Treatment planning optimization based on response-surface modeling of cost function versus multiple constraints

摘要: In treatment planning optimization and inverse planning, a solution is governed by operator-specified objectives constraints, which are usually in conflict with each other. Currently, the process involves manual iterations of testing different combinations constraints and/or objectives, evaluating resulting solutions. this work, authors developing more efficient paradigm to automate process. The key component their approach model cost be optimized as function all competing constraints. Initially evaluated technique for optimizing angles weights coplanar beams conformal therapy pancreas prostrate cancer. minimum tumor dose was maximized multiple dose-volume other critical structures. An initial set established such that fell within desired limits. Subsequently, active constraint changed small step rerun obtain new solution. Effectively, sampled space obtained database solutions variety values. They then modeled multi-variable doses structures using response-surface method. This allowed visualize how competed affected dose. Using information, one can interactively navigate through quickly balance among focuses on clinical considerations during improves efficiency