S714 I. J. Radiation Oncology d Biology d Physics Volume 75, Number 3, Supplement, 2009

Materials/Methods: We present a global to local 3D shape registration framework between the planning 3D CT images and treat-ment 3D CBCT images. The global registration is based on maximizing mutual information and the local registration is based a B-spline based incremental free form deformation model to minimize a sum of squared differences measure. The strengths of thismethod are that it preserves shape topology during local deformation, develops a novel 3D distance map algorithm and producessmooth, continuous and one-to-one correspondence local registration fields. The method is applied on 7 datasets consisting of bothplanning CT and CBCT images of breast patients. The target volumes which are used in the registration method in the planning 3DCT images and treatment 3D CBCT images are delineated by radiation oncologists manually and are chosen as the gold standards.

Results: The registration method is evaluated using both qualitative and quantitative methods. The qualitative method is based onvisualization by radiation oncologists. The quantitative methods are based on distance-based estimators and volume overlap ratio tocompare the differences between the gold standard of the treatment CBCT images and the registered planning images. The quan-titative results show that between the gold standard of the treatment CBCT images and the registered planning images, the meandiscrepancy in distance ranges from 1.21 mm to 2.32 mm, the root-mean-square error (RMSE) ranges from 1.35 mm to 2.54 mmand the mean breast volume overlap ratio ranges from 83.2% to 94%.

Conclusions: We present a novel and fast deformable registration method to capture the transformation between the planning andtreatment images for external beam radiotherapy. We also present promising results of our method applied to clinical datasets.These preliminary results show that the proposed method is robust and reasonably fast for the registration of the deformablesoft tissue of breast, and for deriving the clinical target volume on the treatment 3D CBCT images from the planning 3D CT imagesduring the partial breast treatment.

Author Disclosure: J. Zhou, None; S. Goyal, None; S. Kim, None; A. Khan, None; S. Jabbour, None; M. Rao, None; B. Haffty,None; N.J. Yue, None.

3184 Inverse Treatment Planning for Volumetric Modulated Arc Therapy using Compressed Sensing

Techniques

L. Zhu, L. Xing

Stanford University, Stanford, CA

Purpose/Objective(s): The volumetric modulated arc therapy (VMAT) includes the field angle as another degree of freedom in thetreatment planning and is therefore able to achieve a superior dose performance. Due to the increased computational complexityand the hardware constraints of beam segments, however, an optimal inverse planning of the VMAT is yet to be found and itsadvantages are not fully exploited in practice. The conventional beamlet-based algorithms obtain high-complexity fluence mapswhich are not suitable for delivery in the VMAT, and the segment-based algorithms are computationally intensive with no guar-antee on the optimality of the final solution. In this work, a novel VMAT inverse planning algorithm is proposed to overcome thedrawbacks of the existing methods.

Materials/Methods: The main idea of the proposed algorithm stems from the fact that the VMAT planning problem is highlydegenerated. Without compromising the final dose performance, the search space of the optimized solution can be significantlyreduced. A new VMAT planning algorithm is then designed using compressed sensing, a method which is commonly used inthe field of signal processing to recover compressible signals. We first divide the projection fields of VMAT into several groupsof fields as in fix-gantry intensity-modulated radiation therapy (IMRT). Each group is then optimized separately using a previouslydeveloped IMRT planning algorithm with total-variation regularization. The derived segments are added into the solution space ofthe final VMAT plan. Another optimization is then carried out in the generated solution space using L-1 norm regularization on theweights of the segments to find a sparse solution.

Results: The proposed algorithm is able to achieve a small number of segments for each projection field. After leaf sequencing, theoptimized plan is deliverable using different schemes such as one single arc, multiple full or partial arcs. The algorithm performanceis demonstrated using a prostate patient study. As compared to the fixed gantry radiation therapy and the conformal Arc therapy, ourapproach achieves a much improved target conformity and dose sparing of the sensitive structures, and truly gains the advantages ofthe VMAT.

Conclusions: The compressed sensing techniques reduce the size of the optimization problem without sacrificing the optimality ofthe final solution, which leads to two major advantages of our VMAT planning algorithm. As compared to beamlet-based algo-rithms, our approach generates a plan with a small number of segments, which is more readily deliverable in VMAT. As comparedto segment-based algorithms, our method has much higher efficiency due to its reduced size of solution space. As arc-based radi-ation therapy becomes more commonly used, the proposed algorithm is very attractive in clinic.

