Medical Visual Question Answering

Also known as VQA

Overview

Medical visual question answering (MedVQA) asks models to interpret images such as X-rays, scans, pathology slides, and surgical videos in response to clinical questions, supporting tasks from finding specific abnormalities to integrating evidence across views or time. Common approaches adapt vision-language models with medical image-question datasets, while newer methods add region grounding, clinical knowledge, stepwise reasoning, or tools for measurements and image navigation. Benchmarks increasingly test spatial, temporal, multilingual, and multi-image reasoning, as well as whether answers actually depend on the visual evidence. A central challenge remains that high answer accuracy can coexist with shortcut use, hallucinations, poor calibration, or weak generalization across modalities and clinical settings.

Latest papers

183 papers8 papers in the last four weeks, down 58% on the four weeks before. 0.1% of all new papers.

Nov 19, 2026

Harnessing Medical Visual-Language Models with Agentic Evidence Management

Medical vision-language agents increasingly rely on segmentation tools, and recent work has focused on acquiring the right mask, often treating a good mask as the end point of tool use. But do these agents use the resulting evidence effectively? To answer this question, we need a benchmark that separates evidence use from segmentation quality, spans multiple medical image types, and allows the same mask to be presented in different ways. Existing medical VQA datasets do not jointly provide these properties. We therefore construct Mask-Evidence MedVQA, a 2.5K-example benchmark spanning four image types, where each question is paired with an answer-relevant ground-truth mask and multiple possible renderings. Our study reveals two limitations of direct tool use. First, evidence presentation matters. Different renderings of the same fixed mask can substantially change model behavior. Second, masks introduce an evidence trade-off. Highlighting a relevant region can expose diagnostic evidence the model previously missed, but the added visual cue can also induce incorrect associations. As a result, even a ground-truth mask can correct some answers while turning initially correct ones into incorrect ones. Together, these findings show that acquiring the right mask is not enough. An agent must also decide how to present the evidence and whether it should change the answer. We therefore introduce TUNE (Tool Use is Not Enough), an evidence-management harness that explicitly manages what evidence to acquire, how to present it, and whether to accept the resulting answer update. Across three public medical VQA benchmarks, evidence management improves the fine-tuned Qwen3.5-9B backbone's mean accuracy by 4.0 percentage points. Full TUNE improves over direct tool use by 9.1 points on average across seven benchmarks and achieves competitive results against existing medical VLMs and agents.
Nov 17, 2026

OmniTongue: A Comprehensive MLLM Benchmark for Automated Tongue Diagnosis in Traditional Chinese Medicine

Multimodal Large Language Models (MLLMs) have excelled in Western medical imaging, yet their potential in Traditional Chinese Medicine (TCM) tongue diagnosis remains systematically unevaluated. Tongue diagnosis requires sophisticated global and local perception, serving as a rigorous touchstone for assessing the visual understanding depth of MLLMs. However, existing datasets are limited by single dimensions, coarse granularity, and ambiguous labels, hindering the evaluation of authentic clinical capabilities. To bridge this gap, we introduce OmniTongue, the first comprehensive, fine-grained Visual Question Answering benchmark for TCM tongue diagnosis. This benchmark integrates 1,351 expert-validated images and 21,616 instruction pairs. Its core design features 16 clinical dimensions for multidimensional evaluation, region-specific annotations for pathological localization, and 66 fine-grained labels for precise diagnostic representation. Extensive evaluations of existing MLLMs reveal significant performance gaps, demonstrating that medical-specific models lack domain advantages, while universally poor Macro-F1 scores expose a reliance on statistical shortcuts rather than genuine understanding of tongue pathology. These findings not only highlight the current limitations of MLLMs in interpreting regional medical features but also underscore OmniTongue's significance for developing robust automated tongue diagnosis models. The code and dataset will be made publicly available in a future release.
Nov 16, 2026

MedJev: A System One Vision-Language Decision Model for Fast and Effective Medical Image Understanding

Medical image understanding with vision-language models (VLMs) typically relies on autoregressively generating reasoning and answer tokens. These unnecessary tokens make the inference cost scale with the response length and force the answer to be parsed from free-form text, incurring avoidable latency and decision errors. To this end, we introduce MedJev, a system one vision-language decision model that maps medical images, questions, and candidate answers directly to decisions in a single multimodal forward pass, without any autoregressive generation process. For fast inference, MedJev removes the upper language layers and the generative output pathway, while keeping the full vision pathway and all visual tokens. On this shortened backbone, a decision head scores each candidate by pretrained query--key matching and refines the scores with a lightweight multilayer perceptron (MLP) residual, coupling the pretrained matching structure with task-specific nonlinear discrimination. Built on Qwen3.5-4B, MedJev-4B improves both accuracy and speed over the original model on each of four medical image classification and visual question answering benchmarks, with an absolute accuracy gain of 7.93% and a \textbf{6.04×\times} end-to-end speedup. Further comparison with 20 cutting-edge models shows that MedJev-4B achieves state-of-the-art performance and inference speed among 4B--9B general and medical VLMs, indicating that direct multimodal decision making is an effective route to fast and effective medical image understanding.
Nov 13, 2026

QCAP-VQA: Question-Conditioned Capped Prior-Mass Supervision for Medical Visual Question Answering

Medical visual question answering models can answer correctly without attending to the anatomy that the question names. Automatic region priors allow spatial supervision without expert masks, but they are noisy, and it is unclear how strongly they should shape attention and how their effect should be verified. We present QCAP-VQA, which supervises a subset of cross-attention heads with a question-conditioned prior built from ranked segmentation proposals. Its capped prior-mass objective penalizes insufficient attention inside the prior only below a threshold. We evaluate attention against SLAKE structure boxes that play no role in training, separating attention on the named anatomy from how attention shifts between questions about the same image. QCAP-VQA matches the accuracy of a baseline without spatial supervision (84.26% vs. 84.23%), and every evaluated prior-supervised model places more attention on the annotated structures. Over a 64-fold range of regularization strength, the capped objective keeps prior mass between 0.67 and 0.77, whereas a pooled region-mass objective drives it to 0.97 and loses accuracy. The cost lies in question conditioning: as the constraint widens or strengthens, attention shifts agree less with the expert-defined contrast. Across 13 configurations, prior-referenced and expert-referenced evaluation rank the models in largely opposite order (Spearman correlation of -0.87), and the capped models it favors track the prior's own errors. Because clinical use of medical visual question answering requires answers grounded in the queried anatomy, supervision from automatic priors should be judged against independent anatomical references.
Nov 7, 2026

EndoSCOP-V: A Multi-Turn Video Understanding Evaluation Framework for Multimodal Models in Endoscopy Reporting and Clinical Reasoning

Endoscopic diagnosis unfolds over native video, where a single positive finding during traversal of the gastrointestinal tract sets off a multi-minute chain of reasoning over location, morphology, severity and intervention. Existing endoscopy multimodal benchmarks test only static images or single-disease videos, which miss the temporal and clinical reasoning that real-world reporting demands. We introduce EndoSCOP-V, a video-grounded multi-turn evaluation framework for structured endoscopy reporting and clinical reasoning. It contains 1,307 real-world endoscopy and colonoscopy videos (63.2 h; 97% 1080p) averaging about 174 s. Its 20,274 question-answer turns are instantiated deterministically from clinician-filled structured endoscopy-reporting EHRs, so no per-case annotation is synthetic or LLM-rephrased. They come from 1,033 clinician-reviewed templates, about 18 per disease, that span a clinically aligned 9-task x 30-sub-task reporting taxonomy over 54 World Endoscopy Organization (WEO) MST 3.0 diseases. 22% of procedures carry co-existing pathologies, up to six per case. A Markovian context-pruned oracle-forcing protocol evaluates the dialogue turn by turn, injecting the gold answer before each subsequent turn so that per-turn reasoning is isolated from cascading dialogue errors. A panel of 15 board-certified endoscopists sets a 78% strict-accuracy baseline on 1,500 questions. Across 14 open-weight LMMs, the best 27B-class model reaches 59.8% strict accuracy, falling to 53.4% once format-trivial disease-rejection items are excluded. On a between-subjects spot-check, the best of 3 frontier proprietary models trails the clinician panel by 23 pp on strict NDR accuracy. Sub-task analysis reveals consistent gaps in quantitative measurement, multi-variable grading and intervention reasoning on long-horizon video. Data: Hugging Face (EndoSCOP-V). Code: GitHub (abhi25t/EndoSCOP-V).
Nov 1, 2026

SurgQ-VLM: Question-Conditioned Dual-Stream Compression and Answer-Type-Aware Reinforcement for Surgical Video Question Answering

Reasoning about surgical video -- which phase is underway, which instruments act, and when events occur -- underpins operative documentation, workflow analysis, and skill assessment. Yet surgical visual question answering is still largely built on single frames and closed labels that cannot express a duration, an ordering, or a count, and the video models that lift the frame limit typically compress every clip the same way regardless of the question, discarding the temporal evidence the answer needs. Worse, the answers are structurally heterogeneous, spanning five incompatible structures -- binary, classification, multi-label, counting, and sentence-temporal -- that no single training signal scores uniformly well. We propose SurgQ-VLM, a two-stage model for surgical video question answering. Stage 1 trains a question-conditioned dual-stream (slow/fast) compressor for a LoRA-adapted Qwen2.5-VL-7B: cross-attention conditions both streams on the question, and each clip's visual tokens shrink 49x from the raw ViT patches (11x vs. post-merge backbone tokens). Stage 2 post-trains with GRPO under an answer-type-aware reward that scores each question on the metric matching its answer structure (micro-F1 for label sets, normalised error for counts, a content-timing blend for temporal windows). On a sealed 21,840-question test split over Cholec80 and CholecT50, SurgQ-VLM reaches 0.779 keyword accuracy and a 10x BLEU-4 gain over zero-shot; GRPO adds 2.5 pp K-Acc (2.0 pp attributable to the answer-type-aware reward), concentrated in the structurally hard types, with answer format unchanged (ROUGE-L). Code is available at https://github.com/DhruvMakwana/SurgQ-VLM.
Oct 31, 2026

Density-Aware Calibration for Resource-Efficient Medical Vision-Language Models

Medical vision-language models (VLMs) show promise for clinical image understanding, but their computational and memory requirements hinder deployment on resource-constrained clinical hardware. Post-training quantization (PTQ) offers a practical compression strategy without retraining, yet its effectiveness depends strongly on the calibration data. We propose CoreCalib, a training-free and quantizer-agnostic framework for constructing compact calibration sets that preserve the target medical VQA distribution. CoreCalib stratifies data by clinical question class, identifies representative visual modes using frozen DINOv2 features with UMAP and HDBSCAN, and expands selected images into class-balanced question-answer pairs. On Kvasir-VQA-x1, we evaluate CoreCalib with MASQuant, GPTQ, and QIG using a fixed 128-pair budget. CoreCalib achieves the best METEOR with all three quantizers and the best ROUGE-L with MASQuant and GPTQ. With W4A8 MASQuant, it improves METEOR from 0.5672 with COCO calibration to 0.6234, recovering approximately 41% of the degradation relative to BF16. These results support CoreCalib as a data-centric component for resource-efficient medical VLM deployment. Project page: https://magnusdtd.github.io/CoreCalib/ .
Oct 25, 2026

