Volumetric modulated arc radiotherapy for carcinomas of the oro-pharynx, hypo-pharynx and larynx: A treatment planning comparison with fixed field IMRT

Radiotherapy and Oncology 92 (2009) 111–117

Contents lists available at ScienceDirect

Radiotherapy and Oncology

journal homepage: www.thegreenjournal .com

Head and neck IMRT

Volumetric modulated arc radiotherapy for carcinomas of the oro-pharynx,hypo-pharynx and larynx: A treatment planning comparison with fixed field IMRT

Eugenio Vanetti a, Alessandro Clivio a, Giorgia Nicolini a, Antonella Fogliata a, Sarbani Ghosh-Laskar b,Jai Prakash Agarwal b, Ritu Raj Upreti b, Ashwini Budrukkar b, Vedang Murthy b,Deepak Dattatray Deshpande b, Shyam Kishore Shrivastava b, Ketayun Ardeshir Dinshaw b, Luca Cozzi a,*

a Oncology Institute of Southern Switzerland, Radiation Oncology Department, Bellinzona, Switzerlandb Departments of Radiation Oncology & Medical Physics, Tata Memorial Hospital, Mumbai, India

a r t i c l e i n f o

Article history:Received 2 September 2008Received in revised form 23 December 2008Accepted 26 December 2008Available online 20 January 2009

Keywords:RapidArcIMRTHead and neck radiation therapy

0167-8140/$ - see front matter � 2009 Elsevier Irelandoi:10.1016/j.radonc.2008.12.008

* Corresponding author. Address: Oncology InstituRadiation Oncology Department, Medical Physics Unland.

E-mail address: [email protected] (L. Cozzi).

a b s t r a c t

Purpose: A planning study was performed to evaluate the performance of volumetric modulated arcradiotherapy on head and neck cancer patients. Conventional fixed field IMRT was used as a benchmark.Methods and materials: CT datasets of 29 patients with squamous cell carcinoma of the oro-pharynx,hypo-pharynx and larynx were included. Plans for fixed beam IMRT, single (RA1) and double (RA2) mod-ulated arcs with the RapidArc technique were optimised. Dose prescription was set to 66 Gy to the pri-mary tumour (at 2.2 Gy/fraction), 60 Gy to intermediate-risk nodes and 54 Gy to low-risk nodal levels.The planning objectives for PTV were minimum dose >95%, and maximum dose <107%. Maximum doseto spinal cord was limited to 46 Gy, maximum to brain stem to 50 Gy. For parotids, mean dose <26 Gy (ormedian <30 Gy) was assumed as the objective. The MU and delivery time were scored to measureexpected treatment efficiency.Results: Target coverage and homogeneity results improved with RA2 plans compared to both RA1 andIMRT. All the techniques fulfilled the objectives on maximum dose, while small deviations were observedon minimum dose for PTV. The conformity index (CI95%) was 1.7 ± 0.2 for all the three techniques. RA2allowed a reduction of D2% to spinal cord of �3 Gy compared to IMRT (RA1 D2% increased it of �1 Gy).On brain stem, D2% was reduced from 12 Gy (RA1 vs. IMRT) to 13.5 Gy (RA2 vs. IMRT). The mean doseto ipsi-lateral parotids was reduced from 40 Gy (IMRT) to 36.2 Gy (RA1) and 34.4 Gy (RA2). The meandose to the contra-lateral gland ranged from 32.6 Gy (IMRT) to 30.9 Gy (RA1) and 28.2 Gy (RA2).Conclusion: RapidArc was investigated for head and neck cancer. RA1 and RA2 showed some improve-ments in organs at risk and healthy tissue sparing, while only RA2 offered improved target coverage withrespect to conventional IMRT.

� 2009 Elsevier Ireland Ltd. All rights reserved. Radiotherapy and Oncology 92 (2009) 111–117

The aim of the present study was to investigate the potentialclinical role for head and neck cancer patients of RapidArc, thenovel radiation treatment technique (Varian Medical Systems),which is based on volumetric intensity-modulated arc delivery,as opposed to intensity modulation which uses fixed gantry beams.

RapidArc falls into the category of intensity modulation therapywith arcs (IMAT) [1–4]. The Yu’s group established the benefit ofusing multiple modulated arcs for complex cases [5,6]. The Ghentgroup applied IMAT techniques with multiple non-coplanar beamsto pelvic treatments [7,8] proving equivalent or superior targetcoverage and improved sparing of OARs compared to conventionalconformal treatments.

d Ltd. All rights reserved.

te of Southern Switzerland,it, 6504 Bellinzona, Switzer-

RapidArc is a technique based on an investigation from K. Otto[9] and it aims to (i) improve OARs and healthy tissue sparing com-pared to other solutions; (ii) maintain or improve the same degreeof target coverage; and (iii) reduce beam-on time per fraction.

Faster treatments could have a clinical impact on patients interms of comfort on couch, immobility and minimisation of inter-nal organ displacement. It could also allow more time for imagingprocedures allowing, in perspective, routine application of adap-tive treatment strategies when changes are observed due toresponse to radiation as in patients with head and neck cancer.

IMRT in head and neck cancer patients has been largely investi-gated at both planning and clinical levels. Excellent reviews fortreatment outcome and major toxicity patterns can be found inGregoire et al. [10], Lee et al. [11] and Popovtzer et al. [12]. Onthe toxicity side, besides the major attention given to spinal cordand brain stem (with toxicity thresholds set in the proximity of45–50 Gy for the first and at 50 Gy for the second), it is generally

112 Volumetric arc modulation for head and neck radiation therapy

known that, for parotids, mean doses inferior to 25–30 Gy corre-late well with substantial recovery of function within two years[13] (higher thresholds were observed for sub-mandibular glandsin the range of 39 Gy [14]). To reduce dysphagia [15], sparing ofconstrictors (with mean dose below 60 Gy) highly correlatedwith improved swallowing, laryngeal elevation and epiglotticinversion. To manage ototoxicity [16], the mean dose to cochleaproved to be a highly significant factor with some threshold ef-fect in the order of 60 Gy between high and low risks of auditorydefects.

Volumetric IMAT has already been investigated for prostate,small brain tumours and cervix uteri cancer [17–19], i.e. on rela-tively simple clinical cases; head and neck is an ideal advancedbenchmark for assessment of its conformal avoidance capabilitiessince the anatomical features of this location require highly sophis-ticated techniques to ensure adequate treatments.

The present study was initiated as a side investigation in theframework of a larger Phase II trial activated at Tata Memorial Cen-tre (TMC) in Mumbai to investigate the role of IMRT vs 3D confor-mal radiotherapy in squamous cell carcinoma (SCC) patients atAJCC stages T1-3, N0-2b, and M0.

Material and methods

Patient selection and planning objectives

CT data (3 mm slice thickness) for a group of twenty-nine con-secutively treated patients from the TMC protocol were selectedfor the purposes of planning. These patients had been randomizedand treated on the IMRT arm of the protocol. The RapidArc plan-ning was carried out on the same dataset at the Oncology Instituteof Southern Switzerland.

Fourteen tumours were localised in the oro-pharynx district,eight in the hypo-pharynx and seven in the larynx. Fourteen pa-tients presented T3 stage, twelve T2 stage and two T1 stage. No-dal involvement was mostly N0 (16), N1 (6) and N2 (7) stages;all patients presented no distant metastasis (M0). The main or-gans at risk (OAR) considered for all patients were ipsi- and con-tra-lateral parotids, spinal cord and brain stem. For some of thepatients, additional organs were outlined by radiation oncolo-gists depending on the indication including oral cavity, oesopha-geal constrictors, base of the tongue, cochlea, mandible, andvocal apparatus. These additional OARs were defined only forthose patients where potential sparing was possible due to thetarget conformation. The healthy tissue was defined as the pa-tient’s volume covered by the CT scan minus the envelope ofthe various target volumes (PTV). Various PTVs were definedfrom the respective clinical target volumes (CTVs) adding7 mm margins with 3D expansion. Dose prescription was set to66 Gy at 2.2 Gy/fraction to the PTV including the primary tumourand the lymph-nodal metastases (PTV66 – high risk). Two addi-tional elective PTVs were defined to be irradiated at 60 Gy(intermediate risk) and 54 Gy (low risk, sub-clinical disease) inthe nodal regions (PTV60 and PTV54). All volumes were to besimultaneously treated according to the simultaneous integratedboost (SIB) approach. Due to target definition criteria, PTVs didnot overlap and were not mutually included. All plans were nor-malised to the mean dose of PTV66. The mean volume of targetvolumes was PTV66: 342 ± 182 cm3, PTV60: 107 ± 81 cm3, andPTV54: 95 ± 54 cm3 for the global group.

For all the PTVs, plans aimed to achieve minimum dose largerthan 95% of the prescribed dose and a maximum lower than 107%.For the spinal cord, a maximum dose of 46 Gy was allowed. For thebrain stem, the limit was set to 50 Gy. In the case of parotids, theplanning objectives aimed to keep the mean dose below 26 Gy (or

D50% < 30 Gy). For other organs at risk, the planning strategy was tominimise their involvement but no specific constraint was set.

