eI

Medical Dosimetry 38 (2013) 5-11

Medical Dosimetry

journal homepage: www.meddos.org

h

Applicator-guided volumetric-modulated arc therapy for low-risk endometrialcancer

Savino Cilla, Ph.D.,* Gabriella Macchia, M.D.,† Domenico Sabatino, Ph.D.,* Cinzia DigesÚ, M.D.,†

Francesco Deodato, M.D.,† Angelo Piermattei, Ph.D.,‡ Marco De Spirito, Ph.D.,* andAlessio G. Morganti, M.D.†,§

*Medical Physics Unit and †Radiation Oncology Unit, Fondazione di ricerca e cura “Giovanni Paolo II,” UniversitÞ Cattolica del Sacro Cuore, Campobasso, Italy; ‡Physics Institute and§Radiation Oncology Unit, UniversitÞ Cattolica del Sacro Cuore, Rome, Italy

A R T I C L E I N F O

Article history:

A B S T R A C T

The aim of this study was to report the feasibility of volumetric-modulated arc therapy (VMAT) in the

Received 14 October 2011Accepted 09 April 2012

postoperative irradiation of the vaginal vault. Moreover, the VMAT technique was compared with 3Dconformal radiotherapy (3D-CRT) and fixed-field intensity-modulated radiotherapy (IMRT), in terms oftarget coverage and organs at risk sparing. The number of monitor units and the delivery time were

analyzed to score the treatment efficiency. All plans were verified in a dedicated solid water phantomusing a 2D array of ionization chambers. Twelve patients with endometrial carcinoma who underwentradical hystero-adenexectomy and fixed-field IMRT treatments were retrospectively included in thisanalysis; for each patient, plans were compared in terms of dose-volume histograms, homogeneity index,and conformity indexes. All techniques met the prescription goal for planning target volume coverage,with VMAT showing the highest level of conformity at all dose levels. VMAT resulted in significantreduction of rectal and bladder volumes irradiated at all dose levels compared with 3D-CRT. No significantdifferences were found with respect to IMRT. Moreover, a significant improvement of the dose conformitywas reached by VMAT technique not only at the 95% dose level (0.74 vs. 0.67 and 0.62) but also at 50% and75% levels of dose prescription. In addition, VMAT plans showed a significant reduction of monitor units bynearly 28% with respect to IMRT, and reduced treatment time from 11 to �3 minutes for a single 6-Gyfraction. In conclusion, VMAT plans can be planned and carried out with high quality and efficiency for theirradiation of vaginal vault alone, providing similar or better sparing of organs at risk to fixed-field IMRTand resulting in the most efficient treatment option. VMAT is currently our standard approach forradiotherapy of low-risk endometrial cancer.

� 2013 American Association of Medical Dosimetrists.

Keywords:VMATIMRTEndometrial cancer

ofsr

Bluls(mvmb

Introduction

Current international guidelines include adjuvant radiotherapy as atreatment option in the treatment of low-risk endometrial cancer.1,2 Re-cent multicenter, randomized trials comparing surgery alone with sur-gery plus adjuvant radiotherapy in endometrial cancer patients (stage I)documented longer progression-free survival, although it failed to showbetter overall survival inmultimodality treatment patients.3

Because themajority of local recurrences in patientswith low-riskendometrial carcinoma is localized on vaginal vault,4,5 exclusive ad-juvant vaginal brachytherapy has been proposed as an effective treat-ment in this clinical setting by several authors,6–8 showing that post-

Reprint requests to: Savino Cilla, Ph.D., Medical Physics Unit, Fondazione di ricercacura, “Giovanni Paolo II,” UniversitÁ Cattolica del Sacro Cuore, 86100 Campobasso,

wtaly.

E-mail: savinocilla@gmail.com

0958-3947/$ – see front matter Copyright � 2013 American Association of Medical Dosimetrttp://dx.doi.org/10.1016/j.meddos.2012.04.004

perative brachytherapy (BRT) alone compared with standardractionation external beam radiotherapy (EBRT) had similar overallurvival and locoregional failure rates, as well as cumulative recur-ence rates and lower toxicity rates.

However, the BRT technique requires the presence of a dedicatedRT unit in any radiation oncology center. In addition, it presents theimit of an intrinsic inhomogeneity of dose distribution. In the partic-lar case of the treatment of vaginal vault, there can be a significantlyower dose delivery to the deeper vaginal layer with respect to theurface of the vaginalmucosa, and if the target volume extends deeplyor has irregular shapes), it becomes very difficult to use BRT treat-ent avoiding organs-at-risk (OAR) overdosage. From this point ofiew, the search for new treatment modalities, such as intensity-odulated radiotherapy (IMRT) or stereotactic radiotherapy haveeen explored extensively.9–14 Some studies have been performed

ith the aim to compare IMRTwith standard conventional radiother-

ists

vbvstcidtmta

tofico

f

m9vutsDss

ptwtptati

T

F

t

tations

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-116

apy (3D-CRT) for pelvic treatments, finding a significant reduction inOARs irradiation with IMRT planning and suggesting a potential de-crease of both acute and late treatment-related toxicities.15–20

Regarding the irradiation of the vaginal vault alone, the IMRT tech-nique was proposed as an alternative to high-dose-rate brachyther-apy (HDR-BRT) in the postoperative radiotherapy of endometrial car-cinoma,21 showing that IMRT plans had comparable planning targetolume (PTV) coverage but lower inhomogeneity and lower rectal/ladder doses than HDR therapy, thus prompting further clinical in-estigations. In our recent study22 we showed that IMRT provided aignificant dosimetric advantage over various 3D-CRT strategies inhe adjuvant treatment of vaginal vault alone in terms of treatmentonformity and rectumand bladder sparing at all dose levels, suggest-ng that IMRT may be safely used to reduce rectum and bladder irra-iation and/or to escalate the dose delivered to the vaginal vault. Onhe basis of these results, the IMRT technique was clinically imple-ented in our institution and the subsequent clinical data confirmed

hat the reduction in normal tissue irradiation translated into an over-ll reduction in acute treatment-related toxicity.23

VMAT is a newer technique of delivering IMRT, in which highlyconformal doses can be realized by varying the speed of gantry rota-tion, themultileaf collimator (MLC) shape, and thedose rate.24,25 Untiloday, a very large body of studies have shown the dosimetric benefitsf VMAT and are able to provide similar plan quality with respect toxed-field IMRT, but with large reduction in treatment time and con-omitant reduction in monitor units (MUs), but to our knowledge,nly one paper addresses gynecologic tumors.26

The aim of this study was to report the feasibility of the VMATtechnique in our department for the irradiation of low-risk endo-metrial cancer; in addition, a dosimetric comparison among VMAT,fixed-field IMRT, and optimized 3D-CRT plans with respect to tar-get coverage, irradiation of OARs, and treatment efficiency wasinvestigated.

Methods and Materials

Study design

At our institution, a short course hypofractionated exclusive fixed-field IMRT treat-ment was routinely performed in patients with low-risk endometrial cancer in the last7 years. In a previous study,23 we showed that in this particular setting, IMRT is aeasible and well-tolerated treatment approach at 30 Gy dose level, 6 Gy/fraction.

In this study, 12 patients with endometrial carcinoma who underwent radicalhystero-adenexectomywith or without pelvic lymphadenectomy and fixed-field IMRTtreatments were selected to be replanned with the VMAT technique. Dosimetric com-parison and treatment efficiency among VMAT, step-and-shoot IMRT, and 3D-CRTplans was performed for all patients. The accuracy of delivered dose was assessed bymeans of bidimensional ion-chambers array and gamma index analysis.

Target definition and patient positioning

The device for immobilization and localization of the vaginal vault was a home-made vaginal applicator composed of polymethyl methacrylate (PMMA) material.Three radio-opaque markers, 1 mm in diameter were inserted into the applicator to

Fig. 1. Field shape arrangements at some representative control points during a full rohielded only when it is in front of the target (red volume).

allow for their X-ray visualization. Further details about this vaginal applicator and itsuse as a device for localization and immobilization are given in a previous study.27

During the simulation process, patients were immobilized in the supine position withan individually fashioned immobilization device; to minimize organ motion, patientswere instructed to empty their bladders 1 hour before computed tomography (CT)simulation, and for daily treatment fractions to drink 300mL ofwater and to empty therectum 2 hours before treatment.

CT images were taken at 3-mm increments. The clinical target volume (CTV) wasdefined as the upper two thirds of the vaginal vault, identified on the simulation CTscans. CTV was expanded uniformly by 5 mm in all directions to produce the planningtarget volume (PTV). Critical organs (OAR) were contoured as follows: (1) The rectumwas contoured from the lower margin of ischial tuberosities to the rectosigmoid junc-tion, with no distinction between the rectal wall and its content; (2) the bladder wascontoured entirely, with no distinction between the wall and its content; (3) heads ofthe femurwere contoured from the cranial extremity to the level of the lowermargin ofthe ischial tuberosities; and (4) the small intestine was defined as all intestinal loopssituated below the promontory (rectosigmoid excluded).

Prescribed dose was 30 Gy in 5 fractions (6 Gy/fraction); the equivalent dose in2-Gy fractions for late effects (a/b ratio: 3) relative to this dose level was 54 Gy.28

All planswere performed primarily to optimize PTV dose coverage and secondarily toinimize dose to the rectum and bladder. The requirement on target coverage was that5% of target volume receive at least 95% of the prescribed dose, and 100% of the targetolume received at least 90% of prescribed dose. Secondary, rectal and bladder dose-vol-me histogram (DVH) goals were �50% of rectum and bladder volumes receiving 50% ofhe prescribed dose and �10% of rectum and bladder volume receiving 90% of the pre-cribeddose. Tertiary,with less priority, dose to both femoral headswere restricted so that10% �50% of the prescribed dose, i.e., not more than 10% of either femoral head volume,hould receive a dose �50% of prescribed dose. Doses to the small bowel were not con-trained because in all patients the small bowel was well outside of treatment fields.

