Explainable vertebral fracture analysis with uncertainty estimation using differentiable rule-based classification

TL;DR

This paper introduces an explainable deep learning method for vertebral fracture assessment that incorporates uncertainty estimation and aligns with clinical criteria, achieving high accuracy and reliability comparable to human experts.

Contribution

The method uniquely combines differentiable rule-based classification with uncertainty estimation for vertebral fracture analysis, improving interpretability and performance over prior approaches.

Findings

Achieves 93% vertebra-level sensitivity.

Attains 97% end-to-end AUC.

Model reliability comparable to human annotators.

Abstract

We present a novel method for explainable vertebral fracture assessment (XVFA) in low-dose radiographs using deep neural networks, incorporating vertebra detection and keypoint localization with uncertainty estimates. We incorporate Genant's semi-quantitative criteria as a differentiable rule-based means of classifying both vertebra fracture grade and morphology. Unlike previous work, XVFA provides explainable classifications relatable to current clinical methodology, as well as uncertainty estimations, while at the same time surpassing state-of-the art methods with a vertebra-level sensitivity of 93% and end-to-end AUC of 97% in a challenging setting. Moreover, we compare intra-reader agreement with model uncertainty estimates, with model reliability on par with human annotators.