Planning techniques

Three sets of plans were compared in this study, all designed onthe Varian Eclipse treatment planning system (TPS) with 6 MV pho-ton beams from a Varian Clinac equipped with a Millennium Multi-leaf Collimator (MLC) with 120 leaves (spatial resolution of 5 mm atisocentre for the central 20 cm and of 10 mm in the outer 2 � 10 cm,maximum leaf speed of 2.5 cm/s and leaf transmission of 1.8%). Plansfor RapidArc were optimised selecting a maximum DR of 600 MU/min, and a fixed DR of 300 MU/min was selected for IMRT.

The Anisotropic Analytical Algorithm (AAA, version 8.6.02) pho-ton dose calculation algorithm was used for all cases [20–23]. Thedose calculation grid was set to 2.5 mm. RapidArc optimisationwas performed with version 8.6.05.

Details for each planning method are as follows

IMRTReference plans were computed by selecting the ‘conventional’

intensity modulation approach as a benchmark, with fixed gantryand intensity-modulated beams delivering the dose by means ofthe sliding window approach [24–26]. Plans were individuallyoptimised using seven or nine coplanar fields. The modulation flu-ences used in the present analysis are equivalent to the fluencesapproved for patient treatment, while final dose calculation wasperformed using the AAA, including heterogeneity management,instead of the Pencil Beam (used at TMC) for coherence with theRapidArc condition. MUs were kept fixed at the value from originalcalculation. It should be noted that in Eclipse, the optimisation pro-cess to generate the fluences and the final dose calculation arecompletely disentangled and therefore the determination of theoptimal fluence is in no way influenced by the algorithm adoptedafterwards.

RapidArc (RA)To achieve the desired level of modulation required, the instan-

taneous dose rate (DR), MLC leaf positions and the gantry rota-tional speed were continuously varied by the RapidArc optimiser.To minimise the contribution of tongue and groove effect, the col-limator rotation in RapidArc was kept fixed to a value differentfrom zero. In the present study, the collimator was rotated to40�. Details of the RapidArc process can be found in [9,15]. Twosets of plans were optimised and analysed. RA1 consisting of a sin-gle 360� rotation and RA2 consisting of two coplanar arcs of 360�were optimised simultaneously, to be delivered with oppositerotation (clock- and counter clock-wise). The application of twocoplanar arcs aims to increase the modulation factor during opti-misation. In fact, since each individual arc is limited to a sequenceof 177 control points (i.e. ‘’elementary’’ fields), the application oftwo independent arcs, simultaneously optimised, could allow theoptimiser to achieve higher target homogeneity and lower OARsinvolvement at the same time, as seen in other IMAT applications[5–8]. In the previous investigations on RapidArc [17,18], the rela-tive simplicity of cases did not require investigation of this featureof the optimiser.

Evaluation tools

Quantitative evaluation of plans was performed by means ofstandard Dose-Volume Histogram (DVH). For PTV, the values ofD98% and D2% (dose received by the 98%, and 2% of the volume)were defined as metrics for minimum and maximum doses in asso-ciation to V95% V107% (the volume receiving at least 95% or at most

E. Vanetti et al. / Radiotherapy and Oncology 92 (2009) 111–117 113

107% of the prescribed dose). The target homogeneity was ex-pressed by D5% � D95% (difference between the dose covering 5%and 95% of the PTV). The degree of conformality of the plans wasmeasured with a Conformity Index, CI95%, which is defined as theratio between the patient volume receiving at least 95% of the pre-scribed dose and the volume of the PTV. For OARs, the analysis in-cluded the mean dose, the maximum dose expressed as D2% and aset of appropriate VX and DY values. For Healthy Tissue, the integraldose, ‘‘DoseInt”, is defined as the integral of the absorbed dose ex-tended to over all voxels excluding those within the target volume(DoseInt dimensions are Gy � cm3). This was reported togetherwith the observed mean dose and some representative VxGy value.

The average cumulative DVH for PTV, OARs and healthy tissuewere built from the individual DVHs obtained by averaging thecorresponding volumes at each dose bin (0.01 Gy in this case). Toappraise the difference between the techniques, the paired, two-tailed Student’s t-test was applied. Data were considered statisti-cally significant for p < 0.05.

