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.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2602.21950 [cs.CL] |
| (or arXiv:2602.21950v3 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2602.21950 arXiv-issued DOI via DataCite |
From: Xudong Liu [view email]
[v1]
Wed, 25 Feb 2026 14:33:33 UTC (5,263 KB)
[v2]
Sat, 11 Apr 2026 05:14:32 UTC (5,267 KB)
[v3]
Fri, 17 Apr 2026 00:38:13 UTC (5,437 KB)
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。