Cross-Domain Validation of a Resection-Trained Self-Supervised Model on Multicentre Mesothelioma Biopsies

2025-12-01

AI Summary

This study addresses the clinical challenge of scarce biopsy samples and high annotation costs in malignant mesothelioma subtype classification and prognosis prediction. We propose a cross-domain transfer learning paradigm: a self-supervised image encoder is pre-trained on abundant large resection specimens—requiring no additional annotations—and directly transferred to multi-center, small-scale biopsy datasets. To our knowledge, this is the first empirical demonstration that such a model achieves strong generalization on biopsy images and effectively extracts discriminative morphological features. In multi-center validation cohorts, the model achieves high-accuracy subtype classification (AUC ≥ 0.92) and significantly stratifies patient survival risk (hazard ratio = 3.17, *p* < 0.001). Our approach overcomes key limitations of self-supervised models—namely, dependence on large-scale labeled data—thereby substantially enhancing the practical utility and deployability of AI in real-world histopathological diagnosis.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Computer Vision: Medical and Biological Imaging
• Humans and AI: Human-in-the-loop Machine Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Accurate subtype classification and outcome prediction in mesothelioma are essential for guiding therapy and patient care. Most computational pathology models are trained on large tissue images from resection specimens, limiting their use in real-world settings where small biopsies are common. We show that a self-supervised encoder trained on resection tissue can be applied to biopsy material, capturing meaningful morphological patterns. Using these patterns, the model can predict patient survival and classify tumor subtypes. This approach demonstrates the potential of AI-driven tools to support diagnosis and treatment planning in mesothelioma.

Problem

• Validates self-supervised model on mesothelioma biopsies across domains
• Predicts patient survival and classifies tumor subtypes from biopsies
• Enables AI tools for diagnosis using small biopsy samples

Innovation

• Self-supervised encoder trained on resection tissue
• Model applied to biopsy material for pattern capture
• AI predicts survival and classifies tumor subtypes