and dose to normal liver

Several dose-to-target and dose-to-normal tissue indexes were calculated fromdose-volume histograms (DVHs), including Vx, conformity index (CI), effectivevolume (Veff), and EUD. Vx was the volume fraction receiving more than a certaindose (x Gy). CI was the ratio of the prescription volume to the target volume, asdefined in the Radiation Therapy Oncology Group (RTOG) radiosurgery guide-lines.15 The CI was defined as follows:

CI¼ VPTV=TVPV

TVPV=VTV¼ VPTV � VTV

TVPV2,

where VTV was the treatment volume covered by the prescribed dose, VPTV was thePTV, and TVPV was the part of the PTV within the VTV.16 The effective volume (Veff)method from Kutcher and Burman17 was used to estimate the equivalent dose andvolume pairs for uniform partial organ irradiation from the DVHs summarizing thenonuniform irradiation. The Veff was defined as follows:

ΔVef f ,i¼ViDi

Dref

� �1n

Vef f ¼∑i

ΔVef f ,i

where Di was each dose level, Vi was the volume fraction of the organ receiving thedose Di, Dref was the maximum dose delivered to the organ, and n (¼ 0.97) was thevolume-effect parameter.14 The concept of EUD was used to determine the uniformdose for any given nonuniform dose distribution that gives the same biologicaleffect. The EUD was defined as follows:18,19

EUD¼ ∑i

V iðDiÞ1=n" #n

, or

EUD¼Dref ðVef f Þn

Statistical analysis

The descriptive statistics of target conformity and organs at risk (OARs) werecompared between patients receiving sIMRT, C-VMAT, and C/NC-VMAT. A pairedStudent t-test was used to evaluate the significance of differences in CI of target,mean liver dose, V10, V20, V30, Veff, EUD for liver minus GTV, and the absolutevolume or volume fraction receiving specific doses for individual organs. p Valueo 0.05 was defined as statistically significant. SPSS software release 17.00 (SPSS,Inc., Chicago, IL) was used to calculate the statistics.

Dose verification of VMAT

Doses were verified using the MapCHECK 2 device version 5.02.00.02 (SunNuclear Corporation, Melbourne, FL). The differences between the planned andmeasured doses were analyzed by gamma tests. The criteria of gamma evaluation

Table 2Parameter comparison of liver minus gross tumor volume (Liver−GTV) between stepmodulated arc therapy (C-VMAT), and combined coplanar/noncoplanar VMAT (C/NC-VMlobe tumors

Tumor location(no.)

Liver-GTVparameter Mean � standard deviation

sIMRT C-VMAT C/NC-V

All tumors(n ¼ 15)

MLD (Gy) 19.7 � 7.5 18.9 � 8.8 17.5 �

V10 (%) 41.1 � 13.4 36.4 � 13.0 33.7 �

V20 (%) 30.2 � 9.9 27.7 � 12.2 25.3 �

V30 (%) 24.7 � 9.9 23.1 � 11.9 21.0 �

Veff (%) 24.3 � 7.1 23.3 � 7.9 21.5 �

EUD (Gy) 19.8 � 7.5 19.2 � 8.6 17.8 �

Left-lobe tumor(n ¼ 7)

MLD (Gy) 20.5 � 9.4 20.0 � 10.4 17. 6 �

V10 (%) 41.3 � 11.8 37.7 � 12.3 32.0 �

V20 (%) 30.8 � 11.5 28.5 � 12.6 24.0 �

V30 (%) 24.1 � 11.7 23.5 � 13.1 19.8 �

Veff (%) 24.7 � 7.9 23.6 � 8.1 20.6 �

Right-lobe tumor(n ¼ 8)

EUD (Gy) 20.4 � 9.4 20.2 � 10.5 17.8 �

MLD (Gy) 19.0 � 6.0 17.9 � 7.7 17.3 �

V10 (%) 40.9 � 15.4 35.3 � 14.2 35.2 �

V20 (%) 29.6 � 9.1 27.0 � 12.7 26.6 �

V30 (%) 25.2 � 8.7 22.8 � 11.7 22.1 �

Veff (%) 24.0 � 6.9 23.1 � 8.3 22.2 �

EUD (Gy) 19.3 � 6.0 18.4 � 7.2 17.8 �

MLD ¼ mean liver dose; Vx ¼ volume fraction receiving more than certain dose (x Gy

were 3% dose difference and 3-mm distance to agreement. Γ ≤ 1 was defined as theverification passing the criteria and satisfying at least 95% of points.

