Enhancing Organ at Risk Segmentation with Improved Deep Neural Networks

TL;DR

This paper proposes architectural improvements to U-Net for organ at risk segmentation, significantly increasing accuracy and generalization across multiple datasets in radiotherapy planning.

Contribution

The study introduces simple yet effective modifications to U-Net's architecture, loss function, and optimization to enhance its segmentation performance.

Findings

Achieved an average of 80% dice score across datasets

Improved U-Net performance from 63% to 80% dice score

Validated improvements on multiple publicly available datasets

Abstract

Organ at risk (OAR) segmentation is a crucial step for treatment planning and outcome determination in radiotherapy treatments of cancer patients. Several deep learning based segmentation algorithms have been developed in recent years, however, U-Net remains the de facto algorithm designed specifically for biomedical image segmentation and has spawned many variants with known weaknesses. In this study, our goal is to present simple architectural changes in U-Net to improve its accuracy and generalization properties. Unlike many other available studies evaluating their algorithms on single center data, we thoroughly evaluate several variations of U-Net as well as our proposed enhanced architecture on multiple data sets for an extensive and reliable study of the OAR segmentation problem. Our enhanced segmentation model includes (a)architectural changes in the loss function, (b)optimization framework, and (c)convolution type. Testing on three publicly available multi-object segmentation data sets, we achieved an average of 80% dice score compared to the baseline U-Net performance of 63%.