Proceedings of the 50th Annual ASTRO Meeting S653

set of patients was estimated by an independent software (Syntegra Pinnacle).We analyzed isocenter translations in all three axes,the absolute magnitude of isocenter dislocations, and the shifts applied in CBCT group.

Results: The isocenter translations in bodyframe treated patients were 1.4 ± 1.4 mm, 1.3 ± 1.7 mm and 0.9 ± 1.8 mm respectivelyfor lateral (X), anterior-posterior (Y) and caudocranial (Z) and the absolute magnitude of isocenter dislocation was 3.0 ± 1.5 mm. Inthe group of patients treated frameless an average shift of 0.6 ± 2.2 m, 2.7 ± 3.7 mm and 0.2 ± 1.0 mm was applied respectively forlateral (X), anteroposterior (Y) and caudocranial (Z) directions, the absolute magnitude of isocenter dislocation was 4.3 ± 2.5 mm.The average rotational corrected deviation around Y axis was 2.3 ± 4.6�. No detectable rotation around the X and Z axes was ob-served fusing CBCT with CT Plan. The second CBCT always confirmed the correct overlapping with the CT Plan after applyingshifts.

Conclusions: Our data seem to confirm that the accuracy of HSRT using bodyframe is comparable with frameless treatment usingCBCT. The high quality of CBCT imaging leads to correct set up error\5 mm to obtain the same positioning of the planning CT.The use of wing-board and vacuum pillow as immobilization devices avoid large rotational errors which are difficult to correct.

Author Disclosure: S. Ursino, None; F. Fiorica, None; S. Lappi, None; S. Fabbri, None; C. Flammia, None; L. Perazzini, None; E.De Guglielmo, None; L. Vignoli, None; F. Cartei, None.

3075 Convergent Radiotherapy: Adaptive Treatment Planning with Delivered Dose Information

R. I. Berbeco, D. Ionascu, F. Hacker, C. Zatwarnicki, S. Park, D. O’Farrell, H. J. Mamon

Brigham & Women’s Hospital, Boston, MA

Purpose/Objective(s): Recent studies have shown the value of beam’s-eye-view imaging during radiotherapy to monitor the tu-mor location. This information can further be used to calculate the delivered dose, even in the presence of motion. ConvergentRadiotherapy is the replanning of a radiotherapy treatment, in-between fractions, with the cumulative delivered dose includedin the re-optimization. The goal is for the delivered dose to converge with the prescription. Although this is distinct from AdaptiveRadiotherapy (ART) in that the measured delivered dose is used for the re-optimization, principles of Adaptive Radiotherapy couldalso be incorporated in Convergent Radiotherapy. We have demonstrated the concept of Convergent Radiotherapy for an SBRTtreatment with poor dosimetric target coverage on the first day of treatment.

Materials/Methods: Exit radiation is passively acquired by operating an electronic portal imaging device (EPID) in cine mode.Anatomical or fiducial landmarks in the beam’s-eye-view cine EPID images are back-projected through the patient model and de-livered dose is calculated for each image frame. The total delivered dose is calculated for each fraction and also summed over theentire course of treatment. This information can be used for clinical delivered dose-treatment response studies, as a basis for sub-sequent courses of radiotherapy, for quality assurance and for Convergent Radiotherapy. For the latter, the delivered dose is cal-culated after each fraction and compared to the planned distribution. Regions of disagreement are taken into account in an iterativeadaptive treatment planning approach. Each iteration of re-planning and delivery brings the cumulative delivered dose closer to thatwhich was planned. The implementation of a corrective strategy like Convergent Radiotherapy becomes all the more importantduring a hypofractionated treatment, as there are fewer fractions with which to recover from an aberration.

Results: We found that a significant setup error on the first day of treatment could be compensated for on subsequent days such thatthe delivered dose coverage converged upon the prescribed coverage. Beginning with the original treatment plan, only slight mod-ifications were needed for each re-plan. Note that the same procedure may also be used for adjacent critical organs whose dosetolerances should not be exceeded. In this case extra avoidance on subsequent treatment days is only necessary when there isan over-coverage seen in the cumulative dose. Giving too little dose to critical structures will not be penalized.

Conclusions: The combination of beam’s-eye-view imaging, 3D delivered dose calculation and Convergent Radiotherapy willprovide a valuable clinical tool for ensuring that the patient receives the prescribed dose.

Author Disclosure: R.I. Berbeco, Varian Medical Systems, Inc., B. Research Grant; D. Ionascu, None; F. Hacker, None; C. Zat-warnicki, None; S. Park, None; D. O’Farrell, None; H.J. Mamon, None.

3076 Full 3D Dose Calculations for Total Body Irradiation: A Comparison Study between Treatment Planning

Systems in Homogeneous and Heterogeneous Conditions

M. Lavallee, L. Gingras, S. Aubin, C. Cote, M. Larochelle, M. Chretien, L. Beaulieu

CHUQ - Hotel-Dieu de Quebec, Quebec, QC, Canada

Purpose/Objective(s): To realize full 3D heterogeneous dose calculations for total body irradiation (TBI) cases and to comparedifferent treatment planning tools.

Materials/Methods: A retrospective dosimetric study was performed on 7 patients. Dose distributions obtained with Pinnacle3

v.7.9u (Philips Medical Systems) were compared with those calculated with our actual TBI planning system Theraplan Plus(TPP) by MDS Nordion/Nucletron. Two different Pinnacle3 models were studied: standard beam commissioning (std_Pinnacle3)and TBI commissioning (TBI_Pinnacle3). For the later case, TBI_Pinnacle3, the commissioning was adapted to the special con-ditions of TBI beam physics: extended SSD of 190cm, large field, acrylic beam spoiler, and out of field dose (OFD). Homogeneousand heterogeneous dose calculation conditions were also compared.

Results: Significant differences between TPPstd_Pinnacle3 and TBI_Pinnacle3 dose distributions were found. For relative homo-geneous mid-line doses, differences up to 12% were obtained and a systematic overestimation of 5% was found in patients extrem-ities (e.g., head and legs) between TPP and TBI_Pinnacle3. Average dose underestimation of 3% was observed betweenstd_Pinnacle3 and TBI_Pinnacle3. Differences in patient extremities are attributed to the OFD contribution which is not correctlycomputed in TPP and std_Pinnacle3. It was observed that this effect is worst for tall patients (i.e., adults) than for small patients(i.e., children). Dose comparison outside the patient mid-line showed greater differences (up to 20%) between models. Becauseonly TBI_Pinnacle3 takes into account the beam spoiler effect, other models underestimate shallow doses. An accurate 3D hetero-geneous dose calculation is possible only with specifically design beam model, the TBI_Pinnacle3 model. Results showed major