Program Information
A Fast Inverse Treatment Planning Strategy Optimizing Dosimetric Measures for Low-Dose-Rate Brachytherapy Including Catheter Optimization
C Guthier1*, D Cail1 , P Orio1 , J Hesser2 , R Cormack1 , (1) Department of Radiation Oncology, Brigham and Women's Hospital, Dana-Farber Cancer Institute, and Harvard Medical School, Boston, MA, (2) Department of Experimental Radiation Oncology, University Medical Center Mannheim, and Heidelberg University, Mannheim
Presentations
WE-AB-605-12 (Wednesday, August 2, 2017) 7:30 AM - 9:30 AM Room: 605
Purpose: Permanent low-dose-rate (LDR) brachytherapy for prostate cancer is associated with excellent biochemical control and minimal toxicity. Treatment outcome linked to the achieved dose-distribution. Plans are generated intraoperatively, where a computer-based initial plan is manually tailored to the patient’s anatomy. This study introduces an inverse treatment planning strategy based on dose-volume-histogram related dosimetric measures (DMs).
Methods: Current r LDR-ITP guarantee fast optimization times, but are known to be non-intuitive while producing plans requiring manual adjustments. We introduce a model that using automatically generated gland subvolumes and DMs. Further we adapted our optimizer MPIP to efficiently solve the resulting non-convex and constrained optimization problem. To increase the plan quality additional structures that are related to the inherent risk of harboring disease are added as additional targets. Under an IRB-approved study the applied plans (APP) of 25 patients (prostate volume 24-77cc) are compared to our proposed method without (manual catheter selection) (S1) and with catheter optimization (S2) using one set of optimization parameters.
Results: The optimization time was App:(20-30) minutes, S1:(13.4±6.3)s and S2:( 24.2±12.0)s. All generated plans fulfill the DMs for the OARs. The V100 of the prostate was found to be APP:(94.4±3.8)%, S1:(96.5±2.1)%, and S2:(96.6±1.3)%. Compared to APP, S1 and S2 show a statistically (Wilcoxon signed rank test) significant improvement (p<.001) in target coverage while S1 and S2 perform equally (p>.05). Based on achieved dose-distribution and dosimetric measures 23 of 25 plans were deemed acceptable for treatment after a single run of the optimizer.
Conclusion: Using clinical data we demonstrate that the proposed regional DM-based optimization allows intuitive and fast plan optimization. Without the need for additional manual planning, the proposed approach has the potential to significantly shorten planning time. Reduced planning time will also reduce the time in the operating room and the time a patient is under anesthesia.
Contact Email: