Clinical reasoning in medicine is a hypothesis-driven process where physicians refine diagnoses from limited information through targeted history, physical examination, and diagnostic investigations. In contrast, current medical benchmarks for large language models (LLMs) primarily assess knowledge recall through single-turn questions, where complete clinical information is provided upfront. To address this gap, we introduce VivaBench, a multi-turn benchmark that evaluates sequential clinical reasoning in LLM agents. Our dataset comprises 1152 physiciancurated clinical vignettes structured as interactive scenarios that simulate a viva voce examination in medical training, requiring agents to actively probe for relevant findings, select appropriate investigations, and synthesize information across multiple steps to reach a diagnosis. We evaluated several state-of-the-art LLMs and found that while models demonstrate competence in diagnosing conditions within well-described clinical presentations, their performance degrades significantly when required to navigate diagnostic uncertainty. Our analysis identified several failure modes that mirror common issues in clinical practice, including: (1) fixation on initial hypotheses, (2) excessive investigation ordering, (3) premature diagnostic closure, and (4) missing critical conditions. These patterns reveal fundamental limitations in how current LLMs manage uncertainty and gather information sequentially. Through VivaBench, we provide a standardized benchmark for evaluating conversational medical AI systems for real-world clinical decision support. Beyond medical applications, we contribute to the larger corpus of research on agentic AI by demonstrating how sequential reasoning trajectories can diverge in complex decision-making environments.

Clinical diagnosis is a highly specialized discipline requiring both domain expertise and strict adherence to rigorous guidelines. While current AI-driven medical research predominantly focuses on knowledge graphs or natural text pretraining paradigms to incorporate medical knowledge, these approaches primarily rely on implicitly encoded knowledge within model parameters, neglecting task-specific knowledge required by diverse downstream tasks. To address this limitation, we propose Retrieval-Augmented Diagnosis (RAD), a novel framework that explicitly injects external knowledge into multimodal models directly on downstream tasks. Specifically, RAD operates through three key mechanisms: retrieval and refinement of disease-centered knowledge from multiple medical sources, a guidelineenhanced contrastive loss that constrains the latent distance between multi-modal features and guideline knowledge, and the dual transformer decoder that employs guidelines as queries to steer cross-modal fusion, aligning the models with clinical diagnostic workflows from guideline acquisition to feature extraction and decision-making. Moreover, recognizing the lack of quantitative evaluation of interpretability for multimodal diagnostic models, we introduce a set of criteria to assess the interpretability from both image and text perspectives. Extensive evaluations across four datasets with different anatomies demonstrate RAD's generalizability, achieving state-of-the-art performance. Furthermore, RAD enables the model to concentrate more precisely on abnormal regions and critical indicators, ensuring evidence-based, trustworthy diagnosis. Our code is available at this repository.

Section C introduces5 the three main tasks defined in our benchmark: abnormality detection, radiological image captioning,6 and clinical reasoning through differential diagnosis.

In many real-world applications, deployed models encounter inputs that differ from the data seen during training. Open-world recognition ensures that such systems remain robust as ever-emerging, previously unknown categories appear and must be addressed without retraining. Foundation and vision-language models are pretrained on large and diverse datasets with the expectation of broad generalization across domains, including medical imaging. However, benchmarking these models on test sets with only a few common outlier types silently collapses the evaluation back to a closed-set problem, masking failures on rare or truly novel conditions encountered in clinical use. We therefore present NOVA, a challenging, real-life evaluation-only benchmark of 900 brain MRI scans that span 281 rare pathologies and heterogeneous acquisition protocols. Each case includes rich clinical narratives and double-blinded expert bounding-box annotations. Together, these enable joint assessment of anomaly localisation, visual captioning, and diagnostic reasoning. Because NOVA is neverused for training, it serves as an extreme stress-test of out-of-distribution generalisation: models must bridge a distribution gap both in sample appearance and insemantic space.

Multimodal large language models (MLLMs) have begun to demonstrate robust reasoning capabilities on general tasks, yet their application in the medical domain remains in its early stages. Constructing chain-of-thought (CoT) training data is essential for bolstering the reasoning abilities of medical MLLMs.

