Evidential Reasoning Advances Interpretable Real-World Disease Screening

2026-05-14

AI Summary

This work addresses the limitations of existing medical image screening models, which often suffer from poor interpretability, suboptimal performance, and an inability to effectively leverage historical cases for transparent reasoning. To overcome these challenges, the authors propose EviScreen, a novel framework that constructs dual knowledge bases and introduces a region-level evidence retrieval mechanism to fuse current imaging data with relevant evidence from historical cases through evidence-aware reasoning. Furthermore, EviScreen employs a contrastive learning–driven approach to generate anomaly maps, enabling retrospective interpretability without relying on post-hoc saliency maps. Evaluated on real-world disease screening benchmarks, EviScreen achieves clinical-grade high recall while significantly improving specificity, thereby delivering both strong performance and enhanced interpretability.

Abstract

Disease screening is critical for early detection and timely intervention in clinical practice. However, most current screening models for medical images suffer from limited interpretability and suboptimal performance. They often lack effective mechanisms to reference historical cases or provide transparent reasoning pathways. To address these challenges, we introduce EviScreen, an evidential reasoning framework for disease screening that leverages region-level evidence from historical cases. The proposed EviScreen offers retrospection interpretability through regional evidence retrieved from dual knowledge banks. Using this evidential mechanism, the subsequent evidence-aware reasoning module makes predictions using both the current case and evidence from historical cases, thereby enhancing disease screening performance. Furthermore, rather than relying on post-hoc saliency maps, EviScreen enhances localization interpretability by leveraging abnormality maps derived from contrastive retrieval. Our method achieves superior performance on our carefully established benchmarks for real-world disease screening, yielding notably higher specificity at clinical-level recall. Code is publicly available at https://github.com/DopamineLcy/EviScreen.

Problem

• disease screening
• interpretability
• medical images
• evidential reasoning
• historical cases

Innovation

• evidential reasoning
• interpretable AI
• disease screening
• contrastive retrieval
• knowledge bank