Leveraging Pathology Co-occurrence for Test-Time Adaptation in Chest X-Ray Diagnosis

2026-07-04

AI Summary

This work addresses the performance degradation of medical imaging models when deployed in new clinical settings due to shifts in imaging devices, acquisition protocols, or patient populations. Existing test-time adaptation methods often overlook the co-occurrence dependencies among multiple pathologies. To bridge this gap, the authors propose Co-occurrence Weighted Adaptation (CoWA), the first approach to incorporate disease co-occurrence structure into test-time adaptation. CoWA estimates a pathology co-occurrence matrix from model predictions, identifies samples that deviate from expected co-occurrence patterns, and dynamically down-weights them during adaptation, combined with entropy minimization for robustness. Evaluated on a chest X-ray domain shift benchmark, CoWA significantly outperforms existing methods, demonstrating that leveraging co-occurrence information effectively enhances both the generalization and adaptation stability of diagnostic models.

Abstract

Medical imaging models often degrade when deployed at new clinical sites due to differences in imaging equipment, protocols, and patient populations. Test-time adaptation (TTA) addresses this by updating a pretrained model using only unlabeled target data, without access to source data. However, existing TTA methods were designed for single-label classification on natural image benchmarks, minimizing entropy uniformly across all samples without considering label dependencies. This overlooks a key property of multi-label medical imaging: pathologies do not occur independently but exhibit structured co-occurrence patterns. In this work, we propose Co-occurrence Weighted Adaptation (CoWA), which leverages disease co-occurrence patterns as a reliability signal for adaptation. CoWA estimates label co-occurrence structure from model predictions and downweights samples that deviate from expected patterns, enabling adaptation to rely more on consistent predictions while reducing the impact of noisy ones. We evaluate CoWA on chest X-ray benchmarks under domain shifts and demonstrate consistent improvements over established baselines.

Problem

• test-time adaptation
• chest X-ray diagnosis
• pathology co-occurrence
• domain shift
• multi-label medical imaging

Innovation

• test-time adaptation
• disease co-occurrence
• multi-label medical imaging
• domain shift
• chest X-ray diagnosis