Author Disclosure: L. Zhu, None; L. Xing, None.

3185 Dosimetric Effects of Non-symmetric Openings of Multi-catheter Breast Brachytherapy Applicators

J. E. Reiff1, D. C. Perlingiero2, L. T. Komarnicky1

1Drexel University College of Medicine, Philadelphia, PA, 2Hahnemann University Hospital, Philadelphia, PA

Purpose/Objective(s): The Strut Adjusted Volume Implant (SAVITM) applicator is an HDR device used for accelerated partialbreast irradiation (APBI). It is designed with multiple catheters, or struts, forming an ellipsoid around a central lumen. Occasionallythe applicator may not open completely or symmetrically resulting in inverted or splayed struts. To investigate the resulting do-simetric consequences, we have run plans which simulate these abnormalities.

Materials/Methods: A CT scan was performed on all four sizes of the SAVITM applicator. In each case the PTV was defined as thevolume extending from the periphery of the applicator to 1.0 cm from the ellipsoid formed by the fully opened struts; the prescrip-tion dose of 3.40 Gy was prescribed to the outer edge of the PTV. Three plans were run off of each scan. The first plan was for thecomplete and symmetric opening of the struts. The second plan excluded one of the struts in the dose optimization simulating aninverted or splayed strut. The third plan excluded 2 adjacent struts, thus simulating two struts splayed away from each other. Foreach plan, the following 8 parameters were studied: the maximum dose encompassing the PTV (D100), V100, V150, V200, V250, V300,V350, and the volume outside the PTV receiving at least the prescribed dose (Vout).

Proceedings of the 51st Annual ASTRO Meeting S715

Results: D100 was greatest for each applicator when it opened symmetrically, ranging between 93.2% and 98.8% of the prescribeddose. When a single strut was excluded from the plan, D100 decreased to between 90.6% and 93.2% of the prescribed dose. D100

remained unchanged when the second strut was removed from the plan except for the largest applicator where it fell to 87% of theprescribed dose. V100 ranged from 96.7% to 99.8% of the PTV over all 12 plans with less than 2% variation observed as asymmetrywas added. V150 and V200 were maximized when the applicator opened symmetrically and decreased as the asymmetry in eachapplicator increased. For the largest applicator, these values fell from 49.8 cm3 to 45.4 cm3, and from 19.5 cm3 to 15.1 cm3, re-spectively. Of course, these values are lower for the smaller applicators. The V250, V300, and V350 values all showed the same pat-tern as asymmetry was introduced into the dosimetric optimization. Finally, Vout either decreased or remained the same as theasymmetry of the applicator increased, indicating that hotspots outside the PTV did not increase as the applicator became increas-ingly asymmetric.

Conclusions: Due to the differential loading capability of the struts as well as the central lumen, no increase in hotspots eitherinside or outside the PTV are observed due to the introduction of applicator asymmetry. In fact, in all the cases studied, the hotspotswere reduced in the asymmetrically optimized applicators while maintaining an acceptable dose to the PTV.

Author Disclosure: J.E. Reiff, Cianna Medical, F. Consultant/Advisory Board; D.C. Perlingiero, None; L.T. Komarnicky,None.

3186 Feasibility Study of Volumetric Modulated Arc Therapy for Malignant Pleural Mesothelioma Tumor

L. Cozzi1, M. Bignardi2, A. Fogliata1, P. Lattuada2, P. Mancosu2, P. Navarria2, G. Urso3, S. Vigorito3, M. Scorsetti2

1Oncology Institute Southern Switzerland, Bellinzona, Switzerland, 2Istituto Clinico Humanitas, Milano, Italy, 3Istituto ClinicoHumanitas, Milano, Italy

Purpose/Objective(s): Malignant pleural mesothelioma (MPM) is an aggressive tumor with disastrous prognosis. Many RT pro-cedures, were evaluated to cover the large and irregular target and to limit the dose at the nearness risk structures. In this work thefeasibility of volumetric modulated arcs therapy with RapidArc (RA) technique for adjuvant radiotherapy treatment of MPM wasevaluated in comparison with fixed field IMRT.

Materials/Methods: A retrospective analysis was performed on 6 patients affected by MPM and enrolled for adjuvant RT (3 left/right sides). The CTV was defined as the region of resection and other areas considered at high risk. The PTV was delineated byuniform margins of 5 mm around CTV. The plans were normalized at PTV mean dose, dose prescription was 54 Gy, with 2 Gy/fraction. For RapidArc two coplanar full arcs using 18 MV were used, whereas 9 fields with mix 6 and 18 MV energies were appliedto IMRT plans. Planning objectives were: PTV: V95%.95%, controlateral lung: V5Gy\60%, V20Gy\10%, mean dose\9.5Gy,Liver: V30Gy\33%, mean dose\31Gy; Heart: V45Gy\30%, V50Gy\20%, DMax\60Gy; Spinal cord: V45Gy\1%, DMax\50Gy,Contralateral kidney: V15Gy\20%, Esophagus: V55Gy\30%, Spleen: V40Gy\50%. The plans were compared by analyzing targetcoverage, homogeneity and normal tissue sparing.

Results: Both IMRT and RA met the target objective on CTV (V95%=95.3±8.5 for IMRT and 98.7±1.4 for RA) and approached iton PTV (V95%=93.3±7.7 and 93.5±3.4). Target homogeneity resulted equivalent between RA (D5%-D95%=5.9±0.8) and IMRT(D5%-D95%=6.0±1.5). All the objectives for OARs were met for both techniques: contralateral mean lung doses were 5.8±0.8and 5.6±0.7 Gy respectively for IMRT and RA, with V5Gy=45.0±12.6% and V20Gy=2.0±0.9% for IMRT, V5Gy =47.9±7.4%and V20Gy=0.4±0.4% for (RA). Heart resulted with IMRT V45Gy=15.6±4.0% and V50Gy=6.1±1.5%, while with RAV45Gy=19.2±11.5% and V50Gy=12.0±7.2%. Liver V30Gy was 15.1±12.8% and 12.8±14.0% with IMRT and RA, respectively.Spleen had V40Gy=39.4±6.3% with IMRT and 40.5±7.4% with RA. Esophagus V55Gy was 8.2±14.4% and 4.4±9.2% withIMRT and RA. V15Gy to contralateral kidney, as well as V45Gy to spine presented 0% volume for both techniques. For healthytissue, integral dose resulted equivalent between the two techniques. Concerning delivery parameters, RapidArc resulted in reduc-ing total monitor units 2195±317 for IMRT and 734±82 for RA.

Conclusions: The dosimetric findings are significantly favorable for RapidArc and, from the clinical perspective, these data couldincrease the population candidate to RT of MPM after radical surgery. An upcoming Phase I trial on patients will start to evaluatethe clinical outcomes with RapidArc.

Author Disclosure: L. Cozzi, Scientific Advisor to Varian Medical Systems, G. Other; M. Bignardi, None; A. Fogliata, None; P.Lattuada, None; P. Mancosu, None; P. Navarria, None; G. Urso, None; S. Vigorito, None; M. Scorsetti, None.

3187 Helical Tomotherapy for Debulking Irradiation Prior to Stem Cells Transplantation in Malignant

Lymphoma

C. Chargari1, J. Tamburini2, S. Zefkili1, M. Fayolle3, A. Fourquet1, Y. Kirova1

1Institut Curie, Paris, France, 2Hopital Cochin, Paris, France, 3Hopital d’Instruction des Armees du Val-de-Grace, Paris,France

Purpose/Objective(s): Preliminary clinical experiences suggested that radiation therapy (RT) may be effectively incorporated intocondition therapy before transplant for patients with refractory / relapsed malignant lymphoma. We investigated the planning anddelivery of debulking selective lymph nodes irradiation prior to autogenic and or allogenic stem cells transplantation (SCT), usingHelical Tomotherapy (HT).

Materials/Methods: Three consecutive patients with refractory malignant lymphoma were referred to our Institute. All patientshad been previously heavily treated, but had tumor persistence despite chemotherapy (CT) intensification. Pt #1 was a 23-year oldfemale, with stage IV Hodgkin lymphoma (HD). At time of HT, she presented with multiple lung and pleural nodes and residualbreast tumor despite 4 lines CT and one autogenic SCT. The only area which did not experience recurrence was previously irra-diated axilla. Pt #2 was a 59-year old female with aggressive B- large cell lymphoma of rhino pharynx and pt #3 was a 54-year oldmale with HD of rhino pharynx. At time of RT, both of them presented with bulky bilateral cervical lymph node attempt and tumorprogression despite several lines of CT. All three patients were referred to debulking HT prior to salvage allogenic for the firstpatient and autogenic SCT for pts 2 & 3.