{"paper":{"title":"Clinically-Informed Modeling for Pediatric Brain Tumor Classification from Whole-Slide Histopathology Images","license":"http://creativecommons.org/licenses/by/4.0/","headline":"Expert-guided contrastive fine-tuning improves fine-grained pediatric brain tumor classification from whole-slide images in low-data settings.","cross_cats":[],"primary_cat":"cs.CV","authors_text":"Ankita Shukla, Chandra Krishnan, Hairong Wang, Jian Yu, Jinrui Fang, Joakim Nguyen, Nicholas Konz, Sanjay Krishnan, Tianlong Chen, Ying Ding","submitted_at":"2026-04-22T20:04:09Z","abstract_excerpt":"Accurate diagnosis of pediatric brain tumors, starting with histopathology, presents unique challenges for deep learning, including severe data scarcity, class imbalance, and fine-grained morphologic overlap across diagnostically distinct subtypes. While pathology foundation models have advanced patch-level representation learning, their effective adaptation to weakly supervised pediatric brain tumor classification under limited data remains underexplored. In this work, we introduce an expert-guided contrastive fine-tuning framework for pediatric brain tumor diagnosis from whole-slide images ("},"claims":{"count":4,"items":[{"kind":"strongest_claim","text":"Through comprehensive experiments on pediatric brain tumor WSI classification under realistic low-sample and class-imbalanced conditions, we demonstrate that contrastive fine-tuning yields measurable improvements in fine-grained diagnostic distinctions.","source":"verdict.strongest_claim","status":"machine_extracted","claim_id":"C1","attestation":"unclaimed"},{"kind":"weakest_assumption","text":"That the expert-identified hard negatives correctly target diagnostically confusable subtypes and that any observed gains in representation geometry and classification metrics are attributable to the contrastive regularization rather than other training choices or dataset artifacts.","source":"verdict.weakest_assumption","status":"machine_extracted","claim_id":"C2","attestation":"unclaimed"},{"kind":"one_line_summary","text":"An expert-guided contrastive fine-tuning framework improves fine-grained slide-level classification of pediatric brain tumors under low-data and class-imbalanced conditions by regularizing representations with clinically informed hard negatives.","source":"verdict.one_line_summary","status":"machine_extracted","claim_id":"C3","attestation":"unclaimed"},{"kind":"headline","text":"Expert-guided contrastive fine-tuning improves fine-grained pediatric brain tumor classification from whole-slide images in low-data settings.","source":"verdict.pith_extraction.headline","status":"machine_extracted","claim_id":"C4","attestation":"unclaimed"}],"snapshot_sha256":"e7fefd9790bcb61ca7479754a76c1871f6c141559d0bfcd487ee7ea13db8008d"},"source":{"id":"2604.21060","kind":"arxiv","version":2},"verdict":{"id":"46cdf7ee-8d7e-463d-98df-e6e63fc5d237","model_set":{"reader":"grok-4.3"},"created_at":"2026-05-09T23:49:51.709894Z","strongest_claim":"Through comprehensive experiments on pediatric brain tumor WSI classification under realistic low-sample and class-imbalanced conditions, we demonstrate that contrastive fine-tuning yields measurable improvements in fine-grained diagnostic distinctions.","one_line_summary":"An expert-guided contrastive fine-tuning framework improves fine-grained slide-level classification of pediatric brain tumors under low-data and class-imbalanced conditions by regularizing representations with clinically informed hard negatives.","pipeline_version":"pith-pipeline@v0.9.0","weakest_assumption":"That the expert-identified hard negatives correctly target diagnostically confusable subtypes and that any observed gains in representation geometry and classification metrics are attributable to the contrastive regularization rather than other training choices or dataset artifacts.","pith_extraction_headline":"Expert-guided contrastive fine-tuning improves fine-grained pediatric brain tumor classification from whole-slide images in low-data settings."},"integrity":{"clean":true,"summary":{"advisory":0,"critical":0,"by_detector":{},"informational":0},"endpoint":"/pith/2604.21060/integrity.json","findings":[],"available":true,"detectors_run":[{"name":"doi_compliance","ran_at":"2026-05-20T01:21:54.175362Z","status":"completed","version":"1.0.0","findings_count":0}],"snapshot_sha256":"e2e20e5701304fdf1808814397d4e1fa21a447251496611efcb661b2aaab0eed"},"references":{"count":0,"sample":[],"resolved_work":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57","internal_anchors":0},"formal_canon":{"evidence_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"author_claims":{"count":0,"strong_count":0,"snapshot_sha256":"258153158e38e3291e3d48162225fcdb2d5a3ed65a07baac614ab91432fd4f57"},"builder_version":"pith-number-builder-2026-05-17-v1"}