When Background Matters: Breaking Medical Vision Language Models by Transferable Attack
For clinical VLM users, this reveals a critical vulnerability in reasoning capabilities that could lead to undetected misdiagnoses in real-world settings.
MedFocusLeak achieves state-of-the-art transferable black-box attacks on medical VLMs, generating imperceptible perturbations in background regions that cause clinically plausible misdiagnoses across six imaging modalities.
Vision-Language Models (VLMs) are increasingly used in clinical diagnostics, yet their robustness to adversarial attacks remains largely unexplored, posing serious risks. Existing medical attacks focus on secondary objectives such as model stealing or adversarial fine-tuning, while transferable attacks from natural images introduce visible distortions that clinicians can easily detect. To address this, we propose MedFocusLeak, a highly transferable black-box multimodal attack that induces incorrect yet clinically plausible diagnoses while keeping perturbations imperceptible. The method injects coordinated perturbations into non-diagnostic background regions and employs an attention distraction mechanism to shift the model's focus away from pathological areas. Extensive evaluations across six medical imaging modalities show that MedFocusLeak achieves state-of-the-art performance, generating misleading yet realistic diagnostic outputs across diverse VLMs. We further introduce a unified evaluation framework with novel metrics that jointly capture attack success and image fidelity, revealing a critical weakness in the reasoning capabilities of modern clinical VLMs.