MedicalPatchNet: A Patch-Based Self-Explainable AI Architecture for Chest X-ray Classification

2025-09-09

AI Summary

To address the limited interpretability of deep neural networks in chest X-ray classification—which hinders clinical trust—this paper proposes an intrinsically self-explanatory architecture. The method partitions input images into non-overlapping patches, independently encodes and classifies each patch using an EfficientNet-style backbone, and aggregates patch-level predictions. This design enables natural, post-hoc-free attribution of decisions to anatomical regions, mitigates shortcut learning, and enhances lesion localization. On CheXpert, the model achieves an AUROC of 0.907—comparable to EfficientNet-B0’s 0.908—while attaining a lesion localization hit rate of 0.485 on CheXlocalize, significantly surpassing the baseline (0.376). The core contribution is the unified realization of end-to-end interpretability and high performance, establishing a novel paradigm for clinically trustworthy AI.

Technology Category

• Computer Vision: Interpretability, Explainability, and Transparency
• Machine Learning: Deep Neural Architectures and Foundation Models
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Deep neural networks excel in radiological image classification but frequently suffer from poor interpretability, limiting clinical acceptance. We present MedicalPatchNet, an inherently self-explainable architecture for chest X-ray classification that transparently attributes decisions to distinct image regions. MedicalPatchNet splits images into non-overlapping patches, independently classifies each patch, and aggregates predictions, enabling intuitive visualization of each patch's diagnostic contribution without post-hoc techniques. Trained on the CheXpert dataset (223,414 images), MedicalPatchNet matches the classification performance (AUROC 0.907 vs. 0.908) of EfficientNet-B0, while substantially improving interpretability: MedicalPatchNet demonstrates substantially improved interpretability with higher pathology localization accuracy (mean hit-rate 0.485 vs. 0.376 with Grad-CAM) on the CheXlocalize dataset. By providing explicit, reliable explanations accessible even to non-AI experts, MedicalPatchNet mitigates risks associated with shortcut learning, thus improving clinical trust. Our model is publicly available with reproducible training and inference scripts and contributes to safer, explainable AI-assisted diagnostics across medical imaging domains. We make the code publicly available: https://github.com/TruhnLab/MedicalPatchNet

Problem

• Improving interpretability in chest X-ray classification
• Providing transparent decision attributions to image regions
• Mitigating shortcut learning risks for clinical trust

Innovation

• Patch-based self-explainable AI architecture
• Independent patch classification with aggregation
• Visualization without post-hoc techniques