Overview

Medical image analysis develops methods to detect disease, segment anatomy, track changes over time, reconstruct images, and estimate clinically useful measurements from modalities such as CT, MRI, X-rays, ultrasound, and retinal imaging. Common approaches include U-Net-style segmentation, transformers, self-supervised and foundation-model learning, multimodal vision-language systems, and methods that incorporate anatomical or physical constraints. Models are tested on tasks and benchmarks such as BraTS for brain tumors, ACDC for cardiac MRI, and EchoNet for echocardiography, with growing attention to external-site testing, calibration, and clinical workflow. A central challenge is that strong benchmark results may not transfer across scanners, hospitals, patient groups, or acquisition protocols, so reliable deployment requires better validation, uncertainty estimates, and safeguards against bias and privacy risks.

Latest papers

309 papers28 papers in the last four weeks, down 10% on the four weeks before. 0.2% of all new papers.

Nov 13, 2026

CardioTrace: Motion-Aware Cine Synthesis via Coupled Tracing and Segmentation

Cardiac cine MRI provides time-resolved visualization of cardiac motion and is essential for evaluating ventricular function and anatomy. In routine clinical acquisition, however, cine sequences are frequently temporally sparse or degraded by arrhythmia, motion, and scan-time constraints, which can hinder reliable quantification and downstream analysis. We present CardioTrace, a motion-aware cine synthesis framework that generates a full cardiac cine sequence from limited input while explicitly modeling and tracing cardiac motion. CardioTrace couples two complementary pathways: a pixel-translation branch that directly predicts cine frames in a residual formulation, and a flow-field branch that estimates time-resolved deformation fields from a reference phase to provide interpretable motion correspondences across frames. To further preserve anatomical plausibility, a segmentation branch supervises clinically meaningful cardiac structures at key phases, and a cross-branch consistency objective encourages the synthesized appearance dynamics to follow the explicit motion guidance, mitigating temporal drift and unrealistic shape transitions over long sequences. The framework also supports prompt-style conditioning to incorporate auxiliary patient or acquisition information, improving robustness across heterogeneous multi-center data. Experiments on public cardiac cine MRI datasets demonstrate that CardioTrace improves perceptual quality and temporal coherence over representative GAN- and diffusion-based baselines, while better preserving ventricular morphology throughout the cardiac cycle. Our codes are available after review.
Nov 13, 2026

X-Strata: Separating Aneurysms and Overlapping Vessels into Layers from 2D X-ray Angiography

An X-ray angiogram sums every structure along a ray into one pixel. When vessels cross an aneurysm, they obscure part of it and mix their contrast into its filling curve. We present X-Strata, a network that separates a cine angiogram into three layers: the complete aneurysm sac, the vessels in front of it, and the vessels behind it. Contrast timing indicates how much each layer contributes to a pixel, and parallax between two views indicates their order, which a single view cannot determine. X-Strata first estimates the layers and their contrast curves, then uses the second view to assign vessels to the front or back layer, and lifts all layers into 3D. It learns from layer labels projected from same-session 3D rotational angiography, but needs only 2D cine runs and their acquisition geometry at inference. On 220 held-out patients from three centers (122 external), two-view X-Strata ranks first among 14 baselines, including VGGT and C2^2RV, in recovering the sac where vessels overlap it (Dice 0.693 internal, 0.641 external) and in ordering the layers (balanced accuracy 0.703, 0.659). Its order accuracy exceeds that of all 11 end-to-end baselines by at least 4.1 points internally and 3.3 externally (Holm-adjusted p<0.001p < 0.001) and grows with view separation. Recovered 3D sac surfaces have a median Chamfer distance of 0.34 mm on the internal cohort. X-Strata may therefore provide layered aneurysm and vessel information, including 3D sac shape, for patients who undergo only 2D angiography.
Nov 12, 2026

Automated PD-L1 Tumor Proportion Scoring with Uncertainty-Guided Triage in Non-Small Cell Lung Cancer