Results

CI of target

C-VMAT, C/NC-VMAT, and sIMRT plans were able to meet thetarget conformity goal and the dose-volume constraints for OARs.The mean CI of PTV was 1.4 � 0.2 for C-VMAT, 1.5 � 0.4 for C/NC-VMAT, and 1.7 � 0.4 for sIMRT. The CI was significantly lower forboth C-VMAT (p ¼ 0.001) and C/NC-VMAT (p ¼ 0.03) than forsIMRT in all patients. Compared with sIMRT, both C-VMAT (1.5 �

0.2 vs 1.9 � 0.4, p ¼ 0.02) and C/NC-VMAT (1.5 � 0.4 vs 1.9 � 0.4,p ¼ 0.03) achieved superior conformity for right-lobe tumors.Compared with sIMRT, C-VMAT (1.3 � 0.1 vs 1.5 � 0.2, p ¼ 0.02)but not C/NC-VMAT (1.4 � 0.4 vs 1.5 � 0.2, p ¼ 0.64) achievedsuperior conformity for left-lobe tumors. The difference in CI foreither lobe was not statistically significant between C-VMAT andC/NC-VMAT (p4 0.05). The meanMUs of sIMRT, C-VMAT, and C/NC-VMAT were 967 � 612, 1037 � 428, and 1007 � 463, respectively.

OARs sparing

Differences in dose sparing of the functional liver (liver minusGTV) were assessed among the 3 techniques (Table 2). Liversparing by C/NC-VMAT was significantly more effective in allparameters including mean liver dose, V10, V20, V30, Veff, andEUD than that by C-VMAT and sIMRT for left-lobe tumors (p o0.05) but not right-lobe tumors (p ≥ 0.05). In contrast, bothC-VMAT and sIMRT for either left-lobe or right-lobe tumorsachieved similar liver-sparing results (p ≥ 0.05). The mean DVHby the normalized dose from 3 techniques (sIMRT, C-VMAT, andC/NC-VMAT) for PTV and liver of all 15 patients was shown in Fig. 1.

For patients with left-lobe tumors, the correlations of thetumor size with the liver-sparing advantage between techniquesrevealed the significant inverse correlations in mean liver dose ofC/NC-VMAT over C-VMAT (correlation coefficient ¼ −0.79, p ¼0.03) and in EUD of C/NC-VMAT over C-VMAT (correlation coef-ficient ¼ −0.80, p ¼ 0.03), as well as insignificant correlations inthe other parameters between techniques (p 4 0.05).

-and-shoot intensity-modulated radiation therapy (sIMRT), coplanar volumetric-AT) for all 15 patients, 7 patients with left-lobe tumors, and 8 patients with right-

p Value

MAT C/NC-VMAT vs C-VMAT C/NC-VMAT vs sIMRT C-VMAT vs sIMRT

7.8 0.02 0.002 0.2310.9 0.03 0.04 0.2110.3 0.02 0.002 0.1010.2 0.02 o0.001 0.126.7 0.007 0.004 0.337.6 0.02 0.004 0.429.5 0.03 0.007 0.4310.7 0.003 0.009 0.1111.0 0.006 0.01 0.2611.3 0.02 0.002 0.537.1 0.006 0.001 0.319.6 0.04 0.003 0.746.6 0.31 0.10 0.3611.6 0.93 0.37 0.4210.3 0.68 0.11 0.269.9 0.45 0.05 0.186.7 0.31 0.24 0.606.1 0.32 0.17 0.49

).

Table 3The treatment-related toxicity in 6 patients treated by volumetric-modulated arctherapy

Toxicity Grade 0 Grade I Grade II Grade III Grade IV

Liver 4 2 0 0 0Blood 2 0 1 3 0Coagulation 5 0 1 0 0Skin 4 2 0 0 0Anorexia 3 3 0 0 0Esophagitis 5 0 1 0 0Gastritis 5 1 0 0 0Diarrhea 6 0 0 0 0

Y-C. Tsai et al. / Medical Dosimetry 38 (2013) 366–371 369

For other OARs (lung, stomach, spinal cord, duodenum, bowels,and kidneys), there were no significant dose differences betweenC/NC-VMAT, C-VMMAT, and sIMRT plans (data not shown).

Dose-escalation potentials by C/NC-VMAT than sIMRT

To estimate the dose-escalation potentials by C/NC-VMAT, wecalculated the escalated prescribed dose in the original C/NC-VMAT plans but kept the same mean liver dose as that in theoriginal sIMRT plans for patients with left-lobe tumors. As a resultof the advantageous liver sparing by C/NC-VMAT, the meanescalated prescribed dose normalized to the fraction size of 2 Gy

Fig. 1. The mean dose-volume histogram by the normalized dose from 3 techniques (sIminus gross tumor volume (Liver−GTV) of all 15 patients.

was 84.5� 26.7 Gy compared with the original 71.0 � 21.2 Gy (p ¼0.016). The mean dose escalation ratio was 19.4%.

Dose verification and treatment by VMAT

The gamma criteria by 3 measurements for each patient (3%/3 mm for more than 95% of dose points [mean: 97.7 � 2.2%; range:95% to 100%]) were fulfilled in all 6 patients actually treated withC/NC-VMAT.

Treatment toxicity after C/NC-VMAT

All 6 patients tolerated C/NC-VMAT, and none of them hadtreatment-related toxicities greater than grade 2 or RILD within3 months after completion of treatment, based on CommonToxicity Criteria version 4.0. The exceptions were 3 patients withpreexisting grade III hematological toxicity due to hypersplenismbefore VMAT. Two had low platelet counts and 1 had lowleukocyte counts. All 3 patients had blood counts that were noworse during and after VMAT. The most common side effectsduring VMAT and within 3 months after completion of VMAT weregrade I anorexia (3 patients), grade I dermatitis (2 patients), andgrade I increase in liver enzymes (2 patients), respectively(Table 3). The average dose delivery time was 382 � 109 seconds(range: 243 to 545 seconds).

MRT, C-VMAT, and C/NC-VMAT) for (A) planning target volume (PTV) and (B) liver

Fig. 2. Dose distributions, beam illustration, and dose-volume histogram (DVH) of liver minus the gross tumor volume between (A) step-and-shoot intensity-modulatedradiation therapy (sIMRT) (thin solid line on DVH), (B) coplanar volumetric-modulated arc therapy (C-VMAT) (thin dashed line on DVH), (C) combined coplanar/noncoplanarVMAT (C/NC-VMAT) (thick solid line on DVH), and (D) DVH of liver minus gross tumor volume (Liver−GTV) for a representative patient (no. 3 in Table 1) with left-lobe tumorof hepatocellular carcinoma.

Y-C. Tsai et al. / Medical Dosimetry 38 (2013) 366–371370

The dose distributions and DVH of liver minus the GTV for arepresentative patient with HCC (no. 3 in Table 1) with a left-lobetumor planned by sIMRT, C-VMAT, and C/NC-VMAT techniques areshown in Fig. 2.

Discussion

In this study, CI of the treated tumor was superior whendelineated by the VMAT technique than by the sIMRT techniquein 15 patients with HCC. More importantly, liver sparing by C/NC-VMAT, C-VMAT, and sIMRT was significantly different only for lefthepatic lobe tumors, based on assessment of several parameters(mean liver dose, V10, V20, V30, Veff, and EUD). Compared withC-VMAT and sIMRT, C/NC-VMAT significantly improved liver spar-ing. For patients with right-lobe tumor, both VMAT designs andsIMRT plans had similar characteristics of dose delivery to thetarget and liver. Additionally, 6 patients were actually treated by C/NC-VMAT with verification of satisfactory dose distribution, andnone of them had excessive hepatic toxicity. To our knowledge,ours is the first report proposing the use of the VMAT technique,especially VMAT with the integrated noncoplanar partial-arcdesign, for patients with HCC.

RT is one of the modalities used in multidisciplinaryapproaches for treating HCC.4,20 The dose to the hepatic tumorcan be suboptimal or if too high may increase the risk of RILD.SBRT has the advantages of delivering higher dose intensity infewer fractions of larger fractional doses and of lowering risk ofRILD by protecting cytokine production from the effects of simul-taneous radiation exposure.21 Several parameters, such as volumefraction of liver receiving certain doses, mean liver dose, Veff, and

EUD, have been shown to be effective in predicting RILD.22-24

Dose-escalation trials also use parameter-driven criteria for doseselection.7

VMAT has been increasingly used in a variety of disease sites.The benefits of VMAT are improved target conformity, reducedMUs (partly by 10-MV photon energy), and shorter delivery timeespecially in patients undergoing SBRT.10 Liver malignancies arenow commonly treated by SBRT,7 potentially with advantages byVMAT. Besides, liver has 2 differently shaped lobes and anasymmetric position in the upper abdomen. Anatomically, the lefthepatic lobe is more superficially located than right lobe, and ismore exposed to noncoplanar beams from the ventral side. Thisadvantage is supported by data showing that liver sparing byC/NC-VMAT is better in patients with left-lobe tumors.

It is of note that the target conformity (CI) of HCC was not assatisfactory as other disease sites. For most Asian HCC patientswith chronic viral hepatitis and preexisting cirrhosis of liver, it isimportant to spare noncancerous liver from even low doses of RT.Thus, the beam/arc design may be constrained by limited pathswith the smallest volume of liver exposed to radiation. The largerCI of HCC is likely from this unique situation. Given only thelimited ranges of noncoplanar beam paths, the compromisedfreedom to avoid the possible mechanical collision makes the CIof C/NC-VMAT design acceptably larger than C-VMAT in this study.

Either VMAT or helical tomotherapy (both arc-basedapproaches) with more gantry angles than sIMRT can be used fordose planning. The unique use of noncoplanar arcs by VMAT isdifferent from tomotherapy (which usually has the highest possi-ble modulation depth) in modulating dose distribution, and makespossible the delivery of noncoplanar arcs by changing couchpositions. Liver is more susceptible to partial-arc plans because

Y-C. Tsai et al. / Medical Dosimetry 38 (2013) 366–371 371

of its asymmetric shape, eccentric location, and relationship to theadjacent duodenum and colon at the hepatic flexure. Therefore,noncoplanar arcs from the right cranial direction may focus moreof the radiation on the left hepatic lobe, but spare the dorsally andcaudally seated right hepatic lobe and bowels from high doses. Thesimple addition of noncoplanar beams into sIMRT might not be asgood in target conformity as arc therapy for the fewer beams bysIMRT than VMAT. Our evidence shows that C-VMAT or C/NC-VMAT, compared with sIMRT, achieves superior target conformityexclusively with partial arcs and significantly better liver sparing inpatients with left-lobe tumors.

There were a few limitations in this study. Some commerciallyavailable planning systems, including the one used in this study, donot allow partial-arc plans to be performed with less than 901 ofgantry rotation. Consequently, the selected partial-arc design,especially one involving the use of integrated noncoplanar beams,might produce suboptimal results. Besides, the current 2-dimensional quality assurance systems have certain directionallimits on accurate dose measurement, which might result ininconsistency of noncoplanar plans and require the modifiedmeasurement by collapsing the gantry/couch angle to 0 or attach-ing the array to the gantry. However, such a modified measurementis not the same dynamic arc delivery as VMAT. Besides, the smallerdose grid than 4 mm might have finer resolution and potentiallygenerate different results. Also the available VMAT system cannotbe combined with an active respiratory control device. The inevi-table respiratory motion was controlled exclusively by passiveabdominal compression in our patients undergoing VMAT. Theongoing development of gated VMAT or smaller arc VMAT possiblyintegrated with an active breathing coordinator may soon providethe potential solution. The actual treatment time (including thetime required for MU delivery, the setup procedure, and the imageguidance process) was not compared between techniques. Lastlythe limited number of patients in this study might be associatedwith selection bias and confound the comparison.

In conclusion, compared with sIMRT, VMAT provided superiortarget conformity in patients treated for HCC. C/NC-VMAT providedsignificantly better liver sparing than did C-VMAT and sIMRT inpatients with left-lobe tumors but not right-lobe tumors. Treatmentof HCC with C/NC-VMAT is feasible, accurate, and tolerable.

Acknowledgments

This work was supported by the research grants from NationalScience Council, Execute Yuan (NSC 99-2628-B-002-071-MY3 and99-2314-B-002-111), and National Taiwan University Hospitalgrants NTUH 99S1361, 99N1425, 100S1552, and Liver DiseasePrevention & Treatment Research Foundation, Taiwan, ROC.

References

1. Cheng, J.C.H.; Chuang, V.P.; Cheng, S.H.; et al. Local radiotherapy with orwithout transcatheter arterial chemoemboliziation for patients with unresect-able hepatocellular carcinoma. Int. J. Radiat. Oncol. Biol. Phys. 47:435–42; 2000.

2. Sugahara, S.; Nakayama, H.; Fukuda, K.; et al. Proton-beam therapy forhepatocellular carcinoma associated with portal vein tumor thrombosis.Strahlenther. Onkol. 185:782–8; 2009.

3. Zhao, J.D.; Liu, J.; Ren, Z.G.; et al. Maintenance of sorafenib following combinedtherapy of three-dimensional conformal radiation therapy/intensity-modulatedradiation therapy and transcatheter arterial chemoembolization in patientswith locally advanced hepatocellular carcinoma: a phase I/II study. Radiat.Oncol. 5:12; 2010.

4. Krishnan, S.; Dawson, L.A.; Seong, J.; et al. Radiotherapy for hepatocellularcarcinoma: an overview. Ann. Surg. Oncol. 15:1015–24; 2008.

5. Hata, M.; Tokuuye, K.; Sugahara, S.; et al. Proton irradiation in a single fractionfor hepatocellular carcinoma patients with uncontrollable ascites. Strahlenther.Onkol. 183:411–6; 2007.

6. Hata, M.; Tokuuye, K.; Sugahara, S.; et al. Proton beam therapy for hepatocel-lular carcinoma patients with severe cirrhosis. Strahlenther. Onkol. 182:713–20;2006.

7. Dawson, L.A.; Tse, R.V.; Hawkins, M.; et al. Phase I study of individualizedstereotactic body radiotherapy for hepatocellular carcinoma and intrahepaticcholangiocarcinoma. J. Clin. Oncol. 26:657–64; 2008.

8. Cheng, J.C.H.; Wu, J.K.; Huang, C.M.; et al. Radiation-induced liver disease afterthree-dimensional conformal radiotherapy for patients with hepatocellularcarcinoma: dosimetric analysis and implication. Int. J. Radiat. Oncol. Biol. Phys.54:156–62; 2002.

9. Sahgal, A.; Roberge, D.; Schellenberg, D.; et al. The Canadian Association ofradiation oncology scope of practice guidelines for lung, liver and spinestereotactic body radiotherapy. Clin. Oncol. R. Coll. Radiol. 24:629–39; 2012.

