MO‐A‐137‐02: Comparative Efficacy of Image‐Guided Adaptive Treatment Strategies for Prostate Radiation Therapy Via Virtual Clinical Trials

Purpose: To determine potential dosimetric benefits of image guided adaptive radiation therapy treatment (IGART) strategies for intermediate risk prostate cancer via automated virtual clinical trials (VCTs). Materials and Methods: A 19 patient cohort with 8‐13 CT images was used to compare different IGART strategies. The IMRT prescription was 46 Gy/23 fractions to the prostate and seminal‐vesicle PTVs, followed by a 40 Gy/20 fractions prostate boost. For each patient, daily IGART imaging was simulated by random selection from available images. The VCTs simulated three different IGART‐IMRT strategies; (A0) initial planning only with 5 mm PTV margins; (A1) daily re‐planning without considering prior dose; and (A2) online daily re‐planning for each fraction considering prior dose obtained via deformable dose mapping (A1 and A2, no PTV margin). For each strategy, daily dose was deformably mapped using Demons‐based displacement vector fields and accumulated to estimate the treatment dose. Strategies were compared via dosimetric adherence to constraints (CTV D98=86.4, D2<94.6Gy), and objectives (CTV D98<86.4, D2<91.2Gy; bladder D2<90, D20<70, D30<56, D50<39, EUD=0 Gy; and rectum D2<83, D20<70, D30<56, D50<39, EUD=0 Gy). Results: A0 had larger doses to 20, 30, and 50% rectal (30±20%) and bladder (40±20%) volumes than A1 and A2 (p<0.001). Rectal and bladder D2s were also respectively higher in A0 by (15±13%) and (24±14%) (p=0.002 and 0.00001 respectively). Comparing A1 and A2, no significant differences were found in prostate D98 (p=0.4) or bladder D2 (p=0.33). Significant differences were found for the rectum (D5,A2>D5A1, p=0.02). Conclusion: IGART utilizing daily re‐planning (A1 and A2) has dosimetric advantages over conventional IMRT for critical structures, particularly high‐dose regions, without compromising PTV coverage. Due to inherent deformation vector field inaccuracies, daily re‐planning based on prior dose (A2) showed poor rectal sparing than A1. NIH grant P01 CA 116602 and Philips Medical Systems.