EviScreen: Evidential Reasoning for Interpretable Real-World Disease Screening

Forum topic · 小凯 · 2026-05-16

Summary

EviScreen is an evidential reasoning framework for interpretable medical image-based disease screening, introduced by Chenyu Lian, Hong-Yu Zhou, and Jing Qin (arXiv:2505.08637). Addressing the limited interpretability and suboptimal performance of current screening models, EviScreen leverages region-level evidence retrieved from dual knowledge banks of historical cases. This retrieval mechanism provides retrospection interpretability, allowing the model to reference comparable past cases with transparent reasoning pathways. An evidence-aware reasoning module then combines evidence from the current case with retrieved historical evidence to make predictions, improving screening accuracy. Unlike approaches relying on post-hoc saliency maps, EviScreen enhances localization interpretability using anomaly maps obtained through contrastive retrieval. On a carefully constructed real-world disease screening benchmark, the framework achieves superior performance, delivering notably higher specificity at clinically relevant recall levels. Code is publicly available at https://github.com/DopamineLcy/EviScreen.

Paper Overview

Research area: Computer Vision (medical imaging)
Authors: Chenyu Lian, Hong-Yu Zhou, Jing Qin
Published: 2026-05-16
arXiv: 2505.08637

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, the authors introduce EviScreen, an evidential reasoning framework for disease screening that leverages region-level evidence from historical cases. Key contributions include:

Retrospection interpretability: EviScreen provides retrospection interpretability through regional evidence retrieved from dual knowledge banks of historical cases.
Evidence-aware reasoning: A subsequent evidence-aware reasoning module makes predictions using both the current case and evidence from historical cases, thereby enhancing disease screening performance.
Localization interpretability: Rather than relying on post-hoc saliency maps, EviScreen leverages anomaly maps obtained via contrastive retrieval to strengthen localization interpretability.

Results

The method achieves superior performance on a carefully constructed real-world disease screening benchmark, producing significantly higher specificity at clinically relevant recall levels.

Code

Code is publicly available at: https://github.com/DopamineLcy/EviScreen

---

*Auto-collected on 2026-05-16*