Recent advances in computational pathology have led to the emergence of numerous foundation models. These models typically rely on general-purpose encoders with multi-instance learning for whole slide image (WSI) classification or apply multimodal approaches to generate reports directly from images. However, these models cannot emulate the diagnostic approach of pathologists, who systematically examine slides at low magnification to obtain an overview before progressively zooming in on suspicious regions to formulate comprehensive diagnoses.

While recent multimodal large language models (MLLMs) have advanced automated ECG interpretation, they still face two key limitations: (1) insufficient multimodal synergy between ECG time series and ECG images, and (2) limited explainability in linking diagnoses to granular waveform evidence. We introduce GEM, the first MLLM unifying ECG time series, 12-lead ECG images and text for grounded and clinician-aligned ECG interpretation. GEM enables feature-grounded analysis, evidence-driven reasoning, and a clinician-like diagnostic process through three core innovations: a dual-encoder framework extracting complementary time series and image features, cross-modal alignment for effective multimodal understanding, and knowledge-guided instruction data generation for generating high-granularity grounding data (ECG-Grounding) linking diagnoses to measurable parameters (e.g., QRS/PR Intervals). Additionally, we propose the Grounded ECGUnderstanding task, a clinically motivated benchmark designed to comprehensively assess the MLLM's capability in grounded ECG understanding. Experimental results on both existing and our proposed benchmarks show GEM significantly improves predictive performance (CSN 7.4%), explainability (22.7%), and grounding (25.3%), making it a promising approach for real-world clinical applications.

There is growing interest in using machine learning (ML) to support clinical diagnosis, but most approaches rely on static, fully observed datasets and fail to reflect the sequential, resource-aware reasoning clinicians use in practice. Diagnosis remains complex and error prone, especially in high-pressure or resource-limited settings, underscoring the need for frameworks that help clinicians make timely and cost-effective decisions. We propose ACTMED(Adaptive Clinical Test selection via Model-based Experimental Design), a diagnostic framework that integrates Bayesian Experimental Design (BED) with large language models (LLMs) to better emulate real-world diagnostic reasoning. At each step, ACTMED selects the test expected to yield the greatest reduction in diagnostic uncertainty for a given patient. LLMs act as flexible simulators, generating plausible patient state distributions and supporting belief updates without requiring structured, task-specific training data. Clinicians can remain in the loop; reviewing test suggestions, interpreting intermediate outputs, and applying clinical judgment throughout. We evaluate ACTMEDon real-world datasets and show it can optimize test selection to improve diagnostic accuracy, interpretability, and resource use. This represents a step toward transparent, adaptive, and clinician-aligned diagnostic systems that generalize across settings with reduced reliance on domain-specific data.

Large language models (LLMs) have achieved significant performance progress in various natural language processing applications. However, LLMs still struggle to meet the strict requirements for accuracy and reliability in the medical field and face many challenges in clinical applications. Existing clinical diagnostic evaluation benchmarks for evaluating medical agents powered by LLMs have severe limitations. Firstly, most existing medical evaluation benchmarks face the risk of data leakage or contamination.

A study of 140,000 people suggests that a broadening of the diagnostic criteria for autism and ADHD explains the sharp rise in diagnoses, but that doesn't mean too many people are being told they are autistic or have ADHD We may be beginning to understand what is behind the recent explosion in diagnoses of ADHD and autism . A study of 140,000 people in Denmark reveals that those recently diagnosed with ADHD or autism have fewer genetic variations associated with them than people diagnosed a decade earlier. This suggests that a broadening of the diagnostic criteria is behind the rise, but it doesn't support claims that ADHD and autism are being overdiagnosed. Diagnoses for autism and ADHD have risen up to tenfold around the world over the past two decades, particularly among girls and adults. Several possibilities have been put forward to explain this, including better awareness and understanding, a broadening of the diagnostic criteria, and even the commercial interests of pharmaceutical companies and private diagnostic clinics.