Segmentation and Risk Score Prediction of Head and Neck Cancers in PET/CT Volumes with 3D U-Net and Cox Proportional Hazard Neural Networks

TL;DR

This study develops a combined deep learning and Cox model approach for tumor segmentation and risk prediction in head and neck cancers using PET/CT images, achieving high accuracy in segmentation and risk scoring.

Contribution

It introduces a novel loss function for improved segmentation and a combined neural network and Cox model for hazard risk prediction from radiomic features.

Findings

Segmentation achieved 0.82 Dice score and 3.16 Hausdorff Distance.

Risk prediction model achieved 0.89 c-index in validation.

Test set risk score c-index was 0.61.

Abstract

We utilized a 3D nnU-Net model with residual layers supplemented by squeeze and excitation (SE) normalization for tumor segmentation from PET/CT images provided by the Head and Neck Tumor segmentation chal-lenge (HECKTOR). Our proposed loss function incorporates the Unified Fo-cal and Mumford-Shah losses to take the advantage of distribution, region, and boundary-based loss functions. The results of leave-one-out-center-cross-validation performed on different centers showed a segmentation performance of 0.82 average Dice score (DSC) and 3.16 median Hausdorff Distance (HD), and our results on the test set achieved 0.77 DSC and 3.01 HD. Following lesion segmentation, we proposed training a case-control proportional hazard Cox model with an MLP neural net backbone to predict the hazard risk score for each discrete lesion. This hazard risk prediction model (CoxCC) was to be trained on a number of PET/CT radiomic features extracted from the segmented lesions, patient and lesion demographics, and encoder features provided from the penultimate layer of a multi-input 2D PET/CT convolutional neural network tasked with predicting time-to-event for each lesion. A 10-fold cross-validated CoxCC model resulted in a c-index validation score of 0.89, and a c-index score of 0.61 on the HECKTOR challenge test dataset.