Results

Dose distributions are shown for one example in Fig. 1 for axial sag-ittal and coronal views. Fig. 2 shows the average DVH for all the PTVscomparing the three techniques for the entire patients cohort. Fig. 3reports theaverage DVHcomputed for the spinal cord, brainstem, ipsi-and contra-lateral parotids and for the healthy tissue. Fig. 4 shows theaverage DVH for selected organs at risks showing the potential differ-ence between IMRT and RapidArc in particular cases.

Tables 1 and 2 report numerical findings from DVH analysis onPTV, on the main organs at risk and for the healthy tissue. Table 3summarises results for some complementary organs at risk and forsmall subgroups of patients.

Data are presented as averages over the investigated patients,and errors indicated inter-patient variability at standard deviationlevel 1.

Fig. 1. Dose distributions on axial, coronal and

Target coverage and dose homogeneity

To simplify the reading of the analysis, data in the tables are re-ported expressing values as percentages of the dose prescribed to eachPTV (e.g. for PTV60, 100% corresponds to 60 Gy), while graphs in the fig-ures are shown with a single normalisation to 100% to 66 Gy.

In general, all the techniques resulted in a similar target coverage.RA2 plans achieved the best homogeneity (D5% � D95%), while

RA1 resulted slightly inferior to both IMRT and RA2. The sametrend was observed for D2%, the maximum significant dose, whichwas inferior to the planning objective of 107%. None of the tech-niques reached in average the objective on minimum dose, andthe same trend was observed as before, with IMRT falling betweenRA2 and RA1.

RapidArc and IMRT showed to be equivalent in terms of confor-mity index CI95%. This parameter resulted in 1.7 ± 0.2 irrespectiveof the technique.

Spinal cordAll plans respected the planning objective of 46 Gy as maximum

dose to the spinal cord; RA2 allowed the largest sparing of spinalcord in terms of D2%.

Brain stemAs for the spinal cord, IMRT, RA1 and RA2 plans showed D2% infe-

rior to the planning objective of 50 Gy, and statistical significancewas observed between each technique. With RapidArc, additionalreduction of D2% was observed compared to IMRT of 11.7 Gy forRA1 and 13.4 for RA2. Similarly, RapidArc plans showed a furtherreduction of the volume of the brain stem irradiated at various doselevels, e.g., V20Gy is reduced of 10.4 Gy for RA1 and 12.4 Gy for RA2.

ParotidsThe analysis was carried out for ipsi- and contra-lateral parotids

separately and, as expected, larger sparing was observed for the con-

sagittal views for one representative case.

Fig. 2. Mean DVHs for the three PTVs for the global cohort of patients.

Fig. 3. Mean DVHs of the spinal cord, brain stem, parotids and healthy tissue for the global cohort of patients.

114 Volumetric arc modulation for head and neck radiation therapy

tra-lateral glands. Results reported refer to the entire glands, regard-less of the (eventual) degree of overlap with the various target vol-umes. The planning objectives on mean or median dose were ingeneral hard to achieve (met only for the median dose D50% of thecontra-lateral glands by RA1 and RA2 plans). In general, as in the case

of the dose delivered to 1/3 or 2/3 of the gland volumes, RA1 and RA2allowed to reduce the parotids involvement compared to IMRT witha more pronounced efficacy of RA2 compared to RA1. On mediandose, (D50%) RA2 allowed an additional sparing of 8.4 Gy and 6 Gyover IMRT in the contra- or ipsi-lateral glands.

Fig. 4. Mean dose volume histograms for selected complementary organs at risk.

E. Vanetti et al. / Radiotherapy and Oncology 92 (2009) 111–117 115

The analysis was repeated for the fraction of parotid glands notincluded in the target volumes, i.e. for the most superficial portionof the parotids, and in this case much stronger sparing was obviouslypossible. For the contra-lateral parotids, in average, 90% of the glandvolume was outside the PTV, while for the ipsi-lateral case the ratiowas reduced to 80%. For the entire patient’s cohort, median doses forIMRT, RA1 and RA2 were 36.7 ± 11.8, 26.3 ± 7.9, and 23.0 ± 8.5 in theipsi-lateral case and 28.0 ± 9.9, 27.2 ± 4.8, and 21.3 ± 3.2 in the con-tra-lateral case, respectively.

Special cases

Fig. 4 reports average dose volume histograms of the base of thetongue (six patients, two from each subgroup), the vocal apparatus(three patients, one hypo-pharynx and two oro-pharynx), the con-tra-lateral cochlea (six patients, one hypo-pharynx and five oro-phar-ynx), the constrictors (six patients, two hypo-pharynx, three oro-pharynx and two larynx) and the mandible (eight patients, one

hypo-pharynx, one larynx and six oro-pharynx). In these limited sub-groups of patients, these special organs were either not included oronly partially included in the target volumes. From the graphs andthe table, it is clear how RapidArc allowed, compared to IMRT, somereduction of mean dose. In particular, RA plans showed a reductionof V50% for the contra-lateral cochlea, the constrictors and the vocalapparatus.

