Premium
TH‐E‐BRCD‐01: Automatic Treatment Planning ‐ an MCO Perspective
Author(s) -
Craft D
Publication year - 2012
Publication title -
medical physics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.473
H-Index - 180
eISSN - 2473-4209
pISSN - 0094-2405
DOI - 10.1118/1.4736347
Subject(s) - deliverable , plan (archaeology) , planner , computer science , pareto principle , perspective (graphical) , multi objective optimization , mathematical optimization , operations research , artificial intelligence , machine learning , systems engineering , engineering , mathematics , archaeology , history
We review three multi‐criteria optimization (MCO) approaches for automatic IMRT treatment planning: goal programming, prioritized optimization, and Pareto surface based MCO. After comparing and contrasting these approaches, we describe a hybrid approach where in step one — using any of those alternatives — a “first pass” plan is constructed. When this plan is presented to the treatment planner (or physician), the planner is allowed to navigate on a Pareto surface centered around that plan, allowing for a tuning of the plan. For this tuning step to be most useful, we require directly deliverable MCO (DMCO), by which we mean navigating on the Pareto surface in such a way that the plan you are currently viewing is ready‐to‐go from a delivery perspective (i.e. it does not need to be leaf sequenced first, the multi‐leaf collimator segmentation has already been done). We discuss current approaches to solving this DMCO problem, and finish with a discussion of how the above ideas apply to automatic volumetric modulated arc therapy (VMAT) planning. Learning objectives: 1. Understand three types of multi‐criteria optimization: goal programming, prioritized optimization, and Pareto surface based MCO 2. Understand the idea of Pareto surface navigation refinement of an automatically generated plan. 3. Understand directly deliverable navigation. Research sponsored in part by RaySearch Laboratories.