GRASP: Group-Adaptive Hybrid RL-SFT Post-Training for Medical VQA without Reasoning Annotations

Adapting general-purpose vision-language models (VLMs) to medical visual question answering (VQA) is valuable for helping clinicians read images and make diagnoses, but annotated data are scarce and provide only short answers, without reasoning annotations. Supervised fine-tuning (SFT) fits these labels but overfits to them and amplifies their biases, eroding out-of-domain (OOD) generalization. Reinforcement learning (RL) better preserves generalization but never uses the gold answer as a training target: GRPO-style algorithms learn from the contrast between correct and incorrect answers within a group of sampled responses, so all-correct or all-wrong groups give no correctness signal. Existing SFT-RL hybrids either apply SFT and RL in separate stages or switch between them per question, or need reasoning traces from experts or stronger models, which medical VQA data lack. We propose GRASP, which combines RL and SFT in every update using only short answers: RL lets the model explore its own reasoning, SFT teaches the gold answer after that reasoning to supply the direct supervision RL lacks, and their balance adapts to how well the model answers each question. Tuning only about 1% of the parameters with LoRA, GRASP outperforms all compared training objectives both in domain and on OOD open-ended questions across Qwen3-VL-8B, InternVL3-8B and Gemma 3-4B. On an OOD benchmark spanning eight imaging modalities, GRASP-trained InternVL3-8B and Qwen3-VL-8B beat all 10 compared general and medical VLMs in open-ended accuracy, including models trained on up to 2,600 times more data or with over 250 times more trainable parameters. We release MedTerse, a dataset of 12.9K short-answer questions, and our code at https://github.com/sea-comet/GRASP-MedVQA.
Oct 25, 2026

Conviction Without Evidence: Verdict Stability Is Independent of Visual Grounding in Medical Vision--Language Models

Vision--language models applied to medical question answering can reach correct diagnoses without using the diagnostic image, inferring instead from the clinical vignette. We ask whether this is visible in how firmly a model holds its answer: does a model defend a diagnosis derived from a blank placeholder as tenaciously as one derived from a real medical image? Crossing an image-substitution protocol with two levels of adversarial pressure on 60 Italian State Examination items, we find no evidence that ungrounded verdicts are more fragile. In a paired item-level analysis, fabricated consensus (L4) overturns 94.4% of verdicts with the real image and 98.1% without (Δ=+0.037Δ= +0.037; one-sided 95% upper bound +0.105+0.105); contentless doubt (L1) overturns 62.5% and 51.8% (Δ=−0.107Δ= -0.107, 95% CI [−0.209,−0.001][-0.209, -0.001]), the point estimate running opposite to the direction a grounding detector requires. The two levels nonetheless produce opposite dynamics: under L4 the model flips once, always to the exact option argued for (110/110 first-turn flips), and then freezes, averaging 1.9 verdict changes over ten turns; under L1 it never settles, averaging 7.4 changes. Instability under contentless doubt tracks intrinsic item ambiguity rather than evidence availability, and so does not serve as a grounding signal. Flips are almost entirely corrupting (99/100 across both levels). The model is not unaware of the missing evidence---it flags the blank image at baseline on 35 of 60 items and abstains on 4--6---yet those abstentions do not survive contact with pressure. None of the epistemic signals a user could plausibly consult distinguishes a grounded diagnosis from a confabulated one.
Oct 25, 2026

MIS-VQA: An Annotation-Driven Surgical Vision-Language Dataset for Complex Surgical Understanding

Modern surgical AI increasingly requires multimodal understanding rather than isolated perception, yet existing datasets remain largely limited to single-task supervision such as recognition or templated question answering. Current medical VQA resources often rely on synthetic templates, lack natural descriptive supervision, or provide sparse annotations that hinder reasoning about spatial relations and procedural context. We introduce MIS-VQA, a surgical vision-language dataset generated through an annotation-driven pipeline that converts heterogeneous surgical annotations into structured textual scene representations. Using EndoVis18 and CholecT50, we transform phase, instrument, interaction, triplet, and segmentation labels into unified contexts that guide large language models to generate clinically coherent question-answer pairs. MIS-VQA supports four response types-concise and detailed descriptions, perception-level grounding, and multi-step reasoning-enabling hierarchical evaluation from atomic visual grounding to compositional reasoning. To address annotation sparsity, we incorporate confidence-filtered segmentation and cross-annotation consistency checks, enabling spatially grounded supervision at scale. MIS-VQA establishes a unified benchmark for surgical vision-language understanding with spatial supervision and multi-tier reasoning evaluation. To demonstrate its utility, we introduce LLaVA-SurgXpert, a family of surgical vision-language models with visual encoders to analyze how visual representations influence multimodal reasoning. Experiments show consistent improvements over existing surgical VQA baselines while revealing capability gaps between perceptual grounding and higher-level reasoning hidden in prior benchmarks.
Oct 24, 2026

Learning When to Collaborate: Selective Multi-Agent Medical Reasoning via Uncertainty-Aware Routing

