A Systematic Analysis of Input Modalities for Fracture Classification of the Paediatric Wrist

TL;DR

This paper systematically evaluates how combining radiographs with additional modalities like bone segmentation, fracture location, and reports improves pediatric wrist fracture classification accuracy using deep learning.

Contribution

It introduces a comprehensive analysis of multiple input modalities, demonstrating their combined benefit over radiographs alone in fracture classification.

Findings

Adding modalities increases AUROC from 91.71 to 93.25

Combining data sources enhances classification performance

Code is publicly available on GitHub

Abstract

Fractures, particularly in the distal forearm, are among the most common injuries in children and adolescents, with approximately 800 000 cases treated annually in Germany. The AO/OTA system provides a structured fracture type classification, which serves as the foundation for treatment decisions. Although accurately classifying fractures can be challenging, current deep learning models have demonstrated performance comparable to that of experienced radiologists. While most existing approaches rely solely on radiographs, the potential impact of incorporating other additional modalities, such as automatic bone segmentation, fracture location, and radiology reports, remains underexplored. In this work, we systematically analyse the contribution of these three additional information types, finding that combining them with radiographs increases the AUROC from 91.71 to 93.25. Our code is available on GitHub.