In our clinical routine, patient setup was checked before every daily treatment byortal imagingverification, using2orthogonal virtual beamsopen square at 0� (anteropos-erior) and 90� (lateral) and comparing the obtained megavoltage portal images (MVPIs)ith the corresponding digitally reconstructed radiography (DRR) of the same beams ob-ained by the treatment planning system (TPS). An initial verification was made by com-aring thebonyanatomywith respect to thevirtual beams toensure the correctpositionofhe isocenter point. Then a check of the 3 applicator radio-opaque markers was made tossure their correctpositionwithrespect to theirpredictedpositionbyTPS.Thepatientwasreated only if the relative variations of the applicator markers position between the 2mages were within 3mm along the 3 spatial directions.

reatment planning

ixed-field IMRT and 3D-CRT plans.3D-CRT and step-and-shoot IMRT plans were generated using the Plato Sunrise

reatment planning system (Plato Sunrise, Nucletron BV, Veenendaal, the Nether-

. (A) 190�, (B) 280�, (C) 0�, (D) 80�, and (E) 170� gantry angles. Rectum (pink volume) is

Table 1PTV: comparison of dose statistics and treatment efficiency for 12 patients (mean� SD)

3D-CRT IMRT VMATp VMAT vs.

3D-CRT

p VMAT vs.

IMRT

Dmean 99.9 � 1.2 102.3 � 0.6 102.5 � 1.5 0.002 0.488V95% 99.9 � 0.3 98.2 � 0.8 97.9 � 0.9 < 0.001 0.194D98% 97.9 � 0.8 96.1 � 1.0 95.0 � 2.0 0.003 0.071HI 4.6 � 1.1 10.8 � 1.2 10.1 � 2.0 < 0.001 0.444CI95 0.62 � 0.06 0.67 � 0.07 0.74 � 0.07 < 0.001 < 0.001CI75 0.38 � 0.02 0.39 � 0.04 0.45 � 0.08 0.027 0.054CI50 0.18 � 0.03 0.14 � 0.02 0.19 � 0.06 0.572 0.023MUs/fraction 655.3 � 29.6 1013.0 � 150.1 729.6 � 69.4 < 0.001 < 0.001Delivery time

(min)4.6 � 0.1 10.8 � 0.7 2.5 � 0.3 < 0.001 < 0.001

All dosimetric data are expressed as percentage values.

wcmf

o

vdtivc9

D9et

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-11 7

lands). In a previous paper,22 we showed that 3D-CRT plans supplied the best resultsith a 5-coplanar, 15-MV energy beam arrangements. For these plans, fields wereonformally shaped using the beam’s eye view (BEV) projections of the PTV byeans of the MLC, and a uniform margin of 0.6 cm to the PTV was added to account

Fig. 2. Axial view of the dose distribution in terms of 50%, 75%, and 95% isodoses ofthe prescribed dose for (A) 3D-CRT, (B) IMRT, and (C) VMAT plans for onerepresentative patient.

or beam penumbra.fb

IMRT plans were generated using the ITP/Plato inverse planning module, a com-mercially available version of the Konrad software, developed by Bortfeld29 using theptimization process described in a study by Cilla et al.22 On the basis of a previous

analysis, a class solution formed by 7 coplanar beams (consisting of the following gan-try: 0�, 50�, 100�, 150�, 210�, 260�, 310�) was selected for all patients. The number ofsegments for every beam usually ranged between 10 and 15.

VMAT plans.VMAT plans were generated with ergo�� TPS version 1.7.3 (Elekta, Crawley,

UK). This is an anatomy-based TPS that supplies a simplified approach to createVMAT plans by predefining a series of aperture shapes using Boolean operations inconjunction with the BEV of the target and OAR. The optimization process can bedivided into 2 steps. In the first step, the aperture shape for each segment within anarc is determined by the BEV of the target and adjacent critical structure. By meansof Boolean operators, OAR can always be automatically shielded or only when theyare in front of the target in the BEVs. An example of automatic OAR shielding isprovided in Fig. 1, where some representative BEVs arrangements at some repre-sentative control points during a full rotation are shown. In the second step, thebeam weights for all the control points are optimized by inverse planning based onthe simulated annealing optimization algorithm (AMOA), which computes theweight of each subarc depending on dose and dose-volumes constraints for thetarget and OAR volumes, thus defining the dose-rate/MU number for each subarc.

In addition to dynamically changing the MLC shape, dose rate and gantry speed,ergo�� also offers the possibility to modified the collimator angle during the rota-tion. In any event, our preliminary analysis of clinical cases did not suggest anadditional gain of optimized collimator angles for the PTV shapes, but we used afixed collimator angle equal to 0 for all plans. All plans were generated with asingle-arc clockwise rotation with the starting angle at 190� and the stop angle at170�. The entire gantry rotation is described in the optimization process by a se-quence of 35 control points (CPs), i.e., one every 10�. The dose calculation wasperformed using the pencil beam algorithm with inhomogeneity correction and adose grid resolution of 0.2 cm.

VMAT plans were exported to the R&V systemMosaiq 1.6 (Impac Software, Elekta)by DICOM radiotherapy export for later irradiation.

Plans evaluation and comparison

All plans were compared using DVH analysis and conformity indexes (CIs) to ana-lyze PTV coverage and avoidance of adjacent OARs.

For all plans, the volumes of PTV receiving �95% (V95) and 107% (V107) of theprescribed dose were recorded. Moreover, targets volume coverage was compared interms ofminimal, maximal, mean dose, and homogeneity index (HI). HIwas defined as.

HI � 100 ��D2 � D98�

Dp, (1)

Where D98% is the dose belowwhich 98% of the PTV volumewas treated, D2% the dosebelow which 2% of the PTV volume was treated (indicating that only 2% of the targetvolume receives this dose or higher), and Dp was the prescribed dose. D98% and D2%were considered surrogate for minimum andmaximum dose to avoid distortion of theabsolute values caused by the finite resolution of the calculation grid. Lower HI valuesindicated a more homogeneous target dose.

For an overall comparison of dose distributions, a CI was calculated following theVan’t Riet definition30:

CI �TVRI

TV�

TVRI

VRI, (2)

Where TVRI was the target volume covered by the reference isodose, TV was the targetolume, and VRI was the volume of the reference isodose. The first part of this equationefines the quality of target coverage and the second part defines the volume of healthyissues receiving a dose greater than or equal to the prescribed dose. In this way, thisndex takes simultaneously into account the irradiation of target and healthy tissueolumes. CI ranges from 0 (complete PTV geographic miss) to the ideal value 1 (perfectonformality of the reference isodose to the PTV). Reference isodose was selected as5% of the prescribed dose (CI95). In addition, to quantify the irradiation of healthy

tissue in lower dose regions, conformity indexes were defined also for reference iso-doses at 50% and 75% of the prescribed dose (CI50 and CI75).

Rectum and bladder avoidance was evaluated using the following parameters:mean, maximum dose (D2%), and absolute organ volume receiving 30%, 50%, 75%,5% and 100% of prescribed dose. For the femoral heads sparing, the chosen param-ters for comparison were Dmean and D10%, i.e., the dose administrated to 10% ofhe volume.

For each patient, VMAT delivery parameters were recorded in terms of MU per

raction and treatment time, this last defined from the start of irradiation of the firsteam or arc to the end of delivery, and excluding patient positioning and imaging

f

D

F2sua1ptOa

bt

gt

time,

TR

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-118

procedures. These datawere comparedwith those relative to conformal and fixed-fieldIMRT plans to score the treatment efficiency.

In all comparisons, Student’s t-test was used for the calculation of statistical differ-ences; statistical significance was assumed at p � 0.05. Statistical analyses were per-ormed using SPSS (IBM, Armonk, NY).

osimetric verification

Dosimetric verificationwas performed using the Seven29 ion chamber array (PTW,reiburg, Germany). This dectector, extensively described in other papers,31 consists of7 � 27 vented cubic ion chambers of 0.5 � 0.5 � 0.5 mL each, with a center-to-centerpacing of 1 cm. For VMAT plans, a dedicated phantom was developed by PTW to besed with the seven29 array.32 This phantom has an octagonal shape to allow datacquisition inmultiple planes and ismade of polystyrenewith physical density equal to.04 g/cm3. A central cavity allows the user to insert the 2D ion chamber array into thehantom in such a way that the plane through the middle of the ion chambers goeshrough the center of the phantom. Every arc was recalculated on phantom using thectavius geometry and density. In particular, the dose was measured on both coronalnd sagittal planes for every arc. Comparison of measured vs. calculated dose distribu-

tions was done with Verisoft software version 4.0 (PTW). The evaluation of the mea-surements was performed using the gamma index evaluation, taking account dosevalues of at least 5% of themaximum dose of themodulated field. Pretreatment dosim-etry was considered optimal if the percentage of points fulfilling gamma index criteria(P� � 1) exceeded 95% using 3% for dose criterion and 3 mm for the distance to agree-ment criterion (DTA). Average andmaximum values of gamma indexwere also consid-ered for comparison.

Fig. 3. PTV box-and-whisker plots of (A) MU number, (B) treatment

able 2ectum: comparison of dose statistics for 12 patients (mean � SD)

3D-CRT IMRT VMATp VMAT vs.

3D-CRT

p VMAT vs.

IMRT

V30% 81.5 � 17.9 68.5 � 14.3 75.8 � 20.3 0.111 0.178V50% 63.0 � 17.7 44.2 � 7.1 40.9 � 11.4 < 0.001 0.157V75% 34.6 � 12.4 24.4 � 7.4 19.9 � 6.1 < 0.001 0.021V90% 20.8 � 8.8 10.2 � 3.3 9.3 � 4.3 < 0.001 0.125V95% 15.4 � 7.6 5.7 � 2.6 5.6 � 3.2 < 0.001 0.367V100% 6.0 � 6.1 1.4 � 1.2 1.8 � 1.2 0.030 0.502Dmean 59.6 � 12.8 49.3 � 7.9 48.1 � 8.8 < 0.001 0.237D2% 100.2 � 2.0 98.9 � 2.5 99.8 � 3.1 0.338 0.676

All dosimetric data are expressed as percentage values.

Results

Patients’ characteristics

Twelve patients previously treated with surgical resection andpostoperative IMRTwere included in this analysis.Median agewas 58years (range 49–74). Pathologic stage was IB 83.3% and IC 16.7%.

PTV coverage and conformity index

Table 1 shows the comparison among 3D-CRT, IMRT, and VMATin terms of PTV coverage and conformity indexes. Figure 1 showsthe dose distributions in a transverse midplane obtained with 3D-CRT, IMRT, and VMAT plans for one representative patient. Figure 2shows an axial view of the dose distribution in terms of 50%, 75%,and 95% isodoses of prescribed dose for (A) 3D-CRT, (B) IMRT, and(C) VMAT plans for one representative patient. Figure 3 shows theox-and-whisker plots of (A) Monitor Unit numbers, (B) treatmentime, (C) HI, (D) CI50, (E) CI75 and (F) CI95.

Quantitative comparison in terms of conformity indexes showedood discrimination between the 3 plans. Themean scores of CI95 forhe 3D-CRT plans was 0.62 (range 0.43–0.68), for the IMRT plans was

(C) HI, (D) CI at 50% level, (E) CI at 75% level, and (F) CI at 95% level.

Table 3Bladder: comparison of dose statistics for 12 patients (mean � SD)

3D-CRT IMRT VMATp VMAT vs.

3D-CRT

p VMAT vs.

IMRT

V30% 79.4 � 21.6 67.8 � 20.2 60.7 � 23.3 � 0.001 0.034V50% 48.0 � 23.2 43.5 � 16.6 37.0 � 20.6 < 0.001 0.023V75% 24.3 � 12.1 21.3 � 9.3 18.0 � 9.2 < 0.001 0.006V90% 14.2 � 7.7 10.5 � 5.2 9.1 � 4.4 < 0.001 0.002V95% 10.4 � 6.1 7.0 � 3.9 5.8 � 3.0 < 0.001 0.005V100% 3.3 � 3.9 2.5 � 2.2 2.2 � 1.9 0.352 0.338Dmean 50.9 � 14.9 48.0�12.9 43.6 � 13.9 < 0.001 0.002D2% 99.8 � 2.2 99.9 � 3.1 99.6 � 2.6 0.607 0.115

All dosimetric data are expressed as percentage values.

fihnam

O

ID3a

ddesb3tmai

ds

Sd2p

we

rcol

D

tce

A

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-11 9

0.67 (range 0.58–0.78), and for the VMATplanswas 0.74 (range 0.60–0.83). Because the 95% isodose line encompassed �95% of PTV for allthe plans, then the improvement of the conformity for the VMATplans was only the result of a reduction of healthy tissues receivinghigh doses. The advantage of VMAT in terms of dose conformity re-mained also at the 50% and 75% levels of the prescribed dose. Signifi-cant differences were observed in the homogeneity index. 3D-CRTplans show a high level of homogeneity compared with IMRT andVMAT plans. For IMRT and VMAT plans, a better conformity was ob-tained together with a worsening of target dose homogeneity. How-ever, for all patients, the dosimetric constraint V95% � 95%was satis-ed, so the homogeneity worsening translated only in an increase ofigh-dose regions into the target volume and was considered to beot clinically relevant. Indeed, the mean dose supplied by the VMATnd IMRT plans was slightly but significantly greater than the confor-al technique (Table 1).

ARs sparing

Tables 2 and 3 compare the rectum and bladder dose for the 3D-CRT,MRT, and VMAT plans. V30%, V50%, V75%, V90%, V95%, mean dose, and2%were reported. Table 4 compares the dosimetric results obtained byD-CRT, IMRT, andVMATplans for the femoral heads in terms of Dmeannd D10%.

Rectum. Figure 4 shows the box-and-whisker plots of (A) meanose, (B) V50%, (C) V75%, and (D) V95%. VMATplans reduced themeanose by about 19% compared with 3D-CRT, with no significant differ-nces with respect to IMRT. With respect to 3D-CRT, VMAT showedignificant reductions for irradiated volumes over all dose range (Ta-le 2). In particular, VMAT reduced V50%, V75%, and V95% by about5%, 42%, and 64%, respectively, comparedwith 3D-CRT.With respecto IMRT, it should be noted that VMAT resulted in a slight improve-ent in dose sparing (although thiswas not statistically significant) inll dose regions. No significant differenceswere observed for themax-mum dose (D2%) among the 3 techniques (Table 2).

Bladder. Figure 5 shows the box-and-whisker plots of (A) meanose, (B) V50%, (C) V75%, and (D) V95%. The advantages of VMAT areignificant over all dose ranges, with respect to IMRT and 3D-CRT.

Table 4Femoral heads: comparison of dose statistics for 12 patients (mean � SD)

3D-CRT IMRT VMATp VMAT vs.

3D-CRT

p VMAT vs.

IMRT

Right femurDmean 36.2 � 9.6 28.7 � 7.5 36.3 � 13.3 0.924 0.008D10% 44.2 � 6.9 42.7 � 8.9 50.1 � 11.3 0.119 0.155Left femurDmean 36.1 � 6.7 28.1 � 6.2 37.7 � 7.7 0.434 0.008D10% 43.4 � 6.0 40.7 � 5.4 46.9 � 7.5 0.266 0.123

ll dosimetric data are expressed as percentage values.

Fig. 4. Rectum box-and-whisker plots of (A) mea

ignificant reductions were observed in the regions of low and me-ium doses, with a reduction of V50%, V75%, and V95% by about 23%,6%, and 44%, and 15%, 15%, and 17%with respect to 3D-CRT and IMRTlans, respectively.

Moreover, VMAT reduced the mean dose by about 14% and 9%ith respect to 3D-CRT and IMRT, respectively. No significant differ-nces were observed for the maximum dose (D2%) (Table 3).