Programmed death-ligand 1 (PD-L1) tumor proportion score (TPS) informs immunotherapy selection in non-small cell lung cancer. Near the 1% and 50% cutoffs, the score should also indicate when closer review is warranted. We designed a three-stage framework to follow the sequence of TPS assessment. Stage 1 identifies tumor regions and excludes macrophage-rich patches. Stage 2, a dual-branch cell counting network (D2Count-Net), counts PD-L1-positive and PD-L1-negative tumor cells in these regions to calculate TPS. Stage 3, TPS-Guard, an uncertainty-guided triage mechanism, reports model uncertainty as the disagreement among independently trained models, refers the most uncertain slides, and flags scores near the 1% or 50% cutoff when model estimates disagree across that cutoff. Flags state the reason for referral or alert the pathologist to possible scoring errors on auto-reported slides. In a retrospective, single-center study, 154 whole-slide images were scored by three pathologists, and the model and each pathologist were compared in turn with the mean of the other two. On the 114 slides unseen in training, model mean absolute error (MAE) was 6.96 percentage points, within the range of the three pathologists (4.24 to 8.29), and model uncertainty correlated with absolute TPS error (Spearman rho = 0.77) and inter-pathologist disagreement (rho = 0.71). After referral of the 15% most uncertain slides, model and pathologist MAE both decreased. Applied without retraining to an external, publicly available dataset, TPS-Guard still directed review to higher-error cases, and referring more uncertain cases raised sensitivity at the 50% cutoff. TPS-Guard provides reasons for reviewing individual automated scores and identifies slides that also produce greater disagreement among pathologists. Prospective reader studies should assess TPS-Guard's effect on TPS reporting.
Nov 10, 2026

MicroBT: Centroid-Guided Synergistic Learning for Multi-Modal Micro Brain Tumor Segmentation and Counting

Multi-modal brain tumor segmentation is essential for clinical diagnosis, radiotherapy planning, and longitudinal monitoring. While recent studies have achieved remarkable overall segmentation performance, they rely on overlap-based objectives that inherently favor large regions, causing sub-3 mm micro-lesions to be systematically suppressed during training. Moreover, existing pixel-level segmentation frameworks lack explicit instance-level modeling, leading to unreliable lesion counting that directly impacts clinical staging and therapeutic decisions. To address these challenges, we propose MicroBT, a unified multi-modal framework for micro brain tumor segmentation and counting. Our core insight is that a fundamental geometric property is inherently scale-invariant: regardless of size, each lesion possesses a well-defined centroid. Leveraging this, we first devise a centroid-anchored synergistic paradigm that utilizes predicted centroids as shared anchors to simultaneously condition dense segmentation and instance-level localization. To counteract the tumor size imbalance, we further devise a Volume-Calibrated Loss (VC-Loss) that adaptively amplifies the gradient contribution of micro-lesions, alongside a Size-Aware Curriculum Optimization (SACO) that stabilizes multi-task training by establishing global representations before progressively unlocking fine-grained localization. Extensive experiments demonstrate that MicroBT outperforms state-of-the-art methods, particularly on metastases, where micro-lesion (≤3\leq 3 mm) localization precision improves by approximately 7% and counting estimation errors decrease by approximately 26%. The code is available at https://github.com/leonlou921/MicroBT.
Nov 9, 2026

Answer or Defer? Testing VLMs' Stated Confidence on Brain MRI

Vision-language models (VLMs) could support medical-image interpretation, but their use requires deciding which interpretations to accept and which images to refer to a clinician. We test whether thresholding stated confidence can retain useful coverage while limiting errors among accepted answers, and how this depends on the population. Eight open-weight VLMs, including two reasoning models, assess 3,888 brain MRI slices from post-treatment glioma patients (BraTS 2024) and healthy volunteers (IXI) for tumor or gross abnormality. BraTS labels are mask-derived; IXI slices are treated as negatives. On BraTS, no threshold selects a nonempty set with at most 5% empirical error for any model, even when chosen retrospectively. This also holds for four models evaluated on a large-lesion, distant-negative subset. At a 10% error budget, Gemma-4-12B retains 24.4% of slices and every other model at most 2.5%. To investigate this limited selection ability, we separate confidence differences between predicted answers from differences within each answer. For four models with available outcome-level summaries, pooled mean confidence differs more between answers than between correct and incorrect responses within each answer. A covariance decomposition explains how this pattern can produce opposite confidence-correctness associations across populations. At a 20% error budget, accepting every "tumor" answer covers more slices than any confidence threshold for two models. These findings caution against using unvalidated stated confidence as a standalone deferral signal and support population-specific validation before deployment. Code is available at https://github.com/amir-sbg/VLM-Answer-or-Defer.
Nov 6, 2026

