Multi-Modality Optimal Radiation Therapy (MMORT)

This repository contains code for the paper A Proof of Principle: Multi-Modality Radiotherapy Optimization. We developed a non-convex bilevel optimization framework for combining different radiation modalities in optimal treatment planning for radiation therapy cancer patients. Multi-modality radiation therapy is not used in current clinical practice and this work strives to reinforce the growing interest in this idea by developing methods for the multi-modality treatment planning and providing a proof of concept as well as some real patient experiments (the latter is ongoing work).

Installation and Use

To clone the repository:

git clone https://github.com/LevinRoman/MMORT
cd MMORT

To run the experiments script

(here alpha, beta, gamma, delta_mean and delta_max are radiobiological parameters described in the paper):

cd mmort
python experiments.py --alpha 0.35 --beta 0.175 --gamma 0.35 --delta_mean 0.07 --delta_max 0.175

Project structure

mmort: Main directory with the scripts
- optimization_tools.py: script for fluence map u optimization
- experiments.py: main script for running the multi-modality optimization to find the optimal fractionation schedule as well as fluence maps for each radiation modality

License

This project is MIT licenced.