CheXRelNet: An Anatomy-Aware Model for Tracking Longitudinal Relationships between Chest X-Rays

TL;DR

CheXRelNet is a novel neural model that leverages anatomical information and visual features to accurately track disease progression between sequential chest X-rays, addressing a key challenge in medical imaging analysis.

Contribution

It introduces CheXRelNet, a model that integrates local/global features and anatomical dependencies for longitudinal change detection in chest X-rays.

Findings

Improved performance over baseline methods on Chest ImaGenome dataset

Effective modeling of anatomical relationships enhances change detection accuracy

Demonstrates potential for better disease progression monitoring in clinical settings

Abstract

Despite the progress in utilizing deep learning to automate chest radiograph interpretation and disease diagnosis tasks, change between sequential Chest X-rays (CXRs) has received limited attention. Monitoring the progression of pathologies that are visualized through chest imaging poses several challenges in anatomical motion estimation and image registration, i.e., spatially aligning the two images and modeling temporal dynamics in change detection. In this work, we propose CheXRelNet, a neural model that can track longitudinal pathology change relations between two CXRs. CheXRelNet incorporates local and global visual features, utilizes inter-image and intra-image anatomical information, and learns dependencies between anatomical region attributes, to accurately predict disease change for a pair of CXRs. Experimental results on the Chest ImaGenome dataset show increased downstream performance compared to baselines. Code is available at https://github.com/PLAN-Lab/ChexRelNet