MediConfusion: Can you trust your AI radiologist? Probing the reliability of multimodal medical foundation models

TL;DR

This paper introduces MediConfusion, a challenging benchmark dataset for medical multimodal models, revealing their vulnerabilities and exposing significant reliability issues in current state-of-the-art models used in healthcare.

Contribution

The paper presents MediConfusion, a novel VQA benchmark that systematically uncovers failure modes of medical MLLMs, highlighting their unreliability and guiding future improvements.

Findings

Models perform below random chance on MediConfusion

State-of-the-art models are easily confused by visually dissimilar images

Common failure patterns identified to improve model trustworthiness

Abstract

Multimodal Large Language Models (MLLMs) have tremendous potential to improve the accuracy, availability, and cost-effectiveness of healthcare by providing automated solutions or serving as aids to medical professionals. Despite promising first steps in developing medical MLLMs in the past few years, their capabilities and limitations are not well-understood. Recently, many benchmark datasets have been proposed that test the general medical knowledge of such models across a variety of medical areas. However, the systematic failure modes and vulnerabilities of such models are severely underexplored with most medical benchmarks failing to expose the shortcomings of existing models in this safety-critical domain. In this paper, we introduce MediConfusion, a challenging medical Visual Question Answering (VQA) benchmark dataset, that probes the failure modes of medical MLLMs from a vision perspective. We reveal that state-of-the-art models are easily confused by image pairs that are otherwise visually dissimilar and clearly distinct for medical experts. Strikingly, all available models (open-source or proprietary) achieve performance below random guessing on MediConfusion, raising serious concerns about the reliability of existing medical MLLMs for healthcare deployment. We also extract common patterns of model failure that may help the design of a new generation of more trustworthy and reliable MLLMs in healthcare.