MEDSYN: Benchmarking Multi-EviDence SYNthesis in Complex Clinical Cases for Multimodal Large Language Models

TL;DR

MEDSYN is a comprehensive benchmark for evaluating multimodal large language models in complex clinical scenarios, highlighting their strengths and failure modes in synthesizing heterogeneous medical evidence.

Contribution

This paper introduces MEDSYN, a new multilingual, multimodal benchmark with complex clinical cases, and analyzes model performance and failure modes in medical diagnosis tasks.

Findings

Top models match or outperform human experts in differential diagnosis.

All models show a significant gap between differential diagnosis and final diagnosis performance.

Smaller evidence sensitivity gap correlates with higher diagnostic accuracy.

Abstract

Multimodal large language models (MLLMs) have shown great potential in medical applications, yet existing benchmarks inadequately capture real-world clinical complexity. We introduce MEDSYN, a multilingual, multimodal benchmark of highly complex clinical cases with up to 7 distinct visual clinical evidence (CE) types per case. Mirroring clinical workflow, we evaluate 18 MLLMs on differential diagnosis (DDx) generation and final diagnosis (FDx) selection. While top models often match or even outperform human experts on DDx generation, all MLLMs exhibit a much larger DDx--FDx performance gap compared to expert clinicians, indicating a failure mode in synthesis of heterogeneous CE types. Ablations attribute this failure to (i) overreliance on less discriminative textual CE ($\it{e.g.}$, medical history) and (ii) a cross-modal CE utilization gap. We introduce Evidence Sensitivity to quantify the latter and show that a smaller gap correlates with higher diagnostic accuracy. Finally, we demonstrate how it can be used to guide interventions to improve model performance. We will open-source our benchmark and code.