MedMMV: A Controllable Multimodal Multi-Agent Framework for Reliable and Verifiable Clinical Reasoning
This work addresses the critical need for reliable and verifiable AI systems in high-stakes clinical domains, representing a novel method for a known bottleneck rather than a foundational advancement.
The paper tackles the problem of instability and hallucination in multimodal large language models for clinical reasoning by introducing MedMMV, a controllable multimodal multi-agent framework that improves accuracy by up to 12.7% on six medical benchmarks and enhances reliability and truthfulness in clinical decision-making.
Recent progress in multimodal large language models (MLLMs) has demonstrated promising performance on medical benchmarks and in preliminary trials as clinical assistants. Yet, our pilot audit of diagnostic cases uncovers a critical failure mode: instability in early evidence interpretation precedes hallucination, creating branching reasoning trajectories that cascade into globally inconsistent conclusions. This highlights the need for clinical reasoning agents that constrain stochasticity and hallucination while producing auditable decision flows. We introduce MedMMV, a controllable multimodal multi-agent framework for reliable and verifiable clinical reasoning. MedMMV stabilizes reasoning through diversified short rollouts, grounds intermediate steps in a structured evidence graph under the supervision of a Hallucination Detector, and aggregates candidate paths with a Combined Uncertainty scorer. On six medical benchmarks, MedMMV improves accuracy by up to 12.7% and, more critically, demonstrates superior reliability. Blind physician evaluations confirm that MedMMV substantially increases reasoning truthfulness without sacrificing informational content. By controlling instability through a verifiable, multi-agent process, our framework provides a robust path toward deploying trustworthy AI systems in high-stakes domains like clinical decision support.