S822 I. J. Radiation Oncology d Biology d Physics Volume 78, Number 3, Supplement, 2010

Results: Twin plans comparison showed a number of points satisfying the tolerance of 1% in both dose and volume differencegreater than 99% for all structures’ curves composing the DVHs. The use of transition volumes improved dose homogeneity atthe junction level within ± 10% of the prescribed dose. In vivo measurements confirmed the uniformity of entrance dose deliveredover the abutment region within the same level of homogeneity.

Conclusions: The proposed technique has the feature to provide a ‘‘full helical’’ dose distribution for TMI treatments which can beconsidered truthful as long as dose identity between LTMI and tLTMI plans exists. Dose homogeneity in the overlapping area re-sults thus evaluable and, if necessary, improved by acting on the inverse planning constraints assigned to the transition volumes.The delivery of TMI with HT for the whole body maintaining an optimal degree of dose homogeneity in the abutting region iseffective by using the proposed planning technique.

Author Disclosure: M. Zeverino, None; S. Agostinelli, None; G. Taccini, None; F. Cavagnetto, None; S. Garelli, None; M. Gusinu,None; S. Vagge, None; S. Barra, None; R. Corvo, None.

3382 Are 2D Detector Arrays Sufficient for VMAT Quality Assurance?

F. Chen, M. Rao, J. Ye, J. Wu, T. Wong, J. Saini, V. Mehta, D. Shepard, D. Cao

Swedish Cancer Institute, Seattle, WA

Purpose/Objective(s): Volumetric modulated arc therapy (VMAT) is a fully dynamic delivery technique that necessitates a rig-orous quality assurance (QA) program. A key to VMAT QA is the appropriate selection of the QA device to be used in patientspecific VMAT verifications. We have evaluated both two dimensional (2D) ion chamber arrays and 2D diode arrays as toolsfor VMAT QA. Specifically, we studied the angular dependency at the individual detector level and investigated the potential gainsfrom correcting for the dependency.

Materials/Methods: A MatriXX (IBA Inc.) and a MapCheck-I (Sun Nuclear Inc.) served as our 2D ion chamber and diode arrays,respectively. To examine the angular dependency, the arrays were irradiated on an Elekta Synergy using square fields from 72 gantryangles. The measured 2D dose distributions were then compared to those generated in the treatment planning system, Pinnacle3. 2Dratio maps were generated by dividing the measured and the planned dose distributions. VMAT plans for 10 cases were examined.

Results: For the ion chamber array, the correction ratio was .1 (101�104%) for irradiations from the anterior beam angles, but\1(96�99%) for the posterior beam angles. In the case of the diode array, the correction ratio was between 102 and 107% at mostgantry angles. The variation between individual detectors was more significant in the diode array.For most VMAT cases, the use of the ion chamber array resulted in a close agreement between the planned and delivered doses. Anangular correction was not needed due to the relatively small correction factors and the fact that the correction factors from opposedbeam angles largely canceled each other out. Greater discrepancies were observed when there was a more asymmetric distributionof the monitor units.. Angular correction using only the ratio from the CAX detector has been found to be helpful for some caseswith gamma passing rate increased from 86% to 95%. The use of a 2D correction algorithm proved sufficient for all cases when theion chamber array was used. The use of a diode array poses additional challenges. First, the correction ratio was found to be de-pendent on the field size of the square fields used to create them and the aperture shape. Larger field size yielded a higher ratio. Asecond concern is the known responses of diode to dose rate, e.g., higher dose rate causing higher diode reading. Although the QApassing rate for some cases was as high as 95%�98% without angular dependency correction, this is in part due to the averaging outof various measurement errors. Consequently there is a risk that errors may be masked out.

Conclusions: Caution should be taken in using a diode array for VMAT QA due to the field size, aperture shape and dose rateeffect. The angular dependency of an ion chamber array can be correct for in a more straightforward manner.This work is supported by an grant from Elekta.

Author Disclosure: F. Chen, None; M. Rao, None; J. Ye, None; J. Wu, None; T. Wong, None; J. Saini, None; V. Mehta, None; D.Shepard, None; D. Cao, None.

3383 Dynamic Jaws and Dynamic Couch Tomotherapy for High Risk Prostate Cancer

J. Lee, M. Michaletz-Lorenz, J. Michalski, S. M. Goddu

Washington University in St. Louis, St. Louis, MO

Purpose/Objective(s): At present helical TomoTherapy (HT) is limited to fixed jaw widths and a constant couch speed (FJCC).The next generation of HT features dynamic motion of both jaws and couch termed Dynamic Jaws and Dynamic Couch (DJDC).We dosimetrically compared multifield IMRT (mf-IMRT) with FJCC and DJDC techniques for differences in PTV coverage andsparing of organs at risk (OARs) in high risk prostate cancer.