Healthy tissue

The planning objectives for healthy tissue were not forma-lised in numerical terms but the strategy was to minimise itsinvolvement. In this respect, RA and IMRT presented similarshapes in the DVH of the healthy tissue. No statistically signifi-cant difference was observed between the two RapidArc groups.The integral dose was found to be improved with RA in compar-ison to IMRT with an average reduction of 7% (RA1 or RA2 vs.IMRT).

Table 1Summary of the dosimetric results for the three PTVs.

Parameter Objective [%] IMRT RA1 RA2 P

PTV66Mean [%] 100 100.0 ± 0.0 100.0 ± 0.0 100.0 ± 0.0 n/aD2% [%] 107 105.4 ± 1.1 106.2 ± 1.4 104.9 ± 1.2 a,b,cD98% [%] 95 92.4 ± 1.2 91.7 ± 1.6 93.2 ± 1.5 a,b,cD5% � D95% [%] – 9.9 ± 1.6 11.1 ± 2.4 8.8 ± 2.0 a,b,c

PTV60Mean [%] 100 100.3 ± 1.6 100.8 ± 1.3 100.7 ± 1.2 –D2% [%] 107 105.9 ± 2.0 106.4 ± 1.6 105.4 ± 1.6 CD98% [%] 95 92.3 ± 2.3 92.5 ± 2.3 93.6 ± 2.3 b,cD5% � D95% [%] – 10.5 ± 1.8 10.6 ± 2.0 8.7 ± 1.9 b,c

PTV54Mean [%] 100 100.0 ± 2.2 100.0 ± 1.3 99.5 ± 1.6 CD2% [%] 107 106.2 ± 3.4 106.4 ± 1.7 105.2 ± 1.5 CD98% [%] 95 92.1 ± 2.8 91.8 ± 2.7 92.1 ± 2.7 –D5% � D95% [%] – 10.5 ± 2.2 10.6 ± 2.3 9.0 ± 2.1 b,c

Statistical significance (p < 0.05) is reported between couples from paired t-testanalysis; a: IMRT vs RA1, b: IMRT vs RA2, c: RA1 vs RA2.

Table 3Summary of the dosimetric results for special organs at risk not included in theoptimisation process.

Organ Patients Parameter IMRT RA1 RA2

Vocalapparatus

3/29 Mean [Gy] 59.8 ± 4.8 54.5 ± 6.6 54.8 ± 8.3V50Gy [%] 92.6 ± 12.1 75.7 ± 27.2 73.0 ± 33.4

Base oftongue

6/29 Mean [Gy] 44.1 ± 21.1 38.8 ± 20.9 39.5 ± 21.1V50Gy [%] 49.5 ± 40.4 36.5 ± 39.1 41.3 ± 38.4

Mandible 8/29 Mean [Gy] 35.9 ± 6.9 36.9 ± 6.1 33.8 ± 7.4D2% [Gy] 63.8 ± 5.5 62.9 ± 6.3 63.2 ± 6.7

Contralateralcochlea

6/29 Mean [Gy] 22.8 ± 27.6 20.9 ± 22.6 14.2 ± 16.8V50Gy [%] 30.5 ± 47.3 16.1 ± 30.0 5.9 ± 13.1

Constrictors 8/29 Mean [Gy] 51.8 ± 13.2 48.2 ± 14.1 48.1 ± 13.9V50Gy [%] 75.7 ± 25.6 66.5 ± 28.0 66.3 ± 27.8

116 Volumetric arc modulation for head and neck radiation therapy

Monitor units and delivery time

The number of MU per fraction of 2.2 Gy resulted to be MU/frIMRT = 1126 ± 333, MU/frRA1 = 463 ± 80 (41% of MU for IMRT),and MU/frRA2 = 584 ± 89 (52% of MU for IMRT). For RapidArc, allindividual arcs could be delivered between 70 and 90 s of beam-on time. IMRT plans showed values of MUs at least roughly dou-bled compared to RapidArc and given the multiple field arrange-ment and the presence of split fields, the delivery time for IMRTis significantly higher since it includes dead times such as the timeneeded to reposition the gantry and to re-program the linac atevery field, giving an overall time for IMRT of �15 min.