Femoral heads. VMAT showed a slightly over-irradiation withespect to IMRT and 3D-CRT, although this was not clinically signifi-ant (Table 4). This behavior could be explained by the fact that in theptimization process the femurs are not constrained, so VMAT is al-owed to irradiate mainly from lateral sides.

elivery efficiency and accuracy

Themean number of MUs was 28% lower for VMAT (729.6 � 69.4)han for fixed-field IMRT (1013.0 � 150.1). The total treatment time,alculated from the start of irradiation of the first beam (or arc) to thend of the last delivery was 2.5 � 0.3 minutes for VMAT compared

with 14.1 � 0.7 minutes for IMRT and 4.6 � 0.1 minutes for 3D-CRT.These values did not include the time necessary for patient position-ing and imaging procedures, common to any technique and not rele-vant to our comparison.

Pretreatment verification was performed for all 24 VMAT arcs (2per patient, each arc was irradiated two times, once in coronal andthen in the sagittal plane). With an agreement criteria of 3% in doseand 3mm in DTA, themeasurements of 24 arcs showed an agreementwith calculated valueswith amean gamma value of 0.37 (range 0.30–0.42) and on average 2.7% (range 0.2–4.8%) of the measured pointswith gamma values �1.0.

Discussion

This paper reports the feasibility study and the clinical implemen-tation in our department of the VMAT technique for the treatment oflow-risk endometrial cancer. In addition, the potential benefit ofVMATwith respect to conformal radiotherapy andfixed-field IMRT, interms of targets coverage, critical structures toxicity and treatmentefficiency were quantified.

In our evaluation of VMAT technique vs. fixed-field IMRT and 3D-CRT, we found that the 3 techniques provide almost the same targetdose coverage, meeting the prescription goal for PTV coverage interms of minimum dose and V95%. VMAT shows the highest level ofconformity compared with other techniques, but with respect to ho-mogeneity, the target dose distribution across the PTV was less ho-mogenous with VMAT (and IMRT) with respect to 3D-CRT. However,in our study no underdosage is observedwith VMAT because themin-imum target dose objectives are well fulfilled, so the heterogeneity iscaused by only the presence of hot spots into the target volume, andthe consequences of nonuniform target dose distribution were con-sidered not relevant. The VMAT technique offers the possibility tosignificantly spare the rectum and bladder with regard to the 3D-CRT

n dose, (B) V50%, (C) V75%, and (D) V95%.

mcaaombwanwTmmit

a

A) mea

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-1110

technique.With regard to rectumavoidance, our results show that theuse of VMAT is associated with a significant reduction in irradiatedrectum volumes at all dose levels and a reduction of mean dose by19%, with respect to 3D-CRT. No significant differences were found inthis dose range with respect to IMRT. With regard to bladder avoid-ance, VMAT demonstrated a clear advantage in terms of bladder spar-ing in all dose levels �30% with respect to 3D-CRT.

In this particular anatomical site, the potential interfractionmove-ent of PTV can produce a geographic miss of target irradiation be-ause of very steep dose gradient of intensity-modulated techniquesnd/or an overdosage to OARs. Moreover, the conventional portal im-ging for tumor localizationmay be limited by low contrast of tissuesf interest. From this point of view, the magnitude of vaginal cuffovement in postoperative patients with endometrial cancer haseen investigated in a recent paper.33 In this work, patients under-ent placement of gold seed fiducial markers in the vaginal cuff apexnd daily megavoltage CT imaging was performed during each exter-al radiation treatment fraction. The daily position of fiducialmarkersere then compared with the original position on the simulation CT.he authors concluded that the magnitude of vaginal cuff move-ent was highly patient-specific with an average and maximalovement of 16.2 � 8.3 mm and 34.5 mm, respectively, and can

mpact target coverage mainly in patients without bladder instruc-ions at simulation.

Following the approach known as “applicator-guided radiother-py” developed by Low,34 we realized a customized homemade int-

racavitary applicator to be used as a localization and immobilizationdevice for the vaginal vault. In this way, the positions of target andOAR become tied to the applicator position, supplying more preciseorgan localization during daily portal imaging verification. In otherwords, treatment delivery is carried out by using an image-guidedsystem able to locate to the position of the vaginal endocavitary ap-plicator in each treatment fraction. As explained in Materials andMethods, the patient setupwas checked daily by not only comparisonof bony anatomy with respect to the virtual beams to ensure the cor-rect position of the isocenter, but also by checking the vaginal appli-cator position by means of the radio-opaque markers, correcting anydeviations �3 mm along the 3 spatial directions.

In our previous clinical experience, the daily treatment reproduc-ibility of fixed-field IMRTwas successfully verified bymeans of endo-cavitary in vivodosimetry.27 In particular,we showed that in vivodosemeasurements for 210 IMRT fields supplied agreement with calcu-lated doses within � 5% for �90% of measurements. In addition, for 2patients the rate of measurements �6%was found to be 23% and 17%,respectively, suggesting that dosimetric discrepancies could becaused in part by the small changes of the filling status of the rectumand bladder and in part by intrafraction organ motion caused by pro-longed times of treatment delivery with a mean time about 11minutes.

