From Linear Probing to Joint-Weighted Token Hierarchy: A Foundation Model Bridging Global and Cellular Representations in Biomarker Detection
This work addresses the need for more interpretable and robust biomarker detection in digital pathology, though it appears incremental as it builds on existing foundation models with a novel integration method.
The paper tackles the problem of AI-based biomarker detection from H&E slides by integrating cell-level morphology with global patch-level embeddings, achieving up to 8.3% higher balanced accuracy and a 1.2% average improvement over prior pathology foundation models across multiple tasks and cohorts.
AI-based biomarkers can infer molecular features directly from hematoxylin & eosin (H&E) slides, yet most pathology foundation models (PFMs) rely on global patch-level embeddings and overlook cell-level morphology. We present a PFM model, JWTH (Joint-Weighted Token Hierarchy), which integrates large-scale self-supervised pretraining with cell-centric post-tuning and attention pooling to fuse local and global tokens. Across four tasks involving four biomarkers and eight cohorts, JWTH achieves up to 8.3% higher balanced accuracy and 1.2% average improvement over prior PFMs, advancing interpretable and robust AI-based biomarker detection in digital pathology.