Dual-Region Slice Aggregation for CT-Based Renal Tumor Classification

Clear cell renal cell carcinoma (ccRCC) is the most common kidney cancer in adults. We present a region-aware architecture for CT-based ccRCC classification. Tumor and 0--10 mm peritumoral context are pooled separately from shared frozen DINOv3 feature maps. A shared position-aware slice aggregator precedes fixed training-only principal component analysis (PCA). The resulting representations are fused with radiomics using two scalar feature gates. Evaluation on KiTS23 and 111 external cases yields internal test and external AUCs of 0.8078\pm0.0154 and 0.9094\pm0.0015. Compared with matched single-tumor readout, mean AUC increases by 1.50 and 2.34 percentage points, respectively. Early-PCA and reconstruction controls yield lower internal mean AUCs, while external AUC is similar for early and late PCA. These findings highlight the joint role of regional readout and slice aggregation in adapting frozen visual features to renal CT.
Nov 5, 2026

VoluCompose: Constructing Compact Volumetric Visual Representations for Medical Image Understanding

Medical vision--language models can interpret volumetric images, but processing full volumes is costly because hundreds of slices produce long, redundant visual-token sequences. Existing compression methods often couple content selection with budget determination. We observe that neighboring slices share visual features and that attention concentration varies across visual encoders. We propose VoluCompose, a training-free, model-agnostic framework that separates what visual content to retain from how many tokens to allocate. Its hyperbolic token composition module groups contiguous slices into ordered support units and uses the Lorentz distance to account for feature magnitude and direction during token allocation, anchor selection, and refinement. Its capacity control module uses visual-encoder attention entropy to allocate more tokens when attention is more dispersed. We evaluate VoluCompose on two tasks across three datasets and four backbones. Across nine calibrated settings, VoluCompose attains the highest mean quality--retention hypervolume among the compared methods. It achieves a mean relative quality of 121.68% against the uncompressed references while retaining 22.78% of visual tokens on average. Compression remains robust even at a 10% token budget. Across settings, the lowest ROUGE-L ratio to the uncompressed reference is 96.4% for VoluCompose. The corresponding minima for the compression baselines range from 15.0% to 70.5%. Code is available at https://github.com/yezanting/VoluCompose.
Nov 5, 2026

TRiST: Topology-Reinforced Swin Transformer for Medical Image Analysis

Medical images contain topological patterns, such as connected components and loops, at scales ranging from small lesions to organ boundaries. Although hierarchical Vision Transformers have advanced medical image analysis, they do not explicitly incorporate topology throughout their multiscale representations. We propose Topology-Reinforced Swin Transformer (TRiST), a framework that integrates topological information into feature construction, refinement, and attention. First, a Hierarchical Multi-Scale Topology Representation algorithm computes 2D persistent homology (H0 and H1) independently at four native Swin stages, producing descriptors aligned with each stage's token grid rather than downsampling fine-scale topology. Second, a Band-Aware H0/H1 Residual Refinement module uses gated residual networks to process short-, medium-, and long-persistence information separately. Third, a Persistence-Band Progressive Attention Bias converts stage-specific descriptors into learned key-saliency scores that directly modify shifted-window attention logits. We instantiate TRiST as Topo-SwinV2-B for classification and Topo-Swin-UNet for segmentation, preserving the underlying Swin-based backbones. Experiments on five medical image datasets show consistent improvements over Swin-based baselines and competitive performance relative to existing topology-augmented methods.
Nov 4, 2026