A major advantage in using the VMAT technique with respect tofixed-field IMRT is the significant reduction in treatment times, with

Fig. 5. Bladder box-and-whisker plots of (

potential advantages in organ motion control. The reduction of deliv-

ery time also remains with respect to the 3D-CRT technique. Thissignificant improvement is caused primarily by the elimination of allthe non–beam-on times, such as MLC movements to realize the var-ious segments of IMRT beams or gantry motion to reach the fixedposition (for 3D-CRT), with the VMAT technique able to irradiate con-tinuously while the gantry rotates around the patient. In our depart-ment, the reduction of mean treatment time from 14 to �3 minutestranslates to a reduction of the whole treatment slot from 24 to �13minutes when patient setup and positioning are considered. Patientcompliance with treatment could be significantly increased and therisk of intrafractional motion is potentially reduced. In addition, thetime gained can be used to increase patient throughput or to increaseimage guidance.

Moreover, our study revealed that VMAT resulted in a smallernumber ofMUs in all 12 plans comparedwith fixed-field IMRT, with apercentage decrease in the mean number of MUs equal to 28%. Be-cause the MU number is correlated with the amount of scatter doseand leakage radiation, the reduced MUs may decrease the risk of ra-diation-induced secondary malignancies.

VMAT plans showed very good agreement between measure-ments and calculation. The average percentage of points passing thegamma test was �95% for every arc (both on coronal and sagittalplanes) with acceptance criteria of 3% to 3mm, then every VMAT planwas considerate clinically acceptable.

On the basis of the results of this paper, we can conclude that inlow-risk endometrial cancer irradiation, the VMAT technique com-bines the advantages of step-and-shoot IMRT with its highly confor-mal dose distribution andOARs sparing and the advantages of 3D-CRTwith its fast delivery and low MU number. The forthcoming clinicaldata will confirm if this reduction in normal tissue irradiation trans-lates into an overall reduction in acute and potentially late treatment-related toxicity.

Conclusions

Twelve patients with low-risk endometrial cancer were selectedfor replanningwith the VMAT technique. This study shows that in theadjuvant treatment of vaginal vault, ergo�� TPS produced planswithsignificant improvement of nearly all planning objectives with regardto target coverage andOAR sparing over fixed-field IMRT and 3D-CRT,suggesting that the VMAT technique may be safely used to furtherreduce rectum and bladder irradiation. In addition, treatment timeswere significantly reduced comparedwith IMRT,with all patients ableto be treated within 3 minutes (for a 6-Gy fraction).

References

1. National Cancer Institute. PDQ � NCI’s comprehensive cancer database. Availableat: http://www.cancer.gov/cancerinfo/PDQ/cancerdatabase. Accessed August2011.

2. National Comprehensive Cancer Network. Clinical practice guidelines in oncology,

n dose, (B) V50%, (C) V75%, and (D) V95%.

v. 2. Available at: http://www.nccn.org/professionals/physician_gls/f_guidelines.asp. Accessed August 2011.

1

1

1

1

1

1

S. Cilla et al. / Medical Dosimetry 38 (2013) 5-11 11

3. Roberts, J.A.; Brunetto, V.L.; Keys, H.M. A phase III randomized study of surgery vs.surgery plus adjunctive radiation therapy in intermediate risk endometrial adeno-carcinoma (GOG 99). Society of Gynecologic Oncologists, 29th Annual Meeting.Gynecol. Oncol. 68:135 (abs 298); 1998.

4. Creasman, W.T.; Morrow, C.P.; Bundy, B.N. Surgical pathologic spread patterns ofendometrial cancer. AGynecologic oncology group study. Cancer. 60:2035–41; 1987.

5. Morrow, C.P.; Bundy, B.N.; Kurman, R.J. Relationship between surgical-pathologicalrisk factors and outcome in clinical stage I and II carcinoma of the endometrium: AGynecologic Oncology Group study. Gynecol. Oncol. 40:55–65; 1991.

6. Alektiar, K.M.; Venkatraman, E.; Chi, D.S. Intravaginal brachytherapy alone for in-termediate-risk endometrial cancer. Int. J. Radiat. Oncol. Biol. Phys. 62:111–7; 2005.

7. Jolly, S.; Vargas, C.; Kumar, T. Vaginal brachytherapy alone: An alternative to adju-vant whole pelvis radiation for early stage endometrial cancer. Gynecol. Oncol.97:887–92; 2005.

8. Chong, I.; Hoskin, P.J. Vaginal vault brachytherapy as sole postoperative treatmentfor low-risk endometrial cancer. Brachytherapy. 7:195–9; 2008.

9. Mundt, A.J.; Lujan, A.E.; Rotmensch, J. Intensity-modulatedwhole pelvic radiother-apy in women with gynecologic malignancies. Int. J. Radiat. Oncol. Biol. Phys. 52:1330–7; 2002.

0. Ahamad, A.; Jhingran, A. New radiation techniques in gynecological cancer. Int. J.Gynecol. Cancer. 14:569–79; 2004.

1. Portelance, L.; Chao, K.S.; Grigsby, P.W. Intensity-modulated radiation therapy(IMRT) reduces small bowel, rectum, and bladder doses in patients with cervicalcancer receiving pelvic and para-aortic irradiation. Int. J. Radiat. Oncol. Biol. Phys.51:261–6; 2001.

2. Wahab, S.H.; Malyapa, R.S.; Mutic, S. A treatment planning study comparing HDRand AGIMRT for cervical cancer.Med. Phys. 31:734–43; 2004.

3. Guerrero, M.; Li, X.A.; Ma, L. Simultaneous integrated intensity-modulated radio-therapy boost for locally advanced gynecological cancer: Radiobiological and do-simetric considerations. Int. J. Radiat. Oncol. Biol. Phys. 62:933–9; 2005.

