MOOZY: A Patient-First Foundation Model for Computational Pathology

2026-03-27

AI Summary

This work addresses the limitations of existing computational pathology foundation models, which are slide-centric, neglect intra-patient multi-slide relationships, and rely on proprietary data and costly paired reports. To overcome these issues, we propose MOOZY, the first patient-case–centric foundation model that leverages a case-level Transformer to capture dependencies among all slides from the same patient. MOOZY employs a two-stage training strategy: first, masked self-distillation pretraining on 77,134 publicly available slides, followed by multi-task alignment across 333 tasks spanning 56 public datasets. Despite having only 85.77 million parameters—approximately 1/14 the size of GigaPath—MOOZY achieves state-of-the-art or tied-best performance on eight held-out tasks and significantly outperforms TITAN and PRISM in macro-averaged metrics, demonstrating an efficient, reproducible, and publicly driven approach to cross-task transfer in computational pathology.

Abstract

Computational pathology needs whole-slide image (WSI) foundation models that transfer across diverse clinical tasks, yet current approaches remain largely slide-centric, often depend on private data and expensive paired-report supervision, and do not explicitly model relationships among multiple slides from the same patient. We present MOOZY, a patient-first pathology foundation model in which the patient case, not the individual slide, is the core unit of representation. MOOZY explicitly models dependencies across all slides from the same patient via a case transformer during pretraining, combining multi-stage open self-supervision with scaled low-cost task supervision. In Stage 1, we pretrain a vision-only slide encoder on 77,134 public slide feature grids using masked self-distillation. In Stage 2, we align these representations with clinical semantics using a case transformer and multi-task supervision over 333 tasks from 56 public datasets, including 205 classification and 128 survival tasks across four endpoints. Across eight held-out tasks with five-fold frozen-feature probe evaluation, MOOZY achieves best or tied-best performance on most metrics and improves macro averages over TITAN by +7.37%, +5.50%, and +7.83% and over PRISM by +8.83%, +10.70%, and +9.78% for weighted F1, weighted ROC-AUC, and balanced accuracy, respectively. MOOZY is also parameter efficient with 85.77M parameters, 14x smaller than GigaPath. These results demonstrate that open, reproducible patient-level pretraining yields transferable embeddings, providing a practical path toward scalable patient-first histopathology foundation models.

Problem

• computational pathology
• whole-slide image
• foundation model
• patient-level representation
• slide dependency

Innovation

• patient-first foundation model
• computational pathology
• case transformer
• multi-task self-supervision
• whole-slide image representation