A 3D foundation model unifying CT interpretation and segmentation

Radiologists interpret computed tomography (CT) scans by jointly determining what a finding represents and where it lies in three-dimensional (3D) anatomy. However, CT models have treated CT interpretation and 3D segmentation as separate tasks, leaving textual findings spatially unverifiable and overlooking potential synergy between clinical understanding and spatial perception. Here we present HoloCT, a generalist 3D foundation model unifying CT interpretation and voxel-level segmentation by jointly modelling whole-volume context and fine-grained anatomical detail, providing 3D masks as evidence. We curated HoloDATA, comprising 195,921 CT scans from 73 public datasets and six clinical cohorts, with 6.2 million instruction dialogues and 2.8 million CT-mask pairs. Across 346 tasks spanning segmentation, lesion screening, visual question answering, and report generation on 61 internal and 12 external datasets, HoloCT outperformed specialist and foundation models. In multicentre validation across 11,455 clinical CT scans from five hospitals, HoloCT outperformed competing models in lesion segmentation and report generation by 18.2 and 15.3 points. It also achieved 85.1% mean accuracy in screening thyroid tumours and colon lesions, exceeding all four participating radiologists in a retrospective reader study. Further analyses revealed bidirectional semantic-spatial synergy: language-guided segmentation delineated unseen anatomical structures and lesions as well as diffuse free-text findings, outperforming the leading approach by 22.1 points, whereas prompting the model to generate corresponding 3D masks alongside its textual outputs significantly improved interpretation performance. HoloCT shows that clinical understanding and spatial perception can be learned together within a single foundation model and offers a path towards more accurate and reliable CT analysis.
Nov 3, 2026

Diagnosis by Composition: Detecting and Localizing Unseen Diseases by Composing Learned Radiological Findings

Deep models for radiological diagnosis are typically trained to predict disease labels directly from images, which limits them to diseases seen during training and makes their decisions hard to interpret in radiologists' terms. Radiology, however, proceeds from visual findings to disease through published guidelines that make this decision rule explicit, interpretable, and reproducible. We introduce diagnosis by composition (DbC), a representation learning framework that encodes this findings-first logic. DbC learns a vocabulary of radiological findings by aligning image features with the text embeddings of the findings, each finding represented as a dense map over the scan. A disease is defined by a set of required findings derived from its guidelines, and its diagnosis map is a fixed, parameter-free composition of the corresponding finding maps. Because the finding-to-disease mapping is derived from guidelines rather than learned, DbC is interpretable and auditable by design, can extend to new diseases by specifying their findings without retraining, and can diagnose unseen diseases by composing their findings. On multi-phase liver CT, DbC detects and localizes unseen diseases, rare diseases from a different institution, and an external cohort significantly better than all comparators, including models that align images with disease names. DbC also improves detection and localization in the fully supervised setting and remains superior across training-set sizes. Although demonstrated on liver CT, DbC applies wherever radiological guidelines define diseases through findings.
Nov 3, 2026

Learning from Unlabeled Photographs for General Oral Image Analysis

Oral photographs contain rich clinical information and hold significant potential for artificial intelligence applications; however, most models are trained for individual tasks and rely on costly expert annotations. We hypothesized that self-supervised training on unlabeled oral photographs could produce transferable visual representations for heterogeneous vision tasks. To test this hypothesis, we developed DOVE (Dental Optical Visual Encoder) by domain-adaptive self-supervised pretraining on approximately 530,000 unlabeled oral photographs collected from publicly accessible sources. DOVE was benchmarked against 3 established baselines across 5 classification datasets and 4 segmentation datasets, covering both frozen backbone evaluation and end-to-end finetuning. DOVE achieved the strongest overall performance, ranking best in few-shot transfer and label efficiency on most benchmarks. Under the 20-shot frozen-backbone protocol, DOVE outperformed the second best DINOv3 by 4.37 percentage points in mean macro-F1 across classification datasets and by 4.49 percentage points in mean macro-Dice across segmentation tasks. Under end-to-end finetuning, DOVE achieved at least 80% of the strongest full-data baseline's performance across most evaluation datasets using only 10% of the labeled training data. Qualitative retrieval and feature analyses further indicated better organization of image-level and spatial representations. These findings support DOVE as a generalizable and label-efficient backbone for oral vision research.
Nov 3, 2026

Improving PI-RADS Prediction and Prostate Cancer Detection from Multiparametric MRI with Anatomical Guidance

