Squeeze-and-Excitation Normalization for Automated Delineation of Head and Neck Primary Tumors in Combined PET and CT Images

TL;DR

This paper introduces a novel segmentation method using Squeeze-and-Excitation Normalization within a U-Net architecture, achieving top performance in automated head and neck tumor delineation on PET/CT images.

Contribution

It presents a new normalization technique integrated into a deep learning model for improved tumor segmentation in medical images.

Findings

Achieved state-of-the-art segmentation accuracy in the HECKTOR challenge

Outperformed competing methods with DSC of 0.759 on test set

Won first prize among 21 teams in the challenge

Abstract

Development of robust and accurate fully automated methods for medical image segmentation is crucial in clinical practice and radiomics studies. In this work, we contributed an automated approach for Head and Neck (H&N) primary tumor segmentation in combined positron emission tomography / computed tomography (PET/CT) images in the context of the MICCAI 2020 Head and Neck Tumor segmentation challenge (HECKTOR). Our model was designed on the U-Net architecture with residual layers and supplemented with Squeeze-and-Excitation Normalization. The described method achieved competitive results in cross-validation (DSC 0.745, precision 0.760, recall 0.789) performed on different centers, as well as on the test set (DSC 0.759, precision 0.833, recall 0.740) that allowed us to win first prize in the HECKTOR challenge among 21 participating teams. The full implementation based on PyTorch and the trained models are available at https://github.com/iantsen/hecktor