Discussion

This study reports on a comparison of the volumetric modu-lated arc therapy, RapidArc technique, with single or doublecoplanar arcs against fixed beam IMRT for head and neck cancerpatients. Similar investigations have been published in the recentpast. Palma et al. investigated RapidArc progenitor on prostateshowing that variable dose rate volumetric arc modulation isbeneficial compared to IMRT or constant dose rate [19]. Cozziet al. [17] and Fogliata et al. [18]. appraised the behaviour ofRapidArc on cervix uteri cancer and on small benign brain tu-mours. In those studies, RapidArc proved to be at least equiva-lent to IMRT in terms of target coverage while showed benefit

Table 2Summary of the dosimetric results for spinal cord, brain stem, parotids and healthy tissue

Organ Parameter Objectives [Gy]

Spinal cord Mean [Gy] –D2% [Gy] 46

Brain stem Mean [Gy] –D2% [Gy] 50

Ipsi-lateral parotid Mean [Gy] <26D50% [Gy] <30D33% [Gy] –D66% [Gy] –

Contra-lateral parotid Mean [Gy] <26D50% [Gy] <30D33% [Gy] –D66% [Gy] –

Healthy tissue Mean [Gy] –V10Gy [%] –DoseIntegral [104 Gy cm3] –

Statistical significance (p < 0.05) is reported between couples from paired t-test analysis

in organs at risk sparing. The planning case selected for thisinvestigation was the head and neck. Since it is a demandingindication, it proved to be an ideal indication for IMRT and dif-ferent strategies have been applied to improve OARs sparing.

Compared to IMRT, only RapidArc with two arcs, RA2, allowed aslight improvement in target dose homogeneity and coverage. BothRA1 and RA2 resulted in a systematic reduction of irradiation ofspinal cord, brain stem and parotids with statistically significant dif-ferences in most of the cases. In addition, single or double RapidArcallowed on average an additional sparing of those organs at risk thatare relevant for quality of life and for important acute and late toxic-ity [15,16] (as base of tongue, vocal apparatus, constrictors and co-chlea) but are normally not subject to optimisation objectives evenin the absence of explicit constraints in the optimisation. Relevanceof the findings of the present study should be validated through aproper clinical protocol that measures toxicity endpoints and con-trol rates (also including the uncertainties derived from fractionateddelivery) but this is obviously beyond the scope of a comparativeplanning study. Some investigations have been performed withexperimental dosimetric measurements to assess reliability of Rap-idArc delivery and its reproducibility, and the first results indicatethat RapidArc does not differ from IMRT in this respect [27,28].

Major limitations of, and strategies to minimise biases in, com-parative studies, such as the present one, have been extensivelydiscussed elsewhere (e.g. [17]) and shall be similarly applied inthe present case. The planning rules were applied as similarly aspossible between techniques, dose calculation algorithms andevaluation tools were unified but, even if a lot of care is taken inminimising arbitrary elements, it is impossible to completely con-trol all potential sources of bias and their influence on plan resultsand comparisons (e.g. the optimisation performed, although usingthe same objectives, by different planners in different institutes).

.

IMRT RA1 RA2 P

30.8 ± 3.4 28.2 ± 3.7 25.3 ± 3.1 a,b,c42.8 ± 2.1 43.7 ± 4.1 39.0 ± 2.6 b,c13.1 ± 10.4 10.4 ± 8.4 9.9 ± 8.6 a,b38.2 ± 15.3 26.5 ± 16.9 24.8 ± 16.3 a,b,c40.1 ± 11.6 36.2 ± 10.8 34.4 ± 11.1 a,b,c40.4 ± 13.8 34.8 ± 14.3 32.0 ± 15.2 a,b,c51.0 ± 12.0 46.4 ± 13.3 44.7 ± 14.5 a,b,c30.1 ± 15.6 25.0 ± 13.8 22.2 ± 14.6 a,b,c32.6 ± 8.4 30.9 ± 7.7 28.2 ± 6.8 b,c30.1 ± 10.4 28.4 ± 9.1 24.1 ± 7.5 a,b,c41.9 ± 10.5 38.6 ± 12.1 36.0 ± 12.1 a,b,c21.2 ± 10.8 20.1 ± 6.3 15.5 ± 4.0 b,c12.2 ± 2.9 11.5 ± 2.4 11.4 ± 2.3 a,b33.1 ± 8.4 30.9 ± 8.0 31.0 ± 8.0 a,b

9.4 ± 3.4 8.7 ± 2.2 8.7 ± 2.2 a,b

; a: IMRT vs RA1, b: IMRT vs RA2, c: RA1 vs RA2.