Accurate prediction of Prostate Imaging Reporting and Data System (PI-RADS) scores from multiparametric MRI (mpMRI) remains challenging, more so than clinically significant prostate cancer (csPCa) detection. Given the importance of lesion location in PI-RADS scoring, we evaluated whether incorporating prostate zonal masks improves PI-RADS prediction and csPCa detection across multiple architectures (3D ResNet-18, ResNet-50, and ViT) using a large-scale internal dataset. Anatomical guidance generally improved performance across architectures, with the best results achieved by ResNet-18 with 2-zone guidance for PI-RADS prediction (QWK = 0.511) and ResNet-50 with 5-zone guidance for csPCa detection (AUC = 0.809). Fine-tuned results on external datasets further demonstrated the benefits of anatomical guidance. Furthermore, models with anatomical guidance improved the utility of model-predicted PI-RADS scores for PI-RADS-based csPCa detection (positive defined as PI-RADS>=3). Combining model-predicted and expert-assigned PI-RADS scores significantly increased the AUC from 0.711 to 0.756 (p < 0.05), suggesting that AI-derived PI-RADS scores may provide complementary information to expert assessment for csPCa detection. Overall, our findings highlight the value of incorporating prostate zonal anatomy into deep learning-based prostate MRI assessment.
Nov 2, 2026

Attention Residual U-Net for Enhanced Cross-Dataset Polyp Localization in Endoscopy

Precise localization of pathologies in endoscopy is a critical task in medical imaging, with deep learning (DL) frameworks now serving as the primary methodology for automated lesion detection. Today, generic object detection methods, such as YOLO, provide state-of-the-art performance in this context. However, their robustness often decreases under real-world clinical conditions due to variations across endoscopic devices, intestinal content, dataset bias, and the presence of small or low-contrast lesions. Motivated by these challenges, we propose a novel polyp detection framework based on a Attention Residual U-Net offering enhanced polyp localization. Instead of directly regressing bounding boxes from global feature representations, the proposed method learns lesion-focused spatial representations and converts them into bounding-box predictions during inference. Unlike conventional encoder-decoder localization frameworks that directly propagate all encoder activations into the decoding pathway, the proposed method selectively suppresses irrelevant background responses before feature reconstruction. This encourages the network to focus on lesion-relevant spatial evidence under heterogeneous imaging conditions. To achieve this, the framework integrates residual feature propagation, attention-guided encoder filtering, and multi-scale contextual learning. The proposed framework was evaluated on multiple public endoscopic datasets acquired using different devices. The comparative experimental results validate its effectiveness over the relevant state-of-the-art methods, with an improvement up to 22.0% in terms of the area under the receiving operating characteristic (AUC).
Oct 30, 2026

GazeRIB: Systematic Evaluation of Multi-Reader Gaze-Guided Chest X-ray Learning

Radiologists' visual attention can supervise medical image learning, but how reader panel composition affects its value remains untested. We evaluated GazeRIB (Gaze-guided Radiologist-Informed Blackout), which masks unviewed regions during training and uses unmasked images at inference. Eight experts repeatedly read the same 80 chest X-rays (CXRs). We compared individual readers with three-, five- and eight-reader panels across seven architectures and 10 seeds, evaluating 910 models on a separate 2,897-CXR cohort for four-class severity prediction. Every panel improved macro F1 over the no-gaze baseline by 5.33-7.47 percentage points (pp; Bonferroni-adjusted p < 0.007); no individual reader reached significance on this endpoint. Three readers captured 73-80% of the eight-reader gain in macro F1, the area under the receiver-operating-characteristic curve (ROC-AUC) and the area under the precision-recall curve (AU-PRC). On external validation without retraining, all four panels improved ROC-AUC on SIIM-FISABIO-RSNA by +1.93 to +4.24 pp (p <= 0.004); on MIMIC-Eye no panel differed significantly from baseline, whereas the single-reader average declined (-1.40 pp, p = 0.008). Reader composition therefore influenced performance, although transfer depended on the cohort.
Oct 29, 2026

DINOv2 for PET: A Pre-Training Domain Benchmark Across Classification, Segmentation, and Prognosis

