Mapis: A Knowledge-Graph Grounded Multi-Agent Framework for Evidence-Based PCOS Diagnosis
This paper proposes Mapis, a multi-agent framework for evidence-based diagnosis of Polycystic Ovary Syndrome (PCOS) that leverages a structured knowledge graph and collaborative workflow.
Why it matters
Mapis represents a novel, evidence-based approach to PCOS diagnosis that could improve clinical outcomes and patient care.
Key Points
- 1Mapis is designed to simulate the clinical diagnostic process for PCOS, a common but complex condition
- 2It decouples diagnostic tasks across specialized agents (gynecological endocrine, radiology, exclusion)
- 3Mapis utilizes a comprehensive PCOS knowledge graph to ensure verifiable, evidence-based decision-making
- 4Experiments show Mapis outperforms traditional ML models, single-agent systems, and previous medical multi-agent frameworks
Details
Polycystic Ovary Syndrome (PCOS) is a significant public health issue affecting 10% of reproductive-aged women. Previous AI-based PCOS detection tools have been constrained by their reliance on large-scale labeled data and lack of interpretability. The authors propose Mapis, a knowledge-grounded multi-agent framework designed specifically for guideline-based PCOS diagnosis. Mapis simulates the clinical diagnostic process, with specialized agents collaborating to verify inclusion criteria, rule out other causes, and leverage a comprehensive PCOS knowledge graph for evidence-based decision-making. Extensive experiments demonstrate that Mapis significantly outperforms competitive baselines, including traditional machine learning models, single-agent systems, and previous medical multi-agent frameworks.
No comments yet
Be the first to comment