Multi-agent systems have emerged as a promising paradigm for multimodal medical reasoning, but existing methods rely on two assumptions that hinder practical deployment: (i) always-active dense collaboration among multiple agents, and (ii) dependence on large cloud-based models that raise privacy and efficiency concerns. In this paper, we challenge both assumptions and propose SMART-Med, a fully local multi-agent method that treats collaboration as a learnable decision rather than a default behavior. Our key insight is that not every medical query requires costly multi-agent deliberation: many can be reliably solved by a single well-trained agent, while only uncertain cases benefit from expert collaboration. Based on this insight, SMART-Med introduces an uncertainty-driven routing mechanism that adaptively decides when to answer directly and when to recruit a small subset of expert agents, effectively converting dense multi-agent reasoning into selective, query-adaptive coordination. To make this routing reliable, we curate a new multimodal medical dataset MedVR-3K with high-quality reasoning traces and a difficulty hierarchy, and then fine-tune local agents using GRPO with two complementary rewards: a certainty-aware accuracy reward and a vision-language reward model for correctness and reasoning quality respectively. Experiments on four medical visual question answering benchmarks show that SMART-Med achieves competitive performance compared to existing MAS approaches while substantially reducing inference cost (reducing inference time by more than 3.8x and token usage by 72%). Moreover, after fine-tuning on MedVR-3K, SMART-Med yields an average improvement of 6.7% across Slake, PathVQA, VQA-RAD, and PMC-VQA.
Oct 16, 2026

Enhancing RAG-based Medical VQA with Gated Multimodal Representations

Medical Vision-Language Models (Med-LVLMs) have shown strong performance in Medical Visual Question Answering (MVQA), yet they may still produce factually inconsistent responses. Retrieval-Augmented Generation (RAG) can mitigate this issue by incorporating external clinical evidence; however, retrieval is often guided by a single modality, potentially overlooking the joint semantics of medical images and clinical questions. We introduce a multimodal RAG framework that learns a joint query representation through a Gated Multimodal Unit, enabling evidence retrieval to leverage information from both modalities while keeping the underlying Med-LVLM frozen. We evaluate the proposed approach on benchmark datasets. Results show that the multimodal query representation consistently outperforms a reference Med-LVLM and remains competitive with state-of-the-art methods. Our findings demonstrate the potential of multimodal query representations for more effective evidence retrieval in Medical VQA.
Oct 16, 2026

When Evidence Vanishes: Distilling Visual-Evidence Dependency in Medical Vision-Language Models

Compact medical vision-language models (VLMs) can answer clinical questions correctly without reliably using the image, relying instead on question wording or learned priors. Standard knowledge distillation copies a teacher's answer preference on the original image, but it leaves unchecked whether that preference is conditional on the visual evidence required by the question. We study this missing target as medical visual evidence dependency: support for the correct answer should decrease when the matched evidence is absent. We introduce Evidence-Conditioned Dependency Distillation (ECDD), which pairs each training image with a control matched to the evidence level and distills the teacher's drop in answer margin across the pair. A lower bound objective transfers positive dependency by penalizing only shortfalls, while original-image supervision preserves task accuracy. The same formulation applies when the evidence is a queried clinical region, a surgical instrument, or whole-image visual content. Across three medical VQA settings, ECDD keeps original-image accuracy competitive with distillation and alignment baselines while making answer support more sensitive to evidence removal. Under the DME headline objective, for example, the positive-case margin response to target rather than random masking rises from 0.786 with KD to 2.023 with ECDD, while original-image accuracy differs by 0.11 points. These results suggest that medical VLM distillation should transfer both the teacher's answer preference and its dependence on visual evidence.
Oct 15, 2026

MedSTP: Preserving Subspace and Transition Evidence for Medical VQA Token Pruning

Medical vision-language models (VLMs) for visual question answering (VQA) often pass long visual-token sequences to the language model, making visual-token pruning attractive but risky for medical images, where answers may depend on localized boundary and texture evidence. We study training free pre-decoding pruning and treat pruning as a subspace stable evidence preservation problem. The retained tokens should remain close to the dominant visual evidence subspace induced by the full token set. To this end, we propose Subspace-Transition Token Pruning (\method{}), which estimates each token's spectral support from principal component space and constructs a compact representation subspace. To better reflect medical image structure, \method{} further uses a bounded token-grid transition prior to promote boundary and texture sensitive tokens. We evaluate \method{} on the closed-ended SLAKE and OmniMedVQA benchmarks and on RadImageNet-VQA, which provides both closed-ended and open-ended evaluations, using Lingshu-7B and Qwen3-VL-8B backbones. \method{} delivers consistently strong performance on both the closed-ended and open-ended RadImageNet-VQA evaluations. These results show that \method{} is an effective training free pruning framework for medical domain VQA.
Oct 13, 2026

PlanPath: Exploration-Guided Orchestration for Patient-Level Pathology Reasoning

A patient's pathology case spans tens of slides from distinct specimens and anatomical sites, each carrying non-redundant, site-specific evidence. Forming a patient-level conclusion therefore requires locating the slides relevant to a given question and integrating complementary findings across them. However, existing pathology agents analyze one slide at a time and describe morphology rather than measure it, leaving evidence acquisition across a patient's specimens and patient-level quantification underexplored. To address this gap, we propose PlanPath, an agentic framework for patient-level pathology visual question answering (VQA). PlanPath turns the question into an exploration plan that specifies the evidence required, routes to the slides likely to carry it, and drives coarse-to-fine acquisition from tissue architecture down to individual cells. Tissue- and cell-level tools supply the counts, proportions, and extents, which PlanPath aggregates across slides into a traceable answer. To enable evaluation, we build two patient-level VQA datasets for renal and gastric cancer, averaging 15 WSIs per patient, with pathologist-verified question-answer pairs derived from diagnostic reports. Experimental results show that PlanPath outperforms the strongest baseline by 9.4% and 12.1% in overall accuracy on the two datasets and further improves existing slide-level agents when plugged into our framework. The dataset and code are available at: https://github.com/JSGe-AI/PlanPath_1.
Oct 13, 2026