DINO (self-DIstillation with NO labels) has become a stan- dard for self-supervised pre-training, producing features that generalise across datasets and downstream tasks. In medical imaging it has been adopted mainly for radiology (X-ray, CT and MRI), where even natural- image DINO features transfer well. In the literature, these features have been reported to transfer worse to functional imaging such as PET. Yet PET-specific DINO models have barely been introduced, and even less validated across multiple tasks and datasets. We present PET-DINOv2-2D, the first DINOv2 model pre-trained on 2D PET, and benchmark it against open-source encoders: the origi- nal DINOv2 (pre-trained on the LVD-142M natural-image dataset) and OmniRad (pre-trained on CT, MRI and ultrasound). Using the same lightweight heads for every backbone, and despite far less pre-training data, PET-DINOv2-2D matches or outperforms both on three down- stream PET tasks: normal/abnormal classification, lesion segmentation and recurrence-free survival prognosis. Its advantage is largest when few labels are available, suggesting that PET pre-training gives more transferable, label-efficient features. Code is made available at https://github.com/nuclivision/pet-dinov2-2d/.
Oct 29, 2026

DOSE: Dominance-Guided Ordinal Severity Estimation from Sparse Relative Annotations

Medical severity estimation commonly relies on absolute grade annotations, which are fundamentally ordinal, yet existing methods often treat the grades as independent categories. Learning-to-rank methods, which can be trained using relative annotations based on severity comparisons of image pairs, can provide continuous severity rankings but cannot directly predict discrete severity grades. We propose Dominance-guided Ordinal Severity Estimation from Sparse Relative Annotations (DOSE), a framework for learning ordinal medical severity from sparse pairwise comparisons without using absolute annotations. DOSE aggregates the observed comparisons through a global comparison graph to estimate a dominance score for each image. Differences between dominance scores set cumulative margins in the embedding space, so that pairs with larger estimated severity differences are separated more strongly, and also generate pseudo-pairs that supervise a scalar severity head alongside the observed comparisons. The learned continuous severity axis is converted into discrete grades via label-free score calibration, with tie-anchored segmentation as the main calibration method. In the experiment, DOSE consistently outperformed pairwise-ranking baselines on ACNE04, IDRiD, and LIMUC under different pair budgets. Code is available at https://github.com/TushirSahu/DOSE
Oct 27, 2026

Convex Potential Learning for Optimal Transport Between Medical Images

Computing optimal transport maps between medical images plays a fundamental role in morphological analysis. However, existing methods are either difficult to scale to large 3D medical images or rely on regularized formulations that provide only approximate transport maps. To improve the accuracy of downstream medical image analysis, we propose a novel approach for estimating optimal transport maps by parameterizing the convex potential function. By construction, the resulting transport map is guaranteed to be curl-free, consistent with Brenier's theorem. We evaluate the proposed method on both 2D and 3D medical imaging datasets. Experimental results demonstrate that our approach achieves lower transport estimation errors and produces higher-quality image reconstructions than existing methods.
Oct 27, 2026

MCP-Net: Real-Time Coronary Perception Network and Topological Analysis for Robotic Guidewire Navigation

Coronary guidewire navigation during percutaneous coronary intervention requires continuous interpretation of vessel anatomy, guidewire position, and target stenosis under two-dimensional X-ray fluoroscopy. While existing approaches commonly address vessel segmentation or device localization as separate tasks, a robotic system requires a more structured perception of the coronary scene to support downstream navigation. In this work, we present a simulation-driven framework centered on a Multi-Task Coronary Perception Network (MCP-Net) for robotic guidewire navigation, integrating coronary vessel segmentation, vascular topology extraction, and roadmap generation. Starting from three-dimensional coronary CT angiography, we generate simulated fluoroscopy images containing vascular structures, simulated guidewires, stenotic lesions with varying severity, and diverse vascular geometries, together with simulation-derived ground truth. MCP-Net jointly performs coronary vessel segmentation, guidewire-tip localization, and stenosis localization from a single fluoroscopic frame. These predictions are subsequently converted into a structured vascular graph containing centerlines, bifurcations, endpoints, branch directions, and branch angles. By associating the predicted stenosis with this graph, the framework generates a roadmap from the automatically selected graph root toward the target lesion. The results demonstrate that the proposed framework can jointly recover coronary perception targets and transform them into an interpretable vascular representation suitable for downstream robotic navigation.
Oct 26, 2026

Targeted Deep Regression - Initial Applications in Whole-Body Water-Fat MRI in UK Biobank

UK Biobank is a British clinical study which, in addition to large amounts of medical data collected, plans to include detailed Magnetic Resonance Imaging (MRI) examina- tions from 100,000 participants. The image data can be analyzed for example by auto- mated analysis with neural networks for image segmentation or classification to expand the collection of image-derived phenotypes for medical research. In this work, we pro- pose, present and evaluate a methodology that uses image-based regression to estimate numerical properties from the collected neck-to-knee water-fat MRI scans. It applies a ResNet50-style network to 2D multi-channel collages that represent targeted and study- specific anatomical sub-regions from the MRI scan. The targeted sub-region sampling is achieved by propagating manually defined subregions from a reference scan to the other >30,000 analyzed scans using image registration. The resulting collages can be compact while retaining high-resolution, for example by providing multiple slices of kidney tissue for a kidney volume regression network. This approach also enables group-wise saliency analysis that highlight image and anatomical regions used by the network for the pre- dictions. The method was applied for the regression endpoints of kidney volume, heart volume, and age. Predictive performance was found between R2 = 0.731-0.984 and the group-wise saliency maps were found to highlight medically relevant regions. We con- clude that the method can be used to standardize and restrict the information presented to a deep regression analysis pipeline, increasing information density and opening up a wide range of future medical inference studies.
Oct 24, 2026

End-to-end prostate lesion segmentation and detection from undersampled diffusion-weighted MRI

Automated prostate lesion detection from diffusion-weighted MR imaging (DWI) conventionally relies on reconstructed images and does not fully use all acquisition information. We present an end-to-end network that uses complex multicoil DWI data and coil sensitivity information for prostate lesion segmentation. We included 269 fastMRI prostate examinations with additional voxel-level lesion annotations. Of these, 64 examinations contained PIRADS 4 or 5 lesions. The proposed network, VarNet-Seg, consists of a k-space variational network (VarNet) with six variational cascades and an image-space 3D segmentation U-Net. Comparator models used the same segmentation architecture with fully coil-combined images (CC-Seg), fixed root-sum-of-squares reconstruction (RSS-Seg), or fixed SENSE reconstruction (SENSE-Seg). We performed ablations with end-to-end training and data consistency (DC) enabled or disabled. We evaluated lesion segmentation using Dice and patient-level detection using AUROC in five-fold cross-validation with full sampling, simulated undersampling, added noise, or both. VarNet-Seg achieved the highest Dice at every undersampling level. At eight-fold acceleration, Dice was 0.22 for VarNet-Seg, 0.14 for SENSE-Seg and 0.16 for VarNet-Seg (no DC). The paired difference from SENSE-Seg was 0.08 (95% CI, 0.04-0.11). With full sampling, VarNet-Seg and RSS-Seg performed similarly (0.40 vs 0.39). Patient-level AUROC decreased from 0.75 with full sampling to 0.63 at eight-fold acceleration; corresponding RSS-Seg AUROCs were 0.76 and 0.64. The proposed multicoil network improved prostate lesion segmentation under simulated undersampling compared to conventional models. Matched ablations supported contributions from both end-to-end training and data consistency. Prospective evaluation is needed to establish whether this approach can be used clinically in accelerated prostate DWI.
Oct 23, 2026

Beyond Discrete Tokens: Preserving Numerical Continuity in Medical Vision-Language Models

