CPathAgent: An Agent-based Foundation Model for Interpretable High-Resolution Pathology Image Analysis Mimicking Pathologists' Diagnostic Logic

2025-05-26

AI Summary

Current computational pathology models fail to emulate pathologists’ multi-scale diagnostic reasoning—specifically, the “low-magnification overview to high-magnification focus” workflow—relying instead on single-scale encoding or end-to-end report generation, thereby lacking interpretable, stepwise inference. This paper introduces the first agent-based pathological diagnostic agent that autonomously performs multi-scale zooming, visual navigation, and collaborative reasoning across patch-, region-, and whole-slide levels to mimic real-world slide review. Key innovations include: (1) an agent-driven, multi-stage training paradigm; (2) a unified three-scale modeling framework; (3) a cross-scale feature alignment mechanism; and (4) PathMMU-HR², the first expert-annotated benchmark for high-resolution regional analysis. Our method achieves state-of-the-art performance across patch-level classification, regional diagnosis, and whole-slide interpretation tasks, generating more comprehensive, traceable, and clinically credible diagnostic reports.

Technology Category

• Computer Vision: Multi-modal Vision
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Multiagent Systems: Other Foundations of Multi Agent Systems

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Search and Retrieval-Augmented AI: Agentic search
• Graph Algorithms and Modeling for the Web: Algorithms and analysis for heterogeneous, signed, attributed, multi-relational, temporal, higher-order, and annotated Web-related graphs

Abstract

Recent advances in computational pathology have led to the emergence of numerous foundation models. However, these approaches fail to replicate the diagnostic process of pathologists, as they either simply rely on general-purpose encoders with multi-instance learning for classification or directly apply multimodal models to generate reports from images. A significant limitation is their inability to emulate the diagnostic logic employed by pathologists, who systematically examine slides at low magnification for overview before progressively zooming in on suspicious regions to formulate comprehensive diagnoses. To address this gap, we introduce CPathAgent, an innovative agent-based model that mimics pathologists' reasoning processes by autonomously executing zoom-in/out and navigation operations across pathology images based on observed visual features. To achieve this, we develop a multi-stage training strategy unifying patch-level, region-level, and whole-slide capabilities within a single model, which is essential for mimicking pathologists, who require understanding and reasoning capabilities across all three scales. This approach generates substantially more detailed and interpretable diagnostic reports compared to existing methods, particularly for huge region understanding. Additionally, we construct an expert-validated PathMMU-HR$^{2}$, the first benchmark for huge region analysis, a critical intermediate scale between patches and whole slides, as diagnosticians typically examine several key regions rather than entire slides at once. Extensive experiments demonstrate that CPathAgent consistently outperforms existing approaches across three scales of benchmarks, validating the effectiveness of our agent-based diagnostic approach and highlighting a promising direction for the future development of computational pathology.

Problem

• Mimic pathologists' diagnostic logic in pathology image analysis
• Address inability to emulate systematic zoom-in/out examination
• Improve interpretable huge region understanding in computational pathology

Innovation

• Agent-based model mimicking pathologists' diagnostic logic
• Multi-stage training for patch, region, and slide analysis
• First benchmark for huge region pathology analysis