E. Vanetti et al. / Radiotherapy and Oncology 92 (2009) 111–117 117

These limitations are, in general, the common features of compar-ative planning studies, particularly when different optimisationengines or algorithms are considered. It should also be mentionedthat, for both techniques, no effort was made to push the optimisa-tion to the maximum achievable but rather the systems were onlypushed to reach the planning objectives when possible. Any furtherimprovement shown in the data was obtained ‘‘per se’’ by the algo-rithms without external driving forces. It is possible that both RAand IMRT might lead to additional sparing of organs with tighterobjectives but this was not the scope of the present investigation.

It is debatable whether single or multiple arcs should be applied torealise proper volumetric modulated arc techniques. Pioneers appliedseveral arcs to generate modulation patterns with multiple intensity lev-els [5–8]; alternative commercial solutions to RapidArc will allow multi-ple arcs and non-coplanar arrangement of arcs. A definitive appraisal ofthis issuecannotbeexploited withsinglestudiessincethe needfor ‘’com-plex’’ modulation patterns depends on several factors, mostly linked tothe clinical indication. Results showed that multiple arcs can improveboth the sparing of organs at risk and target coverage. Nevertheless, itis not per se evident if there is a general clinical relevance of furtherOAR sparing between single and double arcs in terms of expected toxicityor if, as it is likelythecase,multiplearcs should beappliedas asolutionforgiven critical cases but not as a general standard as single arcs may be lar-gely sufficient in most of the cases.

One of the objectives of RapidArc was the capability to delivertreatments in short times. For the cases under investigation,beam-on time was estimated to be less than 1.5 min per arc andconsiderably less than IMRT delivery times, which are mostly dom-inated by the large (seven to nine) number of fixed gantry fields.Literature results from other IMAT approaches [7,8,29] show deliv-ery time ranging from 6.3–14 min for C-arm linac based techniquesto �11 min for helical tomotherapy depending on the clinical indi-cation. Reduced beam-on and effective treatment times may have astrong impact on the clinical throughput, on the individual man-agement of patients and on the ability to perform systematic imageguidance on large groups of patients.

Conclusions

RapidArc was investigated for head and neck cancer, and RA2 led tosignificant sparing of organs at risk and healthy tissue sparing withuncompromised target coverage compared to conventional IMRTand RA1. The potential benefit of the better physical dose distributionis combined with the shorter delivery time and smaller number ofMUs required. These facts might have an impact on both biologyand logistic issues, but whether these reductions translate into clinicalbenefits will be answered by the clinical outcome studies.

Disclosure

Dr. L. Cozzi acts as a Scientific Advisor to Varian Medical Sys-tems, and is the Head of Research and Technological Developmentto Oncology Institute of Southern Switzerland, IOSI, Bellinzona.

Acknowledgements

The Phase II randomised trial mentioned in the introduction ispart of a Research Cooperation Agreement between TMC and Var-ian Medical Systems. The current investigation was also partiallycovered by a Varian research grant to IOSI.

References

[1] Cameron C. Sweeping-window arc therapy: an implementation of rotationalIMRT with automatic beam-weight calculation. Phys Med Biol2005;50:4317–36.

[2] Cotrutz C, Kappas C, Webb S. Intensity modulated arc therapy (IMAT) withcentrally blocked rotational fields. Phys Med Biol 2000;45:2185–206.

[3] Earl M, Shepard D, Naqvi S, Li X, Zu CX. Inverse planning for intensitymodulated arc therapy using direct aperture optimization. Phys Med Biol2003;21:1075–89.

[4] MacKenzie MA, Robinson DM. Intensity modulated arc deliveriesapproximated by a large number of fixed gantry position sliding windowdynamic multileaf collimator fields. Med Phys 2002;29:2359–65.

[5] Yu CX. Intensity-modulated arc therapy with dynamic multileaf collimation:an alternative to tomotherapy. Phys Med Biol 1995;40:1435–49.

[6] Yu CX, Li XA, Ma L, et al. Clinical implementation of intensity-modulated arctherapy. Int J Radiat Oncol Biol Phys 2002;53:453–63.

[7] Duthoy W, De Gersem W, Vergote K, Botenberf T, Derie C, Smeets P, et al.Clinical implementation of intensity modulated arc therapy (IMAT) for rectalcancer. Int J Radiat Oncol Biol Phys 2004;60:794–806.

[8] Duthoy W, De Gersem W, Vergote K, Coghe M, Botenberf T, De Deene Y, et al.Whole abdominal radiotherpay (WAPRT) using intensity modulated arctherapy (IMAT): first clinical experience. Int J Radiat Oncol Biol Phys2003;57:1019–32.

[9] Otto K. Volumetric modulated arc therapy: IMRT in a single arc. Med Phys2008;35:310–7.

[10] Gregoire V, De neve W, Eisbruch A, Lee N, Van den Weyngaert D, Van Gestel D.Intensity modulated radiation therapy for head and neck carcinoma. TheOncologist 2007;12:555–64.

[11] Lee N, Puri D, Blanco AI, Chao KS. Intensity modulated radiation therapy inhead and neck cancers: an update. Head Neck 2007;29:387–400.

[12] Popovtzer A, Eisbruch A. Advances in radiation therapy of head and neckcancer. Expert Rev Anticancer Ther 2008;8:633–44.

[13] Li Y, Ten Haken R, Eisbruch A. The impact of dose on parotid recovery in headand neck cancer patients treated with radiation therapy. Int J Radiat Oncol BiolPhys 2007;67:660–9.

[14] Murdoch-Kinch C, Kim H, Vineberg K, Ship J, Eisbruch A. Dose effectrelationships for the submandibular salivary glands and implications fortheir sparing by intensity modulated radiotherapy. Int J Radiat Oncol Biol Phys2008;72:373–82.

[15] Feng F, Kim H, Lyden T, Feng M, Chepeha D, Eisbruch A. Intensity modulatedradiotherapy of head and neck cancer aiming to reduce dysphagia: early dose-effect relationships for the swallowing structures. Int J Radiat Oncol Biol Phys2007;68:1289–98.

[16] Bhandare N, Antonelli P, Morris C, Malayapa R, Mendenhall W. Ototoxicityafter radiotherapy for head and neck tumors. Int J Radiat Oncol Biol Phys2007;67:469–79.

[17] Cozzi L, Dinshaw KA, Shrivastava SK, Mahantshetty U, Engineer R, DeshpandeDD, et al. A treatment planning study comparing volumetric arc modulationwith RapidArc and fixed field IMRT for cervix uteri radiotherapy. RadiotherOncol 2008;89:180–91.

[18] Fogliata A, Clivio A, Nicolini G, Vanetti E, Cozzi L. Intensity modulation withphotons for benign intracranial tumours. A planning comparison of volumetricsingle arc, helical arc and fixed gantry techniques. Radiother Oncol 2008 [e-pub].

[19] Palma D, Vollans E, James K, Nakano S, Moiseenko V, Shaffer R, et al.Volumetric modulated arc therapy for delivery of prostate radiotherapy.Comparison with intensity modulated radiotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys2008;72:996–1001.

[20] Bragg C, Windate K, Conway J. Clinical implications of the anisotropicanalytical algorithm for IMRT treatment planning and verification. RadiotherOncol 2008;86:276–84.

[21] Fogliata A, Vanetti E, Albers D, et al. On the dosimetric behaviour of photondose calculation algorithms in the presence of simple geometricheterogeneities: comparison with Monte Carlo calculations. Phys Med Biol2007;52:1363–85.

[22] Knöös T, Wieslander E, Cozzi L, et al. Comparison of dose calculationalgorithms for treatment planning in external photon beam therapy forclinical situations. Phys Med Biol 2006;51:5785–807.

[23] Ulmer W, Pyyry J, Kaissl W. A 3D photon superposition/convolution algorithmand its foundation on results of Monte Carlo calculations. Phys Med Biol2005;50:1767–90.

[24] Chui C, LoSasso T, Spirou S. Dose calculation for photon beams with intensitymodulation generated by dynamic jaw or multileaf collimations. Med Phys1994;21:1237–44.

[25] Convery D, Rosenbloom M. The generation of intensity-modulated fields forconformal radiotherapy by dynamic collimation. Phys Med Biol1992;37:1359–74.

[26] Spirou S, Chui C. A gradient inverse planning algorithm with dose-volumeconstrains. Med Phys 1998;25:321–33.

[27] S. Korreman, J. Medin, F. Kjaer-Kreistoffersen. Dosimetric verification ofRapidArc treatment delivery. Acta Oncol 2008 [epub].

[28] Nicolini G, Vanetti E, Clivio A, Fogliata A, Korreman S, Bocanek J, et al. TheGLAaS algorithm for portal dosimetry and quality assurance of RapidArc, anintensity modulated rotational therapy. Radiat Oncol 2008;3:24.

[29] Bijdekerke P, Verellen D, Tournel K, Vinh-Hung V, Somers F, Biesman P, et al.Tomotherapy: implications on daily workload and scheduling patients.Radiother Oncol 2008;86:224–30.