10. Grills, I.S.; McGrath, S.D.; Matuszak, M.M.; et al. Volumetric modulatedarc therapy for delivery of hypofractionated stereotactic lung radiotherapy:a dosimetric and treatment efficiency analysis. Radiother. Oncol. 95:153–7;2010.

11. Roa, D.E.; Schiffner, D.C.; Zhang, J.; et al. The use of rapidarc volumetric-modulated arc therapy to deliver stereotactic radiosurgery and stereotacticbody radiotherapy to intracranial and extracranial targets. Med. Dosim.37:257–64; 2012.

12. Ali, A.N.; Dhabaan, A.H.; Jarrio, C.S.; et al. Dosimetric comparison of volumetricmodulated arc therapy and intensity-modulated radiation therapy for pancre-atic malignancies. Med. Dosim. 37:271–5; 2012.

13. Fiveash, J.B.; Clark, G.M.; Popple, R.A.; et al. Feasibility of single-isocentervolumetric modulated arc radiosurgery for treatment of multiple brain meta-stases. Int. J. Radiat. Oncol. Biol. Phys. 76:296–302; 2010.

14. Dawson, L.A.; Eccles, C.; Craig, T. Individualized image guided ISO-NTCP basedliver cancer SBRT. Acta. Oncol. 45:856–64; 2006.

15. Shaw, E.; Kline, R.; Gillin, M.; et al. Radiation-therapy oncology group—radio-surgery quality assurance guidelines. Int. J. Radiat. Oncol. Biol. Phys. 27:1231–9;1993.

16. Han, C.H.; Liu, A.; Schultheiss, T.E.; et al. Dosimetric comparisons of helicaltomotherapy treatment plans and step-and-shoot intensity-modulated radio-surgery treatment plans in intracranl stereotactic radiosurgery. Int. J. Radiat.Oncol. Biol. Phys. 65:608–16; 2006.

17. Kutcher, G.J.; Burman, C.; Brewster, L.; et al. Histogram reduction method forcalculating complication probabilities for 3-dimensional treatment planningevaluations. Int. J. Radiat. Oncol. Biol. Phys. 21:137–46; 1991.

18. Niemierko, A.; Goitein, M. Calculation of normal tissue complication probabilityand dose-volume histogram reduction schemes for tissues with a criticalelement architecture. Radiother. Oncol. 20:166–76; 1991.

19. Delana, A.; Menegotti, L.; Bolner, A.; et al. Impact of residual setup error onparotid gland dose in intensity-modulated radiation therapy with or withoutplanning organ-at-risk margin. Strahlenther. Onkol. 185:453–9; 2009.

20. Seong, J.; Koom, W.S.; Han, K.H.; et al. Is local radiotherapy still valuable forpatients with multiple intrahepatic hepatocellular carcinomas? Int. J. Radiat.Oncol. Biol. Phys. 77:1433–40; 2010.

21. Chou, C.H.; Chen, P.J.; Jeng, Y.M.; et al. Synergistic effect of radiation andinterleukin-6 on hepatitis B virus reactivation in liver through STAT3 signalingpathway. Int. J. Radiat. Oncol. Biol. Phys. 75:1545–52; 2009.

22. Dawson, L.A.; Normolle, D.; Balter, J.M.; et al. Analysis of radiation-induced liverdisease using the lyman NTCP model. Int. J. Radiat. Oncol. Biol. Phys. 53:810–21;2002.

23. Cheng, J.C.H.; Wu, J.K.; Lee, P.C.T.; et al. Biologic susceptibility of hepatocellularcarcinoma patients treated with radiotherapy to radiation-induced liverdisease. Int. J. Radiat. Oncol. Biol. Phys. 60:1502–9; 2004.

24. Jiang, G.L.; Liang, S.X.; Zhu, X.D.; et al. Radiation-induced liver disease in three-dimensional conformal radiation therapy for primary liver carcinoma: the riskfactors and hepatic radiation tolerance. Int. J. Radiat. Oncol. Biol. Phys.65:426–34; 2006.