Seeing Through the Forgery Is Not Enough: Regressive Sycophancy under Fabricated Evidence in Medical Vision-Language Models

A medical vision-language model may analyze a given CT scan and correctly detect both kidneys in view. Continuing this conversation, if a user provides a prescription with an opposite direction, medical VLMs often sycophantically choose to realign their diagnosis. In this work, we study regressive sycophancy under fabricated evidence, where a model abandons a correct answer when pushback comes with a fake document. We test four open VLMs on 400 VQA-RAD yes/no questions in 19,200 multi-turn conversations, with and without a fabricated prescription. Without a disclosure that documents are synthetic, the models abandon 80.1% of 6,488 correct answers. Notably, under one script each, the same prescription raises the flip rate of Qwen2.5-VL-7B from 45.9% to 89.6% and lowers that of Qwen3-VL-8B from 77.7% to 50.9%. A model rejected the prescription while still preserving the correct answer in 915 of 3,232 conversations with the prescription (28.3%), and 40% of these flipped in later turns. System-prompt sentences that disclose the documents as synthetic lower one model's flip rate from 77.8% to 47.3%. Thus, we conclude that awareness about a forgery does not prevent a model from adopting its answer.
Oct 12, 2026

PathPilot: An Adaptive Question-Conditioned Multi-Magnification Navigation Agent for Whole-Slide Image Analysis

The diagnostic capability of WSI agents relies on effective slide navigation to identify diagnostically relevant regions and efficiently integrate evidence across the slide. While in clinical practice, pathologists dynamically examine slides across multiple scales, integrating global and cellular evidence in a sequential manner, existing AI navigation approaches mostly operate at a single fixed magnification or rely on predefined magnification traversal. This mismatch prevents existing methods from modeling cross-magnification interactions and adaptive magnification selection inherent to real diagnostic workflows. We propose PathPilot, a clinically consistent multi-magnification WSI navigation agent that explicitly models cross-magnification localization and evidence-adaptive exploration. Specifically, we introduce a CrossMagLoc that learns magnification-specific diagnostic localization from pathologist viewing behavior and integrates information from adjacent magnifications to identify the diagnostically most relevant region within the current ROI, thereby providing a diagnostically informed starting point for subsequent question-conditioned exploration. We further introduce an EviRegionSearch that starts question-conditioned exploration from this region, progressively accumulating question-relevant evidence and adaptively determining the next navigation action according to the remaining evidence requirement. Their iterative interaction enables PathPilot to sequentially explore across ROIs and magnifications until sufficient evidence is acquired. Extensive QA evaluations across five cohorts demonstrate consistent superiority over competing SOTA methods on inter/external cohorts, respectively, validating question-conditioned evidence acquisition across regions and magnifications that mirrors the sequential, knowledge-driven navigation of pathologists.
Oct 7, 2026

HeiCo-FOCUS: A Clinically Grounded Dataset for Long-Context Video Understanding

Recent advances in Vision-Language Models (VLMs) have led to rapid progress in video understanding across a wide range of benchmark tasks. However, existing evaluations largely focus on short-term reasoning, failing to assess a critical capability: maintaining cumulative temporal consistency over extended time horizons. To close this evaluation gap, we introduce HeiCo-FOCUS, a clinically grounded dataset for evaluating long-context video understanding through the task of Foreign Object Contextual Understanding in Surgery. Built on a dataset of Heidelberg Colorectal surgeries, this task requires models to continuously track multiple objects as they are inserted, manipulated, occluded, and removed over procedures lasting up to hours. HeiCo-FOCUS comprises 30,000 visual question answering (VQA) pairs covering five core capabilities: object recognition, temporal grounding, aggregation, event and procedural understanding, and complex reasoning. The dataset was constructed through a rigorous multi-stage annotation pipeline involving large-scale crowd annotation and 39 surgical domain experts to ensure high quality and clinical relevance. To systematically probe model behavior, we introduce a multi-track evaluation framework that progressively increases temporal and contextual demands from single frames to full procedures. Experiments with ten frontier VLMs show that HeiCo-FOCUS tasks are far from solved: only around half of the models clearly outperform a text-only baseline. Across the video tracks, models perform best on event and procedural understanding (mean Accuracy: 56.5% across all models), while temporal grounding remains particularly challenging for all evaluated models (mean Accuracy: 19.7%). We therefore expect HeiCo-FOCUS to serve as a catalyst for the development of models capable of reliable, temporally consistent reasoning over hours-long videos.
Oct 5, 2026

A doctrine-grounded visual question answering dataset for Tactical Combat Casualty Care

Tactical Combat Casualty Care (TC3) requires responders to connect visual observations of injuries and interventions with established clinical guidance. Developing vision-language models to support this process requires supervision that links visible evidence to traceable doctrine. We present TC3-VQA, a dataset constructed from public instructional and field TC3 videos and authoritative TC3 documents. It contains 581 items spanning 11 concepts, with 1,860 questions covering intervention recognition, doctrine, clinical reasoning, procedural guidance, and refusal when visual information is insufficient. Doctrine-based answers preserve verbatim source passages and character offsets. Construction combines visual annotation, passage retrieval, entailment checks, and verification across model families. Equipment boxes, anatomical labels, temporal segments, and source metadata accompany the question-answer pairs. Automated audits and ratings by two physicians and two medical students characterize annotation quality, with human ratings available for 88 retained items. The dataset provides a resource for adapting vision-language models to TC3, studying the connection between visual evidence and clinical knowledge, and evaluating recognition, doctrine recall, and abstention.
Oct 5, 2026

MedPrune: Topology-Efficient Multimodal Multi-Agent Communication Evolution for Medical VQA Tasks

While medical multimodal large language models (Med-MLLMs) advance medical visual question answering (VQA), existing clinical workflow-inspired multi-agent frameworks suffer from interaction patterns and excessive computational overhead caused by redundant communication topologies. In this paper, we propose MedPrune, an efficient medical multimodal multi-agent collaboration framework that dynamically prunes both nodes and edges from the communication topology to enhance reasoning ability and token efficiency. Specifically, we first formulate the diagnostic process as a heterogeneous communication graph, where nodes represent specialist agents from various departments and edges capture intra- and inter-departmental interactions. Building on this graph, we introduce two sparsification mechanisms to enable adaptive collaborative evolution: (1) Heterogeneous Node Sparsification, which eliminates task-irrelevant specialist agents irrelevant to the current multimodal question via reinforcement learning-driven topological optimization, and (2) Heterogeneous Edge Sparsification, which selectively retains only the most diagnostically salient intra- and inter-departmental connections by jointly optimizing task performance and topological complexity. Extensive medical VQA experiments under full-set and few-shot training settings prove MedPrune surpasses multi-agent baselines and boosts token efficiency with strong adversarial robustness.
Oct 5, 2026

BrainTRACE: Tracing Longitudinal, Multimodal, and Volumetric Evidence in Brain MRI Clinical Reasoning

Brain MRI interpretation is a longitudinal clinical reasoning problem: radiologists compare serial studies, integrate information across MRI sequences, localize findings within volumetric anatomy, and translate this evidence into report-grounded assessments. Existing medical VQA and 3D imaging benchmarks capture important parts of this workflow, but often evaluate brain MRI through isolated images, static volumes, or ungrounded report-style answers, thereby obscuring failures in the evidence chain that support clinical validity. We introduce BrainTRACE, a report-grounded benchmark for evaluating whether vision-language models can trace the evidence structure required for longitudinal brain MRI interpretation. BrainTRACE contains 7,273 scored VQA instances derived from 1,778 longitudinal patients, 7,299 MRI studies, and approximately 29k co-registered 3D MRI sequence volumes. The benchmark is organized by five levels of clinical reasoning, from acquisition recognition to case-level synthesis, and by evidence demands covering longitudinal comparison, report-grounded references, multi-sequence integration, and volumetric spatial evidence. BrainTRACE supports rendered inputs compatible with standard VLM interfaces, a 3D-evidence condition, and a decomposed case-reasoning track that audits six steps in a longitudinal evidence chain. Evaluation of 20 VLM configurations shows that current systems can identify isolated visual cues but rarely compose them into grounded longitudinal interpretations. We release the benchmark specification, evaluation lists, scoring implementation, scoring rubrics, and audit-record format to support reproducible progress in brain MRI VLM evaluation.
Oct 3, 2026

Localization Lens for Improving Medical Vision-Language Models

Medical Vision-Language Models (Med-VLMs) have demonstrated strong capabilities in clinical tasks. However, they often struggle to understand anatomical structures and spatial positioning, which are crucial for medical reasoning. To address this, we propose a localization-aware enhancement to the Med-VLM pipeline, introducing improvements at three levels: data,architecture, and alignment. First, we introduce localization lens, a set of expert-validated representations that provide richer anatomical and positional context. However, as these representations increase input complexity, we integrate pixel shuffle within the model architecture to filter and refine representations, enhancing spatial information processing while preserving anatomical continuity. Lastly, to effectively align the localization lens representations with textual features, we incorporate decoupled contrastive loss (DCL) alongside the standard loss function. This ensures better feature discrimination and robustness, particularly in data limited medical settings. Through extensive evaluations on medical visual question answering (Med-VQA) datasets, we show that our methodology improves localization-driven performance across different Med-VLM architectures. Our analysis of localization-based questions further reveals that improvements in anatomy and spatial reasoning directly enhance the overall accuracy of Med-VQA upto 6.2%. The proposed approach is model-agnostic and can be seamlessly integrated into existing Med-VLM pipelines. The dataset, code, and trained models will be made publicly available at https://github.com/CVLABLUMS/localizationlens.
Oct 1, 2026

CineMR: Tool-Integrated Vision-Language Reasoning for Quantitative Cardiac MRI Assessment

Cardiovascular magnetic resonance (CMR), including cine imaging, is a reference standard for the noninvasive assessment of cardiac morphology and ventricular function. Cine CMR interpretation integrates qualitative visual assessment with quantitative measurements of ventricular volumes, ejection fraction, myocardial mass, wall thickness, and regional wall motion. Current medical vision-language models (VLMs) cannot reliably derive quantitative measurements from multidimensional cine images without analysis tools. We present CineMR, a tool-augmented VLM that invokes cardiac image-analysis tools and integrates their outputs into interleaved reasoning for quantitative CMR assessment. We also construct a multi-cohort visual question answering benchmark covering quantitative metric extraction, multiclass diagnosis, and differential diagnosis, together with tools for segmentation, phase selection, volumetry, morphometry, and regional wall motion analysis. CineMR is trained with supervised fine-tuning (SFT) on tool-interaction traces followed by Group Relative Policy Optimization (GRPO) with conditional tool-use rewards. On the multi-cohort cine CMR benchmark, CineMR achieves 35.9% pass@1 and 58.9% pass@4, compared with 1.5% pass@1 for the Qwen3-VL-8B backbone and 0.0% and 7.0% pass@1 for LLaVA-Med v1.5 and MedGemma-4B, respectively. Correct tool invocation reaches 99.8% after GRPO, up from 78.9% after SFT. Live tool outputs improve ventricular measurement accuracy by 20.4--23.7% over direct model predictions, and removing all tools reduces pass@1 from 35.9% to 27.9%. These results highlight the importance of reliable tool use for quantitative cine CMR reasoning and support CineMR as a promising approach for assistive cardiac image assessment. Code, benchmark resources, and model weights are available at https://github.com/AI-MIND-Lab/CineMR.
Sep 30, 2026

Learning Where to Look: Anatomical Grounding and Guided Attention for Cardiac MRI Vision-Language Models

Cardiac magnetic resonance imaging (CMR) enables assessment of cardiac anatomy, ventricular function, and myocardial tissue characteristics. Clinicians interpret these images by identifying cardiac structures and focusing on the regions relevant to each clinical question, motivating anatomically guided vision-language models (VLMs). Yet CMR-specific supervision for anatomical localisation and clinical question answering remains limited. To address this gap, we investigate fine-grained CMR visual question answering through anatomical grounding and guided attention. We construct 128,915 anatomical-grounding and 42,799 clinical QA pairs across short-axis cine, late gadolinium enhancement, and long-axis cine. These datasets support anatomical recognition, localisation, and clinical assessment without requiring paired reports for individual training images. To help the model learn where to look, we introduce Cardiac Anatomy-Routed Attention (CARA), which selects predicted anatomical priors according to the question and guides decoder attention with learned task-specific strengths. Combining anatomical grounding pretraining with CARA yields our model, CARA-VL. Experiments demonstrate CARA-VL's strengths in clinical assessment and regional localisation across CMR imaging settings, with promising generalization to an external clinical cohort. Together, our data and method provide a practical framework for studying and advancing cardiac visual understanding in VLMs. We will release the QA data derived from public datasets upon publication.
Sep 30, 2026

CRAFT: Causal Responsibility and Failure Tracing in Medical Vision Language Models

As vision language models are increasingly deployed in clinical diagnosis, understanding how they internally resolve competing visual and textual signals becomes a safety imperative. Existing mechanistic analyses remain confined to unimodal text and offer no explanation for why a single misleading sentence can override a correct image based diagnosis, or why a model commits to a confident answer despite insufficient visual evidence. We find that these two safety risks, arbitration failure where textual context overrides visual grounding and brake failure where the model commits without adequate evidence, are mediated by spatially disjoint attention head populations: arbitration heads form a mid-to-deep wideband reflecting cross-layer evidence competition, while brake heads concentrate in a narrow middle-to-late layer band that regulates evidence sufficiency and abstention behavior. To ground these observations in causal circuitry, we introduce CRAFT, which localizes each failure mode to a minimal causal head set via dual criteria and verifies necessity and sufficiency through temporal probes and Tuned Lens trajectory analysis. Excising arbitration heads sharply reduces conflict following with negligible degradation on clean inputs, while excising brake heads restores appropriate abstention under degraded visual evidence. The two interventions target spatially disjoint head sets and produce distinct corrective effects, underscoring the mechanistic separability of the failure modes. Experiments across multiple medical VQA benchmarks and VLM architectures validate both the localization and interventions, demonstrating that the identified heads causally drive each failure mode and that targeted modulation generalises without retraining. The code is available at https://github.com/zhcz328/CRAFT.
Sep 28, 2026

See, Measure, and Reason: Learning Visually Grounded Reasoning in Pathology

Pathological assessment relies on recognizing fine-grained visual details in histological images. Vision-language models (VLMs) increasingly support pathology interpretation, yet their ability to perceive these details remains inadequate. This weakness leads to inaccurate cellular observations that can persist even when final answers are correct. In this paper, we propose ASPECT to improve visually grounded reasoning through explicit supervision of cellular appearance and abundance. ASPECT trains intermediate visual tokens through pathology feature reconstruction, cell feature alignment, and count supervision. Three-stage supervised fine-tuning teaches the model to perceive, generate visual tokens, and reason, followed by reinforcement learning that rewards answer correctness and consistency with reported measurements. We also introduce PathoVernier, a benchmark of 759 expert-reviewed questions from five pathology datasets covering four cellular composition tasks. It evaluates both final answers and intermediate measurements to expose errors hidden by answer accuracy. On PathoVernier, ASPECT achieves relative accuracy gains of approximately 19.2% over the strongest baseline, Gemini-3.1-Pro, and 99.3% over its Qwen3-VL-8B backbone, while reducing RAWR, which measures counting errors within correct responses, by 28.1% and 42.7%, respectively. ASPECT also improves over its backbone on three external pathology benchmarks covering classification and question answering beyond cellular composition tasks.
Sep 26, 2026

EyeVQA: Benchmarking Ophthalmic Vision-Language Models from Recognition to Spatial Grounding

Vision-language models (VLMs) have shown increasing potential for medical image understanding, yet their capabilities in ophthalmic imaging remain insufficiently characterized. Existing ophthalmic datasets are typically designed for individual diseases or specialized tasks, making it difficult to systematically evaluate whether VLMs can move beyond disease recognition toward comparative reasoning and fine-grained spatial grounding. We introduce EyeVQA, a unified visual question answering benchmark for comprehensive evaluation of ophthalmic VLMs. EyeVQA is constructed from 21 available ophthalmic datasets and contains 20,000 question-answer pairs spanning six disease groups and seven question types: Single-Choice, Multi-Select, Variable-Select, True-False, Ranking, Point Location, and Bounding Box. Gold answers are deterministically derived from source-provided diagnoses, severity grades, clinical findings, segmentation masks, bounding boxes, and anatomical landmarks, enabling reproducible evaluation without relying on model-generated annotations. Notably, 44.5% of the questions require reasoning across multiple images, extending evaluation beyond conventional single-image medical VQA. We benchmark fourteen representative general-purpose, scientific, and medically specialized VLMs under a unified zero-shot protocol. The best-performing model only achieves an overall score of 62.8, while substantial gaps remain in spatial grounding and cross-task generalization. These results highlight the limitations of current VLMs in comprehensive ophthalmic visual understanding and establish EyeVQA as a diagnostic benchmark for developing more reliable and spatially grounded ophthalmic multimodal models. The project page is available at https://github.com/PKUTHM/EyeVQA.
Sep 26, 2026

Contamination, Prior, or Evidence? Decomposing and Training Evidence Use in Whole-Slide Vision-Language Models

Pathology vision-language models (VLMs) are conventionally evaluated by accuracy, but accuracy alone does not measure evidence use: it may conflate dataset contamination, prior knowledge, and image evidence. In a motivating study of lymph-node metastasis prediction, we found that most public pathology VLMs showed minimal differences when changing from feeding the models with whole-slide images, an annotated lesion, or no image at all. To better understand the specific features leveraged by these models, this paper presents two contributions aimed at disentangling these factors. First, we present CleanSlide, a TCGA-based VQA benchmark designed to eliminate image- and question-side contamination. It contains 149K audited multiple-choice questions over 9,985 slides, with patient- and tissue-source-disjoint splits. Every question is audited for option shortcuts, stem leakage, cross-split duplication, and blind solvability. Second, we propose Pair-DPO, a preference loss over counterfactual slide pairs from the same question and source. By controlling for shared confounding factors, Pair-DPO cancels out the question-attributable signal and leaves image evidence as the source of preference. Specifically, each pair consists of two real slides with opposite, verified findings, introducing neither editing artifacts nor unverified labels for diffuse or graded features such as invasion, necrosis, and tumor grade. Experiments show that our method gains 15.29% from image evidence on the CleanSlide, compared with 2.81% for the best published model. On the external CPTAC and BCNB cohorts, our method achieves accuracies of 57.6% and 59.0%, outperforming all other evaluated models by 9.7% and 3.4%, respectively. We will release the benchmark and code.
Sep 25, 2026

From Reward Signal to Visual Utility: A Controlled Audit of Medical VLM Post-Training

Medical vision-language model (VLM) post-training is commonly evaluated through answer accuracy. We examine how changes in accuracy and training objectives relate to image-conditioned decisions in a controlled Qwen2.5-VL-3B study on PMC-VQA. We compare supervised fine-tuning (SFT) with low-rank adaptation (LoRA) restricted to the language model, expanded multimodal adaptation scopes, standard answer-only Group Relative Policy Optimization (GRPO), and a counterfactual evidence objective. On 2,000 clean-test questions, language model LoRA SFT changes correct-image accuracy by +1.10 percentage points (95% paired bootstrap CI:-0.85 to +3.05), while visual-benefit events decrease by 2.40 points and image sensitivity decreases by 5.60 points. Paired records reveal 155 acquired and 203 lost visual-benefit events. Broader adaptation yields lower correct-image accuracy than language-model LoRA SFT. Standard GRPO produces mixed-reward groups and parameter updates, with an uncertain clean test accuracy change. A generation audit reveals that canonical option scores can follow a different token path from generated answers. With scores taken along the greedy generation path, the evidence target improves on the training set; its gains over standard GRPO remain inconsistent on validation data at matched training doses. Sample-level analyses trace how evidence scores, decision margins, and generated answers change during post-training. This empirical and measurement audit identifies gaps between optimization activity, target acquisition, and useful held-out visual behavior.
Sep 21, 2026

Representation-guided in-context learning for medical image interpretation with multimodal large language models

Medical image interpretation is central to diagnosis and care, yet adapting general-purpose multimodal large language models (MLLMs) often requires resource-intensive domain-specific fine-tuning. Here we introduce representation-guided in-context learning (RG-ICL), a training-free inference framework that retrieves query-aligned demonstrations using frozen encoders, without task-specific parameter updates. Across eight datasets spanning histopathology, radiology and retinal fundoscopy, RG-ICL improved classification (mean gain 20 percentage points) and visual question answering (VQA) (mean gain 13 percentage points) over no-context and conventional ICL, approaching or exceeding training-based comparators. Which cases were retrieved mattered more than how many: 6 query-aligned cases outperformed up to 32 randomly selected ones, whereas fixed or random cases often reduced accuracy below baseline. For VQA, aligning reference cases with both image content and question intent produced further gains. These findings indicate that for medical image interpretation, curating which reference cases an MLLM sees is a practical alternative to retraining it.