Preview

PDF, English - main document Download (11MB) | Terms of use

Abstract

Intensity modulated proton therapy (IMPT) plans precisely balance thousands of proton beamlets, giving high dose to the tumor while trying to spare healthy tissues. However, plan quality is affected by factors including: 1) dose calculation inaccuracies, 2) underestimation of the biological effect of the dose in sensitive areas and geometrical changes like 3) patient movement or 4) changes in posture and anatomy. All these factors are addressed in the projects here presented.

Project 1, in collaboration, introduces an upgraded version of a Monte Carlo package for graphics processing units (GPU-MC) to provide fast and accurate dose calculations. This package is extended to serve as the unique dose calculation engine in the following projects. Project 2, in collaboration, presents a prioritized optimization method to reduce the potential biological effect of the radiation in organs at risk near the tumor.

Project 3 compares computationally efficient strategies to take into account the patient respiratory motion by defining planning target volumes based on a 4DCT of the patient. Density overwrites considering water-equivalent-path-length to voxels across the 4DCT targets works best.

Project 4 demonstrates an online algorithm that maintains IMPT plan quality through treatment, adapting it to the daily patient posture and anatomy using GPU-MC calculations.