Program Information
Reduced Order Prioritized Optimization for IMRT Planning
L Rivera1*, E Yorke2, A Kowalski3, R Radke4, A Jackson5, (1)Rensselaer Polytechnic Institute, Troy, NY, (2) Memorial Sloan-Kettering Cancer Center, New York, NY, (3) Memorial Sloan Kettering Cancer Center, New York, NY, (4) Rensselaer Polytechnic Institute, Troy, NY, (5) Mem Sloan-Kettering Cancer Ctr, New York, NY
MO-A-137-9 Monday 8:00AM - 9:55AM Room: 137Purpose:
A prioritized optimization scheme is introduced to improve the previously proposed Reduced Order Constrained Optimization (ROCO) method for generating clinically competitive IMRT plans. The improvement eliminates manual repetition of the ROCO constrained optimization stage to automatically produce an IMRT plan with target coverage equivalent to and organ-at-risk (OAR) doses lower than those in the original plan.
Methods:
The prioritized optimization consists of three stages. The first is creation of a baseline ROCO plan with initial constraints for all the OARs based on the clinical protocol and the planner's expertise. This process involves constrained optimization over a low-dimensional parameter space and is able to directly satisfy requested dose limits. In the second stage, certain critical OAR dose constraints are selected to be further reduced, and the constrained optimization phase of ROCO is re-applied. These constraints are reduced simultaneously by a constant dose until PTV coverage metrics (D95) are violated (maintaining all other constraints from the previous stage). The third stage sequentially pushes each individual OAR constraint on a priority list until PTV D95s are violated (maintaining all other constraints from previous stages).
Results:
The prioritized optimization scheme was tested on the simulation image sets of 10 head-and-neck patients. The prioritized ROCO scheme automatically improved OAR protection with equivalent target coverage (i.e., 6.5% average lower dose in cord, 7.7% in brainstem, 4.4% in the left parotid and 5.4% in the right parotid) automatically in comparison to the first-stage plans. While 26 minutes of constrained optimizations on average were required to obtain these superior plans, no human interaction is involved until plan evaluation.
Conclusion:
After initial site- and patient-specific decisions by the planner, the proposed prioritized optimization method runs automatically and produces superior plans than the original ROCO method, avoiding intermediate manual evaluation.
Funding Support, Disclosures, and Conflict of Interest: This publication was supported in part by Grant Number 1R01CA148876-03 from the National Cancer Institute (NCI). Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute, National Institutes of Health.
Contact Email: