Label-Free Foundational Model Selection for Medical Image Classification under Distribution Shift via Pseudo Label Discrepancy

Foundation models are increasingly deployed for medical image analysis. However, under the inter-institutional distribution shift typical of deployment, their performance varies widely and cannot be known without target-domain labels, which are rarely available. This leaves a practical question unresolved: given several candidate foundational models and labeled-data from a source domain, which one to deploy in an unlabeled target domain? We propose a label-free selection criterion built on SUDO, a framework for evaluating clinical AI systems without ground-truth annotations. SUDO partitions the unlabeled target data by predicted probability and, for each region, measures a pseudo-label discrepancy reflecting class contamination; aggregated across regions, this yields a score (AURCC) requiring neither target annotation nor fine-tuning. We show that AURCC can be used to rank a variety of vision-language models (BioMedCLIP, CXR-CLIP, CheXzero, MedCLIP, MedImageInsight, CLIP) on chest X-ray classification across three inter-hospital shift scenarios, under zero-shot and MLP-probe regimes. The AURCC ranking recovers the ground-truth ranking with Spearman rho up to 0.943 (p<0.05). Against the natural baseline of ranking by held-out source accuracy, AURCC is competitive when the labeled source is large and yields a more accurate ranking once it is small; the regime of interest in resource-constrained settings.

Clinically Structured Surrogate Rewards for Post-SFT Medical Image Captioning

Medical image captioning requires translating heterogeneous visual evidence into concise clinical descriptions, where errors in findings, assertion states, or anatomical relations can alter clinical meaning despite surface-level fluency. Sequence-level policy optimization can directly optimize complete captions, but common rewards rely on global text similarity, direct image-caption compatibility, or unordered concept overlap, leaving visual neighborhoods and clinical-claim structure implicit. We propose a clinically structured surrogate reward framework for post-SFT medical image captioning. The framework combines biomedical semantic and short-range lexical fidelity with two structured rewards: distributional image-neighborhood alignment, which matches the medical-image-bank distributions induced by reference and generated captions, and clinical graph consistency, which applies maximum-weight one-to-one matching to entities, assertion states, and typed relations. The four rewards are independently normalized within each rollout group, combined with fixed relative weights, and optimized with GDPO. Across organizer-evaluated hidden test sets for the Standard and Synthetical ImageCLEFmedical Caption tracks and three vision-language backbones, the method improves Overall, Relevance, and Factuality over matched SFT baselines in all six backbone-track combinations, with average relative gains of 3.4%, 2.1%, and 5.8%, respectively. Ablations and paired diagnostics indicate that the structured rewards provide complementary signals, reducing image-neighborhood divergence and improving entity-assertion-relation consistency.

VLM- and LLM-Driven Multi-Agent System for PET Image Denoising

Positron emission tomography (PET) imaging suffers from limited spatial resolution and low signal-to-noise ratio, which can compromise quantitative accuracy and lesion detectability. Deep learning-based denoising methods have demonstrated strong potential for improving PET image quality. However, their practical deployment in real-world settings remains challenging, often requiring multiple specialized models and expert interventions, such as identifying motion-induced misregistration artifacts, estimating noise levels to select an appropriate denoiser, and performing lesion-focused quantitative assessment after denoising. Recent advances in vision-language models (VLMs) for image quality understanding and large language models (LLMs) for contextual reasoning provide new opportunities for automated, decision-driven workflows. Inspired by expert workflows for PET image quality enhancement, we propose an VLM- and LLM-driven multi-agent PET denoising framework that dynamically assesses image quality and lesion status, autonomously selects optimal denoising models and parameters, and enables closed-loop feedback with rollback mechanisms. Experiments were conducted on Siemens Biograph Vision Quadra PET/CT data with 1/20 and 1/50 low-dose settings. Individual module evaluations demonstrated the reliability of the agentic components, while the complete framework achieved higher PSNR and SSIM than UNet, GAN, and DDPM baselines at both dose levels. These preliminary results demonstrate the feasibility of using a closed-loop multi-agent framework to adapt PET denoising strategies to different image conditions.

Representing Visual Evidence for Item Difficulty Prediction: Visual Textualization and Image-Native Modeling

Predicting item difficulty from content can provide an initial estimate for newly developed questions before sufficient student responses are available. Existing approaches typically represent the question stem and answer choices as text. When mathematics items contain visual components, a common pipeline first textualizes that evidence and then applies a text predictor. We ask: how should visual evidence be represented for item difficulty prediction? We compare question text alone, visual textualization, which expresses visual evidence in language, and image-native modeling, which retains the original image. Using Eedi items with difficulty calibrated from student responses, we train large language models (LLMs) and vision-language models (VLMs) directly for difficulty regression. Both visual interfaces achieve the lowest point estimates, although the leading systems cannot be reliably ordered. Open-VLM textualization yields lower RMSE point estimates for all evaluated LLMs, while broader adaptation does so for all image-native VLMs. Test-time interventions show dependence on the paired full-item image, but do not isolate the additional visual component. The two visual interfaces also make partially complementary item-level errors and differ substantially in computational workflow. Thus, textualization should not be treated as the only practical interface: image-native modeling is a competitive alternative whose effectiveness depends on how the VLM is adapted.

MedARC: Training-Free Adaptive Redundancy Compression of Visual Tokens for 3D Medical Vision-Language Models

Integrating 3D medical images with vision-language models (VLMs) holds substantial promise for computer-aided diagnosis. However, volumetric images generate prohibitively long visual-token sequences with considerable spatial and inter-slice redundancy. Existing token compression methods typically apply uniform reduction or rely on a single importance signal, increasing the risk of removing regions that are clinically relevant to the query or structurally distinctive. To address this limitation, we propose MedARC, a unified, training-free framework for Adaptive Redundancy Compression of visual tokens in 3D medical VLMs. MedARC estimates token importance by integrating three complementary cues: self-attention from the VLM vision encoder, which reflects the model's intrinsic visual focus; similarity between projected visual tokens and text embeddings, which identifies query-relevant regions; and deviations of local visual foundation model features from the volume-level feature center, which highlight structurally distinctive anatomy. The resulting importance distribution guides a saliency-aware merging strategy that preserves informative tokens while consolidating redundant ones rather than simply discarding them. Experiments on CT-RATE and MR-RATE show that MedARC reduces visual-token overhead and inference time while preserving or improving diagnostic performance. Its multi-cue scoring cost is outweighed by the savings from processing fewer tokens, with greater benefits expected for larger language models.

Rethinking Clinical Relevance in Chest X-ray Machine Learning: How Evaluation References Define Performance

Chest X-ray (CXR) machine learning relies heavily on automated evaluation using reference standards that aim to approximate clinical judgment. However, commonly used report-derived labels for pathology classification or generic image quality metrics for reconstruction may not reliably reflect clinical judgment. We systematically investigate how evaluation-reference choices affect model performance and ranking in both pathology classification and image quality assessment (IQA). To enable controlled comparison across evaluation references, we collected paired expert image- and report-derived labels for thoracic findings from a clinical cohort at Cambridge University Hospitals (CUH) and curated a subset of the public MIMIC-CXR dataset, along with expert ratings of diagnostic image quality. We show that for supervised image classifiers (ResNet, DenseNet), several zero-shot and fine-tuned vision-language models (e.g., MedKLIP, GLoRIA, and ConVIRT), changing the label source leads to substantial differences not only in performance estimates but also in model rankings. In parallel, alignment of IQA measures with expert judgment depends heavily on the choice of measure, and commonly used IQA metrics such as SSIM and PSNR often fail to align with expert assessments of diagnostic usability. Our results demonstrate that evaluation choices are crucial: they can determine which models and methods appear best and are therefore selected for further development or deployment. The selection of evaluation references should therefore be treated as a central component of clinical validity in CXR machine learning, and justified with respect to the pathology, imaging task, and intended downstream clinical use.

PathSelect: Sequential Token Selection for Whole Slide Pathology

Gigapixel Whole-Slide Images (WSIs) present a fundamental computational bottleneck for vision-language models (VLMs) due to extreme sequence lengths. Existing approaches predominantly rely on spatial sampling or training-free pruning, which risk diluting weak but informative signals, leading to the loss of critical diagnostic evidence due to the spatially diffuse nature of pathological cues. We reformulate WSI token pruning as a sequential selection process, enabling the model to autonomously learn an optimal routing strategy rather than relying on static heuristics. We herein propose a decoupled routing framework integrated as an active plugin into the fully pre-trained SlideChat base model, leaving both the slide encoder and large language model frozen. To provide continuous gradients for the non-differentiable pruning operation during training, we introduce PathSelect. PathSelect employs a variance-preserving noise gate to modulate each patch's information flow via a differentiable Soft Top-K operator, paired with a diagonal-attention Denoiser that recovers the perturbed representations without semantic leakage. At inference, the PathSelect module is entirely detached. Relying solely on the trained Scorer, a deterministic Hard Top-K operator executes adaptive, data-dependent trajectory termination, significantly accelerating downstream generative processing with exceptionally low sequential token selection latency. Driven by an empirical average of only 44.86 tokens under a maximum constraint of K = 128, our framework achieves 74.00% overall accuracy on SlideBench (TCGA), representing an approximate 36.6x spatial token reduction relative to the uncompressed baseline average while consistently outperforming sampling-based counterparts.

Explaining BiomedCLIP with Weighted Banzhaf Interactions Supported by Tree-Gram Parsing

Vision-Language Models (VLMs) are demonstrating significant capabilities in medical tasks like radiology analysis, yet providing faithful and interpretable explanations remains a key consideration for their responsible deployment in clinical settings. However, existing explanation methods, such as the widely used FIxLIP framework, often struggle with the fine-grained nature of modern tokenizers. The tokenization problem fragments clinical concepts---splitting terms like "saddle embolus" into scattered, meaningless subwords---which leads to noisy, semantically incoherent cross-modal attributions. Such fragmentation also results in a combinatorial explosion of interaction possibilities, obscuring the model's true reasoning. To address this, we introduce ParseFIxLIP, an extension that incorporates the Tree-Gram Parsing into the Banzhaf interaction game used by FIxLIP. This semantically informed strategy utilizes dependency parsing trees to define explanation players by grouping related text tokens into semantically coherent units. Our smart_depth grouping strategy, merging tokens according to spaCy token dependency tree, successfully mitigates concept fragmentation, yielding substantially more interpretable cross-modal interactions by unifying complex medical concepts. Quantitatively, while baselines struggled with the high dimensionality of long captions, our parsing approach maintained statistical robustness and semantic parsimony. Qualitative analysis on BiomedCLIP, validated on medical imagery (ROCOv2) and general examples, confirms that the approach accurately captures the synergistic influence of grouped words on model predictions. In conclusion, our work offers intuitive and clinically relevant insights into VLM decision-making, fulfilling the critical need for coherent explanations in the medical domain.

Memory-Supported Synergistic Adaptation for Training-Free Test-Time Medical Image Segmentation

Test-time adaptation (TTA) aims to mitigate distribution shifts by adapting models with unlabeled target data at inference time. While TTA with vision-language models (VLMs) has shown promising results in classification, extending it to medical image segmentation remains challenging. In this setting, the adaptation gains from optimizing on VLM-generated predictions are often outweighed by the degradation to the VLM's strong pretrained features caused by noisy, update-driven learning, resulting in limited and unstable improvements. We therefore propose Memory-Supported Synergistic Adaptation (MSSA), a novel training-free TTA framework for medical image segmentation. Without updating model parameters, MSSA dynamically selects reliable image-text predictions to construct an online memory, uses them as text-guided semantic priors, and couples them with cross-image structural alignment for robust adaptation. Specifically, MSSA consists of (i) a noise-aware memory construction module that filters and stabilizes cross-modal predictions, and (ii) a relevance-driven prototype alignment module that aligns the target sample with structurally consistent memory samples and their reliable predictions to improve adaptation. Extensive experiments on multiple medical segmentation benchmarks demonstrate that MSSA consistently improves VLM-based segmentation models and outperforms existing fine-tuning-based TTA methods by a clear margin, with gains of up to 12.2% DSC and 11.7% mIoU. Project page: https://lingrayy.github.io/MSSA/ .

Model Merging for Medical LVLMs: A Benchmark and a Winner-Take-All Approach

Large vision-language models (LVLMs) can be adapted to specialized medical imaging tasks via parameter-efficient fine-tuning approaches such as low-rank adaptation (LoRA), leading to a growing ecosystem of expert models tailored to specific imaging modalities and clinical scenarios. However, deploying multiple expert LVLMs in practice incurs substantial computational and operational overhead. Model merging provides a promising solution by consolidating multiple experts into a single model without retraining, yet it remains largely unexplored in the medical domain. In this work, we present the first systematic study of model merging for medical LVLMs. We introduce MergeMedBench, a comprehensive benchmark spanning eight imaging modalities and diverse clinical task types, comprising 16 LoRA fine-tuned models built upon two mainstream architectures. We conduct an extensive evaluation of existing merging methods and further propose winner-take-all, a simple and hyperparameter-free approach that retains only the most dominant parameters across expert models. By preserving the critical parameters that govern model behavior and discarding weaker ones, our method avoids the information dilution inherent in averaging- or alignment-based strategies. Despite its simplicity, winner-take-all consistently outperforms existing approaches, offering both a new perspective on LoRA merging and a strong practical baseline for future research.

When Can Test-Time Adaptation Help Zero-Shot CT Vision-Language Models?

3D CT vision-language models (VLMs) classify abnormalities from text prompts in a zero-shot manner, enabling cross-institution deployment where labels are scarce and clinical tasks shift faster than supervised models can be retrained. A real CT scan, however, typically contains several co-occurring abnormalities, and the reliability of zero-shot multi-label prediction under distribution shift remains poorly understood. Test-time adaptation (TTA) updates a model on unlabeled target scans without source data or target annotations, yet existing TTA methods target multi-class softmax prediction on natural images or 2D medical segmentation, and none addresses unsupervised multi-label adaptation for zero-shot 3D CT VLMs. We study when TTA helps zero-shot 3D CT VLMs. A controlled diagnostic analysis shows that TTA is conditional: the volumetric input must preserve the encoder's depth structure, and the base representation must transfer to the target cohort, with depth reduction alone lowering internal AUROC by more than 0.12. We then focus on the regime where the base model already separates present from absent abnormalities. We introduce CARVE (Cardinality-Aware Retained-View Entropy), the first TTA method for this setting. CARVE estimates a sample-specific positive-label cardinality $\hat{k}$, optimizes a top-$\hat{k}$ objective to preserve co-occurring abnormalities, and performs memory-efficient multi-view adaptation by scoring weak 3D views without gradients before updating on a retained subset. Across contrastive CT-CLIP and anatomy-aware fVLM, CARVE provides the most consistent improvements across multi-label, three-class, and binary CT tasks when the base model is already discriminative. These results establish multi-label TTA for zero-shot 3D CT VLMs as a distinct problem and CARVE as a cardinality-aware solution.

TMF-RSE: Tri-Modal Fusion with Regional Semantics and Evidential Uncertainty for Lung Severity Scoring

Accurate quantification of lung disease severity from chest imaging is critical for clinical decision-making and resource allocation. We propose a tri-modal deep learning framework, TMF-RSE (Tri-Modal Fusion with Regional Semantics and Evidential Uncertainty), that combines appearance features from two-dimensional chest inputs, structural features from lung segmentation masks, and semantic features from vision-language models (VLMs) for severity quantification. Our approach employs complementary fusion mechanisms that integrate semantic guidance, structural priors, and hierarchical interactions across modalities. The model employs evidential regression to provide both severity predictions and uncertainty estimates. Experiments on the Per-COVID-19 CT and RALO datasets show that TMF-RSE outperforms recent transformer-based baselines, achieving MAE of 4.02 and Pearson correlation of 0.9629 on Per-COVID-19 validation, and 0.339 MAE / 0.973 PC on RALO geographic extent.

BioMedVR: Confusion-Aware Mixture-of-Prompt Experts for Biomedical Visual Reprogramming

Recent advances in vision-language models (VLMs) such as CLIP have demonstrated strong generalization across natural-image domains. However, adapting these models to biomedical imaging is non-trivial: full-model fine-tuning is computationally expensive, while medical data are often scarce and exhibit subtle, fine-grained inter-class differences, making parameter-efficient adaptation particularly critical. Visual Reprogramming (VR) offers a parameter-efficient alternative by injecting learnable perturbations into the input space, but existing VR approaches for VLMs mainly focus on positive class prompts and overlook confusing negatives, leading to miscalibrated predictions in fine-grained medical scenarios. We present BioMedVR, the first VR-based framework for biomedical imaging, enabling few-shot adaptation of pretrained VLMs through compact learnable VR modules. To mitigate class confusion, we introduce a Confusion Minimization Mechanism that leverages LLM-generated confusion-aware attributes together with a Confusion-Suppression Loss to explicitly reduce false-positive alignment. Moreover, the designed Mixture-of-Prompt Experts combines a positive expert for main-class discrimination and a negative expert for confusion suppression, balanced via adaptive gating. Extensive experiments on 18 datasets, including 11 biomedical datasets and 7 natural image benchmarks, demonstrate that BioMedVR achieves superior accuracy and generalization, effectively bridging VR and VLMs in biomedical domains.

CADRE: Stable, Parameter Efficient Adaptation of Medical Vision Language Models with Bounded Forgetting and Prior Drift

Medical vision-language models (VLMs) such as BiomedCLIP generalize broadly, but adapting them to a clinical service is as much a safety problem as an accuracy one. Updating a deployed model for a new imaging modality can fail silently in two ways that harm patients: it can forget modalities it already handled (catastrophic forgetting), and it can drift from its trustworthy pretrained prior toward modality-specific shortcuts. We study parameter-efficient continual adaptation through these two properties rather than leaderboard accuracy, presenting CADRE: a frozen-backbone framework combining low-rank adaptation (LoRA) with an online, self-scaling, similarity-aware elastic weight consolidation term that bounds retained-competence loss, and an anchor-to-prior penalty bounding embedding drift from the frozen prior. Two short guarantees, a bound on total consolidation mass and a scale-invariance property, remove the scale-related sources of vanilla EWC's order fragility. Using breast cancer across three maximally dissimilar modalities (histopathology, ultrasound, chest radiography) as a controlled cross-modality stress test, under a multi-seed, multi-order protocol with paired significance testing and training approximately 0.23% of parameters, CADRE attains the highest accuracy, SPQ, and backward transfer and the lowest forgetting among adapting methods, reducing forgetting roughly sevenfold versus the strongest regularized baseline (0.075 to 0.011; paired p=0.023) and achieving positive backward transfer where every baseline is negative. We frame these as stability properties aligned with clinical-safety desiderata, not a deployment guarantee; robustness to distribution shift and adversarial inputs is out of scope.

Hallucination Detection and Correction in Medical VLMs via Counter-Evidence Verification

Vision-Language models (VLMs) reliability in medical diagnosis is challenged by trust-undermining hallucinations. Existing hallucination detection approaches mainly focus on identifying factual inconsistencies between generated text and reference data. While some studies analyze where models attend in images, they seldom verify whether such attention truly reflects the visual evidence supporting the generated text. To address this gap, we propose Co}unter-Evidence Verification (CoEV), a training-free plug-and-play framework that detects and corrects hallucinations through evidence-based factual consistency verification. CoEV performs bidirectional verification between textual assertions and visual evidence, testing whether each statement is supported by its corresponding evidence region, and assigns each statement into a four-quadrant diagnostic map capturing combinations of text factuality and visual grounding. CoEV detects hallucinated content and serves as a post hoc refinement tool, correcting hallucinations without retraining. Extensive experiments on four medical datasets show that CoEV combats hallucinations in VLMs.For hallucination detection, CoEV consistently outperforms existing methods, improving average PR-AUC and ROC-AUC by 3.0% and 3.9% absolute points respectively, with notable gains of up to 18.5% in specific VQA scenarios. For hallucination correction, it improves Micro-F1 by up to 12.5%, reduces hallucination rates by over 11.9% on medical report generation, and also boosts medical VQA accuracy. These results show that CoEV enables reliable detection and correction of hallucinations, providing clinicians with dependable, evidence-based cues for diagnosis. Code will be released upon acceptance.

ReportQA: QA-Based Radiology Report Evaluation

Radiology report evaluation is essential for advancing automated report generation. Natural language generation metrics have limited clinical relevance. Clinical efficacy (CE) metrics evaluate important medical findings, but focus mainly on presence and cover only a limited set of entities. Due to heavy reliance on manual annotations, it is difficult for CE metrics to extend clinical entities or attributes. In clinical practice, radiology reports serve as a medium for information transfer. Clinicians use them to perform downstream diagnostic tasks without directly inspecting images. Based on this insight, we propose ReportQA, a clinical-related and flexible radiology report evaluation framework, supporting detailed quantitative analysis of radiology report generation systems. We first collect datasets covering multiple imaging modalities and anatomical regions. We then construct knowledge trees of clinical entities and attributes with radiologist guidance, and use large language models (LLMs) to extract structured information from raw reports. Next, we generate QA pairs from predefined templates and apply quality control through self-filtering and report-based filtering. During evaluation, the report is treated as context, and an LLM acts as a judge model to answer the QA pairs. Based on the resulting QA accuracy, we introduce QAScore metric. Compared with existing metrics, QAScore shows better alignment with radiologist judgments. Experiments on multiple state-of-the-art vision-language models reveal that current report-based inference paradigms struggle to learn fine-grained clinical representations and exhibit strong negative prior biases. In contrast, question-driven inference provides a more effective alternative. For reproducibility and extensibility, we release the knowledge trees, structured reports, and QA pairs, along with the pipeline code for QA construction and evaluation.

Wasserstein-Aligned Localisation for VLM-Based Distributional OOD Detection in Medical Imaging

Zero-shot anomaly localisation via vision-language models (VLMs) offers a compelling approach for rare pathology detection, yet its performance is fundamentally limited by the absence of healthy anatomical context. We reformulate zero-shot localisation as a comparative inference problem in which anomalies are identified through structured comparison against reference distributions of normal anatomy. We introduce WALDO, a training-free framework grounded in optimal transport theory that enables comparative reasoning through: (i) entropy-weighted Sliced Wasserstein distances for anatomically-aware reference selection from DINOv2 patch distributions, (ii) Goldilocks zone sampling exploiting the non-monotonic relationship between reference similarity and localisation accuracy, and (iii) self-consistency aggregation via weighted non-maximum suppression. We theoretically analyse the Goldilocks effect through distributional divergence, and show that references with moderate similarity minimize a bias-variance trade-off in comparative visual reasoning. On the NOVA brain MRI benchmark, WALDO with Qwen2.5-VL-72B achieves $43.5_{\pm1.6}\%$ mAP@30 (95\% CI: [40.4, 46.7]), representing a 19\% relative improvement over zero-shot baselines. Cross-model evaluation shows consistent gains: GPT-4o achieves $32.0_{\pm6.5}\%$ and Qwen3-VL-32B achieves $32.0_{\pm6.6}\%$ mAP@30. Paired McNemar tests confirm statistical significance ($p<0.01$). Source code is available at https://github.com/bkainz/WALDO_MICCAI26_demo .