Histopath-C: Towards Realistic Domain Shifts for Histopathology Vision-Language Adaptation

2026-01-18

AI Summary

This study addresses the significant performance degradation of vision-language models on digital histopathology images caused by real-world domain shifts such as staining variations, contamination, blur, and noise. To this end, the authors introduce Histopath-C, the first dynamic evaluation benchmark tailored for histopathology, which realistically simulates distributional shifts through synthetic perturbations. They further propose LATTE, a test-time adaptation strategy specifically designed for medical vision-language models, integrating low-rank adaptation with multi-text template fusion to enhance robustness. Extensive experiments demonstrate that LATTE substantially outperforms state-of-the-art test-time adaptation methods—originally developed for natural images—across multiple histopathology datasets, confirming its effectiveness and generalization capability in real-world clinical settings.

Technology Category

Computer Vision: Large Vision ModelsMachine Learning: Transfer, Domain Adaptation, Multi-Task LearningNatural Language Processing: (Large) Language Models

Application Category

Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for rankingGraph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphsUser Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Medical Vision-language models (VLMs) have shown remarkable performances in various medical imaging domains such as histo\-pathology by leveraging pre-trained, contrastive models that exploit visual and textual information. However, histopathology images may exhibit severe domain shifts, such as staining, contamination, blurring, and noise, which may severely degrade the VLM's downstream performance. In this work, we introduce Histopath-C, a new benchmark with realistic synthetic corruptions designed to mimic real-world distribution shifts observed in digital histopathology. Our framework dynamically applies corruptions to any available dataset and evaluates Test-Time Adaptation (TTA) mechanisms on the fly. We then propose LATTE, a transductive, low-rank adaptation strategy that exploits multiple text templates, mitigating the sensitivity of histopathology VLMs to diverse text inputs. Our approach outperforms state-of-the-art TTA methods originally designed for natural images across a breadth of histopathology datasets, demonstrating the effectiveness of our proposed design for robust adaptation in histopathology images. Code and data are available at https://github.com/Mehrdad-Noori/Histopath-C.

Problem

• domain shift
• histopathology
• vision-language models
• image corruption
• medical imaging

Innovation

• Histopath-C
• Test-Time Adaptation
• Vision-Language Models
• Domain Shift
• Low-Rank Adaptation