Clinically-Informed Modeling for Pediatric Brain Tumor Classification from Whole-Slide Histopathology Images

Accurate diagnosis of pediatric brain tumors remains a major challenge for deep learning due to severe data scarcity, class imbalance, and fine-grained morphological overlap across subtypes. To address this, researchers have developed an expert-guided contrastive fine-tuning framework that adapts pathology foundation models to weakly supervised classification of whole-slide images (WSI). The method integrates contrastive learning into slide-level multiple instance learning (MIL) to explicitly regularize the geometry of slide-level representations during downstream fine-tuning.

The framework introduces both a general supervised contrastive setting and an expert-guided variant that incorporates clinically informed hard negatives targeting diagnostically confusable tumor subtypes. Comprehensive experiments under realistic low-sample and class-imbalanced conditions show measurable improvements in fine-grained diagnostic distinctions. The expert-guided approach promotes more compact intra-class representations and improved inter-class separation, highlighting the value of explicitly shaping slide-level representations for robust classification in data-scarce pediatric pathology settings.