Automated Treatment Planning in Radiation Therapy using Generative Adversarial Networks

TL;DR

This paper introduces a GAN-based method for automated radiation therapy treatment planning that predicts high-quality 3D dose distributions without relying on site-specific features, improving clinical satisfaction and similarity metrics.

Contribution

The study presents a novel GAN approach that directly predicts 3D dose distributions, bypassing traditional feature engineering and low-dimensional plan representations.

Findings

Significantly outperforms previous methods on clinical satisfaction criteria

Achieves higher similarity metrics in dose distribution predictions

Demonstrates effectiveness on oropharyngeal cancer patient data

Abstract

Knowledge-based planning (KBP) is an automated approach to radiation therapy treatment planning that involves predicting desirable treatment plans before they are then corrected to deliverable ones. We propose a generative adversarial network (GAN) approach for predicting desirable 3D dose distributions that eschews the previous paradigms of site-specific feature engineering and predicting low-dimensional representations of the plan. Experiments on a dataset of oropharyngeal cancer patients show that our approach significantly outperforms previous methods on several clinical satisfaction criteria and similarity metrics.