1546. Distributionally Robust Chance Constraints for VMAT Treatment Planning
Invited abstract in session WA-11: Resource and treatment planning, stream OR in Healthcare (ORAHS).
Wednesday, 8:30-10:00Room: Clarendon SR 1.03
Authors (first author is the speaker)
| 1. | Houra Mahmoudzadeh
|
| Department of Management Sciences, University of Waterloo | |
| 2. | Stoyan Hristov
|
| University of Waterloo | |
| 3. | Johnson Darko
|
| Grand River Regional Cancer Centre | |
| 4. | Ernest Osei
|
| Grand River Regional Cancer Centre |
Abstract
Radiation therapy (RT) seeks to irradiate a cancerous tumour while minimizing damage to the nearby organs at risk (OARs). Throughout a treatment session, a patient's geometry might change unpredictably, which can degrade the treatment quality. Volumetric Modulated Arc Therapy (VMAT) is a modern form of RT in which the beam follows a path around the patient while continuously delivering radiation. Despite better OAR sparing and shorter treatment times, VMAT planning results in a large-scale nonlinear mixed integer program (NLMIP) that becomes even more complex when geometric uncertainty is incorporated. We propose a distributionally robust chance constraints VMAT model and a heuristic solution scheme that outputs near-optimal treatment plans which are robust to uncertainty in tumour position. Finally, we compare the robustness of nominal, robust, and distributionally robust plans and discuss tradeoffs.
Keywords
- Health Care
- Optimization Modeling
- Robust Optimization
Status: accepted
Back to the list of papers