Materials/Methods: Three high risk prostate cancer patients originally planned using mf-IMRT targeting the prostate, seminalvesicles (SV) and pelvic lymph nodes (LN) were retrospectively planned with FJCC (2.5 cm field width) and DJDC techniqueson TomoTherapy’s research platform. Segmentation, dose, and fractionation (1.8 Gy) were kept identical for each patient. PTV45Gy

covered the prostate, SV and pelvic LN. PTV77.4Gy covered the prostate and SV. Dose volume histograms were compared for PTVcoverage and dosing to OARs, with analysis by ANOVA and t-test.

Results: PTV45Gy was 846 ± 35 mL and PTV77.4Gy was 81 ± 12 mL. Values below are listed in the order of mf-IMRT, FJCC, DJDC,respectively, unless otherwise noted. V45Gy for PTV45Gy were 99.8 ± 0.1%, 99.7 ± 0.2% and 99.2 ± 0.3%. V77.4Gy for PTV77.4Gy were99.3 ± 0.5%, 97.9 ± 0.7% and 98.8 ± 0.2%. Homogeneity expressed as (D5%-D95%)/77.4Gy was 6.1 ± 0.3% for FJCC vs. 18.7 ± 8.5% forDJDC. Rectal V40Gy were 37.3 ± 2.8, 36.4 ± 9.0, and 34.4 ± 8.6Gy. Rectal V65Gy were 14.1 ± 1.2, 12.5 ± 0.9, and 12.7 ± 0.7Gy. Rectummean doses (MD) were 36.3 ± 2.9, 22.8 ± 2.9, 21.1 ± 2.4Gy. Bladder V40Gy were 51.7 ± 6.2, 45.9 ± 7.8, 38.6 ± 4.9Gy. Bladder V65Gy

were 18.7 ± 4.3, 18.8 ± 3.4, 19.5 ± 3.4Gy. Bladder MD were 45.2 ± 3.1, 29.4 ± 2.1, 30.3 ± 1.9Gy. Penile bulb V50Gy was negligible in allgroups. Penile bulb D70% was 7.7 ± 1.8, 19.9± 11.5, 7.2 ± 1.4Gy. Penilebulb D90% was 4.1 ± 1.8, 15.8± 9.3, 6.6 ± 1.3Gy. Penile bulb MDwere 13.0 ± 3.4, 15.6 ± 5.9, 5.4 ± 1.0Gy (p = 0.305). Bowel V50Gy were 2.9 ± 2.9, 20.9 ± 19.9, and 29.5 ± 16.6 mL. Bowel MD were 24.0± 3.9, 22.4 ± 2.9, 21.7 ± 3.6Gy. Femoral head (FH) V50Gy were 0.12 ± 0.06%, 0.21 ± 0.09%, 0.49 ± 0.31%. MD to FHs were 21.3 ± 1.7,12.7 ± 0.6 and 14.0 ± 0.4 Gy (p\0.001 for both HT types vs. IMRT). PTV45Gy treatment times for FJCC and DJDC were 371 ± 17 and253 ± 5 seconds per fraction (p = 0.023). PTV77.4Gy treatment times for FJCC and DJDC were 224 ± 4 and 180 ± 9 seconds (p = 0.049).

Proceedings of the 52nd Annual ASTRO Meeting S823

Conclusions: DJDC provided target coverage while sparing bladder, rectum, and bowel similarly to mf-IMRT and FJCC. BothDJDC and FJCC reduced dose to FHs compared to IMRT. DJDC appeared to reduce the penile bulb MD compared to IMRT,but in this study it was not statistically significant. Treatment time for DJDC is 32 and 20% faster compared to FJCC for initialpelvic and boost fields, respectively.

Author Disclosure: J. Lee, None; M. Michaletz-Lorenz, None; J. Michalski, None; S.M. Goddu, None.

3384 Does the Choice of Isotope, Co60 or Ir192, Affect Treatment Planning Techniques and Outcomes for High

Dose Rate (HDR) Brachytherapy?

A. Palmer B. Mzenda

Portsmouth Hospitals NHS Trust, Portsmouth, United Kingdom

Purpose/Objective(s): The long half-life of Co60 compared to Ir192 has economic and practical benefits, however the affect of thehigher energy of Co60 on treatment planning approaches and outcomes in gynecological HDR brachytherapy has not been fullyinvestigated. This work compares the physical source characteristics of the two isotopes used in the IBt Bebig MultiSource after-loading system, and investigates differences in planning techniques to optimize tumor coverage and critical organ (OAR) sparingfor the two isotopes.

Materials/Methods: Monte Carlo data was used to determine the dose differences from Co60 and Ir192 sources due to anisotropy,radial dose and geometric effects. The high risk clinical target volume (HR-CTV), bladder, rectum and sigmoid were outlined onCT images of 10 gynecological patients. Eight 3D treatment plans were produced for each patient on the HDRplus treatment plan-ning system (TPS). The planning variables used in this study were: the choice of isotope, Co60 or Ir192; the choice of dose pre-scribing and normalization (to Manchester point A or GEC-ESTRO HR-CTV); standard and non-standard loading patterns; andTPS algorithm dwell-time optimization. Quantitative analysis of the resulting treatment plans was performed using isodose distri-butions, dose volume histograms, HR-CTV D90 and V100%, doses to point A, ICRU rectum and bladder, and D0.1cc, D1cc andD2cc for the OARs.

Results: Comparing treatment plans generated using Co60 and Ir192, the isodose distributions showed typical shifts of up to 1.2mm posteriorly and 3.5 mm superiorly between the two isotopes, as a result of anisotropy effects. Evaluation of doses to OARsacross the range of treatment plans showed that the choice of normalization/prescription technique and the optimization parametersused were more significant than the physical differences between the two isotopes, with HR-CTV D90 being near-identical in allcases. For the patients considered in this study, normalizing to point A resulted in up to 5.0% increase to OARs from Co60 com-pared to Ir192. However, renormalizing to the HR-CTV resulted in higher OAR doses from Ir192 compared to Co60 by up to 3.6%.

Conclusions: Despite significant differences in physical characteristics, particularly energy and resulting source anisotropy, ourresults show that optimization and application of appropriate planning techniques achieves clinically comparable plans whenCo60 and Ir192 are used in 3D image based HDR brachytherapy.

Author Disclosure: A. Palmer, None; B. Mzenda, None.

3385 Dosimetric Benefit of Non-coplanar VMAT Delivery

D. Cao, F. Chen, M. Rao, J. Ye, T. Wong, V. Mehta, D. M. Shepard

Swedish Cancer Institute First Hill, Seattle, WA

Purpose/Objective(s): A key feature of VMAT is the ability to deliver non-coplanar arcs. Although the current implementations ofVMAT have focused on axial coplanar delivery, non-coplanar arcs may bring additional dosimetric benefits for specific treatmentsites. This work was carried out to study non-coplanar VMAT delivery for both intra-cranial and extra-cranial targets in terms ofplan quality and delivery efficiency.

Materials/Methods: Three GBM cases were selected for this study as examples for intra-cranial targets. Three extra-cranial caseswith treatment sites of lung, prostate and head-and-neck were also included in this study. The Pinnacle3 SmartArc inverse planningalgorithm was used to generate all the VMAT plans. A single arc was used in the axial coplanar VMAT plan for the five relativelysimple cases, while two arcs were used for the H&N case. While for the non-coplanar VMAT plans, 3 to 4 arcs were used for eachcase with different couch angles and arc paths. Plan quality was compared between the two sets of data in terms of target dosecoverage and OAR sparing. VMAT deliveries on an Elekta Synergy linac were also performed to compare the delivery efficiency.

Results: For all six cases, the target dose coverage in both coplanar and non-coplanar VMAT plans remained essentially un-changed. As to the OAR sparing, the non-coplanar plans showed significant improvement for the cases with intra-cranial targets.For example, on average, the maximum and mean doses to brainstem were reduced from 3277 and 1016 cGy to 2524 and 668 cGyin the non-coplanar plans. These correspond to a 23% and 34% reduction. The average maximum dose to the chiasm was alsoreduced from 876 cGy to 346 cGy in the non-coplanar plan, a 61% reduction. For extra-cranial cases, however, there are no con-sistent improvements in OAR sparing for the non-coplanar plans. For the lung case, the maximum cord dose was reduced from 926to 579 cGy in the non-coplanar plan at the cost of the mean esophageal dose increasing from 484 to 521 cGy. For the prostate case,mean dose to the femoral heads decreased from 3389 to 2304 cGy in the non-coplanar plan, which was the only significant dif-ference between the two plans. For the H&N case, the max dose to the cord and brainstem were reduced by 8.3% at the cost ofthe mean dose to these two structures increasing by 14.3% in the non-coplanar plan. The average treatment delivery time increasedfrom 1.9 to 4.5 minutes for the non-coplanar plans.

Conclusions: For intra-cranial cases, the use of non-coplanar VMAT can significantly improve the OAR sparing while main-taining the target dose coverage. For extra-cranial cases, only modest gains in OAR sparing were observed particularly for struc-tures that are in close proximity to the targets. For all cases, these dosimetric benefits come at the cost of a prolonged treatmentdelivery time.

Author Disclosure: D. Cao, Elekta, B. Research Grant; F. Chen, Elekta, B. Research Grant; M. Rao, Elekta, B. Research Grant; J.Ye, Elekta, B. Research Grant; T. Wong, Elekta, B. Research Grant; V. Mehta, Elekta, B. Research Grant; D.M. Shepard, Elekta, B.Research Grant.