Program Information
Fast Computational Optimization Method for Rotating-Shield Brachytherapy Treatment Planning
M Cho1*, X Wu1,2 , H Dadkhah3 , R Flynn2 , Y Kim2 , W Xu1 , (1) Department of Electrical and Computer Engineering, University of Iowa, Iowa City, IA, (2) Department of Radiation Oncology, University of Iowa, Iowa City, IA, (3) Department of Biomedical Engineering, University of Iowa, Iowa City, IA,
Presentations
MO-F-FS1-4 (Monday, July 31, 2017) 4:30 PM - 6:00 PM Room: Four Seasons 1
Purpose: To provide a new, fast computational method for brachytherapy treatment planning, by deriving a first-order convex optimization algorithm based on the proximal graph solver (POGS). The proposed POGS was validated using volumetric, inverse optimization procedure for rotating shield brachytherapy (RSBT) but is also applicable to clinical HDR and IMRT optimization.
Methods: The proposed POGS algorithm was validated by two different RSBT approaches for five MRI-guided HDR cervical cancer patients whose high-risk CTV was > 40 cc: multi-helix RSBT (H-RSBT) and single-shield RSBT (S-RSBT). Treatment plans were generated for all patients using POGS method and compared to the previously used commercial IBM ILOG CPLEX Optimization solver, in terms of the computational time and plan quality. The rectum, bladder, sigmoid colon, HR-CTV, and HR-CTV boundary were regions-of-interests considered in the asymmetric dose-volume optimization with smoothness control. Dose calculation resolution was 1mm x 1mm x 3mm for all cases. The H-RSBT applicator had 6 helices, with 33:3 mm of translation along the applicator per helical rotation and 1.7 mm spacing between dwell positions, yielding 17.5 degree emission angles per 5 mm along the applicator. The S-RSBT used 45 degree, 22.5 degree and 5 mm for azimuthal angle, rotation angle, and distance between adjacent dwell positions.
Results: POGS was around 20 times faster than CPLEX for both H-RSBT and S-RSBT. For all patients, total optimization times were 35.6-99.6 seconds for CPLEX and 2.0-3.6 seconds for POGS. In terms of plan quality, HR-CTV D90, HR-CTV D100, rectum D2cc, sigmoid D2cc, and bladder D2cc for all plans matched within 1% for CPLEX and POGS regardless of S-RSBT and H-RSBT. Also, we obtained similar EQD2 figures between CPLEX and POGS.
Conclusion: POGS substantially reduced volumetric, inverse optimization time by around 20 times for both H-RSBT and S-RSBT within less than 1% dosimetric differences, compared with CPLEX.
Funding Support, Disclosures, and Conflict of Interest: This work is supported by NIH 1R01EB020665.
Contact Email: