A histomorphological atlas of resected mesothelioma discovered by self-supervised learning from 3446 whole-slide images
Authors:
- David A. MooreDMDavid A. Moore
Preprint
Abstract
1 Mesothelioma is a highly lethal and poorly biologically understood disease which presents diagnostic challenges due to its morphological complexity. This study uses self-supervised AI (Artificial Intelligence) to map the histomorphological landscape of the disease. The resulting atlas consists of recurrent patterns identified from 3446 Hematoxylin and Eosin (H&E) stained images scanned from resected tumour slides. These patterns generate highly interpretable predictions, achieving state-of-the-art performance with 0.65 concordance index (c-index) for outcomes and 88% AUC in subtyping. Their clinical relevance is endorsed by comprehensive human pathological assessment. Furthermore, we characterise the molecular underpinnings of these diverse, meaningful, predictive patterns. Our approach both improves diagnosis and deepens our understanding of mesothelioma biology, highlighting the power of this self-learning method in clinical applications and scientific discovery.Keywords
All Related Versions
- Published version: Nature Communications, 16 (1), 1.
This publication has 40 references indexed in Scilit:
- Transformer-based unsupervised contrastive learning for histopathological image classificationMedical Image Analysis, 2022
- Malignant Mesothelioma Subtyping of Tissue Images via Sampling Driven Multiple Instance Prediction2022
- Deep-learning based classification distinguishes sarcomatoid malignant mesotheliomas from benign spindle cell mesothelial proliferationsLaboratory Investigation, 2021
- Comprehensive Molecular and Pathologic Evaluation of Transitional Mesothelioma Assisted by Deep Learning Approach: A Multi-Institutional Study of the International Mesothelioma Panel from the MESOPATH Reference CenterJournal of Thoracic Oncology, 2020
- Interobserver variation in the assessment of the sarcomatoid and transitional components in biphasic mesotheliomasLaboratory Investigation, 2020
- EURACAN/IASLC Proposals for Updating the Histologic Classification of Pleural Mesothelioma: Towards a More Multidisciplinary ApproachJournal of Thoracic Oncology, 2020
- Hover-Net: Simultaneous segmentation and classification of nuclei in multi-tissue histology imagesMedical Image Analysis, 2019
- In situ growth in early lung adenocarcinoma may represent precursor growth or invasive clone outgrowth—a clinically relevant distinctionLaboratory Investigation, 2019
- Isometric Logratio Transformations for Compositional Data AnalysisMathematical Geology, 2003
- Dealing with Zeros and Missing Values in Compositional Data Sets Using Nonparametric ImputationMathematical Geology, 2003