4. MollÁ, M.; Escude, L.; Nouet, P. Fractionated stereotactic radiotherapy boost forgynecologic tumors: An alternative to brachytherapy? Int. J. Radiat. Oncol. Biol. Phys.62:118–24; 2005.

5. Mundt, A.J.; Lujan, A.E.; Rotmensch, J. Intensity-modulatedwhole pelvic radiother-apy in women with gynecologic malignancies. Int. J. Radiat. Oncol. Biol. Phys. 52:1330–7; 2002.

16. Portelance, L.; Chao, K.S.; Grigsby, P.W. Intensity-modulated radiation therapy(IMRT) reduces small bowel, rectum, and bladder doses in patients with cervicalcancer receiving pelvic and para-aortic irradiation. Int. J. Radiat. Oncol. Biol. Phys.51:261–6; 2001.

17. Ahamad, A.; D’Souza, W.; Salehpour, M. Intensity-modulated radiation therapyafter hysterectomy: Comparison with conventional treatment and sensitivity ofthe normal-tissue-sparing effect to margin size. Int. J. Radiat. Oncol. Biol. Phys. 62:1117–24; 2005.

18. Heron, D.E.; Gerszten, K.; Selvaraj, R.N. Conventional 3D conformal versus intensi-ty-modulated radiotherapy for the adjuvant treatment of gynecologic malignan-

cies: A comparative dosimetric study of dose-volume histograms small star, filled.Gynecol. Oncol. 91:39–45; 2003.

19. D’Souza, W.D.; Ahamad, A.A.; Iyer, R.B. Feasibility of dose escalation usingintensity-modulated radiotherapy in posthysterectomy cervical carcinoma. Int. J.Radiat. Oncol. Biol. Phys. 61:1062–70; 2005.

20. Nutting, C.M.; Convery, D.J.; Cosgrove, V.P. Reduction of small and large bowelirradiation using an optimized intensity-modulated pelvic radiotherapy techniquein patients with prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 48:649–56; 2000.

21. Aydogan, B.; Mundt, A.J.; Smith, B.D. A dosimetric analysis of intensity-modulatedradiation therapy (IMRT) as an alternative to adjuvant high-dose-rate (HDR)brachytherapy in early endometrial cancer patients. Int. J. Radiat. Oncol. Biol. Phys.65:266–73; 2006.

22. Cilla, S.; Macchia, G.; DigesÚ, C. 3D-conformal versus intensity-modulated postop-erative radiotherapy of vaginal vault: A dosimetric comparison. Med. Dosim. 35:135–42; 2010.

23. Macchia, G.; Cilla, S.; Ferrandina, G. Postoperative intensity-modulated radiother-apy in low-risk endometrial cancers: Final results of a phase I study. Int. J. Radiat.Oncol. Biol. Phys. 76:1390–5; 2010.

24. Otto, K. Volumetric modulated arc therapy: IMRT in a single gantry arc.Med. Phys.35:310–7; 2008.

25. Palma, D.A.; Verbakel, W.F.; Otto, K. New developments in arc radiation therapy: Areview. Cancer Treat. Rev. 36:393–9; 2010.

26. Cozzi, L.; Dinshaw, K.A.; Shrivastava, S.K. A treatment planning study comparingvolumetric arc modulation with RapidArc and fixed field IMRT for cervix uteriradiotherapy. Radiother. Oncol. 89:180–91; 2008.

27. Cilla, S.;Macchia, G.; DigesÚ, C. Endocavitary in vivo dosimetry for IMRT treatmentsof gynecologic cancers.Med. Dosim. 36:455–62; 2011.

28. Withers, H.R.; Thames, H.D.; Peters, L.J. A new isoeffect curve for change in dose perfraction. Radiol. Oncol. 1:187–91; 1983.

29. Bortfeld, T.; Schlegel, W.; Rhein, B. Decomposition of pencil beam kernels for fastdose calculations in three-dimensional treatment planning. Med. Phys. 20:311–8;1993.

30. Van’t Riet, A.; Mak, A.C.; Moerland, M.A. A conformation number to quantify thedegree of conformality in brachytherapy and external beam irradiation: Applica-tion to the prostate. Int. J. Radiat. Oncol. Biol. Phys. 37:731–6; 1997.

31. Spezi, E.; Angelini, A.L.; Romani, F. Characterization of a 2D ion chamber array forthe verification of radiotherapy treatments. Phys. Med. Biol. 50:3361–73; 2005.

32. Van Esch, A.; Clermont, C.; Devillers, M. On-line quality assurance of rotationalradiotherapy treatment delivery by means of a 2D ion chamber array and theOctavius phantom.Med. Phys. 34:3825–37; 2007.

33. Ma, D.J.; Michaletz-Lorenz, M.; Goddu, S.M. Magnitude of interfractional vaginalcuff movement: Implications for external irradiation. Int. J. Radiat. Oncol. Biol. Phys.82: 82, Nos. :1439–1444; 2012.

34. Low, D.A.; Grigsby, P.W.; Dempsey, J.F. Applicator-guided intensity-modulated ra-diation therapy. Int. J. Radiat. Oncol. Biol. Phys. 52:1400–6; 2002.