Objective: To develop a unified vision-language model for prompt-conditioned quantitative ultrasound measurement that addresses the mismatch between continuous measurement targets and discrete token-based numerical prediction. Methods: This article presents HAD-VLM, a hybrid autoregressive-diffusion framework that combines autoregressive text generation with continuous measurement prediction within a shared backbone. Visual prompt tuning adapts the vision encoder to ultrasound images. A diffusion process predicts a five-component representation comprising caliper endpoint coordinates and measurement length. A post-processing strategy combining the proposed coordinate-length consistency mechanism with ensemble prediction is introduced to refine measurement estimates. We evaluate four echocardiographic measurement tasks on EchoNet-LVH, with additional assessments of medical question answering and X-ray vertebral measurement. Results: HAD-VLM achieves mean absolute errors of 1.36, 2.77, 3.05, and 1.16 mm for IVSd, LVIDd, LVIDs, and LVPWd, respectively, reducing errors by 21.4-37.6% relative to the Vicuna-based autoregressive baseline. It also outperforms fine-tuned Qwen3-VL and direct MLP regression across all four tasks. Medical question answering performance is preserved on the evaluated dataset. For X-ray vertebral measurement, HAD-VLM reduces mean absolute error from 11.4 to 6.7 pixels. Conclusion: Integrating autoregressive language generation with diffusion-based continuous prediction improves quantitative medical image measurement while retaining prompt-conditioned task flexibility. Significance: HAD-VLM provides a unified approach to combining language interaction with anatomically grounded numerical prediction for quantitative medical image analysis.
Oct 23, 2026

VTRL-Med: Progressive Visual Refinement with Topology-Guided Regularization for Medical Image Clustering

Deep clustering offers a promising approach for analyzing large unlabeled medical image datasets, yet prevailing methods still face two limitations: they may neglect class-agnostic structural priors that capture object boundaries and regions, and they do not explicitly encode topological patterns (e.g., connectivity and holes). These sources of information play distinct and complementary roles: visual-structural features provide the primary discriminative representation for clustering, while topological features provide additional global geometric cues for residual visual ambiguities. We introduce VTRL-Med, a progressive visual-structural learning framework with topology-guided regularization for medical image clustering. VTRL-Med first adapts the Segment Anything Model (SAM) to extract image features that retain structural cues inherited from mask-supervised pretraining. It then generates reliable pseudo-labels through a two-step filtering process that combines prototype-based selection with k-nearest neighbor consistency validation. Finally, it progressively refines the visual branch while an auxiliary topology branch, constructed from persistent homology (PH) descriptors, provides additional geometric regularization during training. Across 13 medical image benchmarks, VTRL-Med achieves state-of-the-art clustering performance, improving average accuracy by 4.48% over existing methods. On three representative datasets, topology-guided regularization further improves accuracy by 1.3--4.3% over visual refinement alone.
Oct 23, 2026

OPTARI: An open-source software framework for clinical translation of optoacoustic imaging

Combined optoacoustic (OptA) and ultrasound imaging is a promising modality across a wide range of pathologies. Its use in clinical research, however, is limited by a lack of analysis software accessible to users without programming expertise. Existing tools are either script-based engineering frameworks or proprietary vendor applications. We present OPTARI, an open-source graphical software application to facilitate large-scale clinical optoacoustic-ultrasound (OPUS) studies. OPTARI builds on the napari image viewer and the PATATO toolkit for OptA data processing. It supports native vendor and open HDF5 data formats, aligns with the IPASC standard, and is distributed as a standalone executable. Features include region-of-interest annotation, reconstruction, chromophore unmixing, deep-learning-based tissue segmentation, and temporal and spectral analysis modules. OPTARI offers a combination of features for clinical study analysis that none of the compared tools provide. The automated tissue segmentation produced results comparable to expert labeling for muscle tissue, and batch processing reduced analysis time from minutes to under one second per scan. The tool is used in ongoing clinical OPUS studies. OPTARI connects engineering frameworks with medical research and provides an open-source basis for reproducible and standardized analysis in clinical OptA imaging studies.
Oct 22, 2026

The Value of System 1 Models in Medicine

Many clinical decisions are a choice among a few fixed options, taken quickly and in large numbers. Such decisions also need an uncertainty estimate to know when to defer to a clinician. Generative language models answer slowly in free text and state their confidence poorly. In contrast, System 1 models answer zero-shot with a probability for each answer. However, their performance across clinical tasks is unknown. Four System 1 models, two proprietary and two with open weights, were evaluated zero-shot on 86 public tasks from 14 clinical domains in 20 languages, spanning examination questions and medical imaging. Tasks were assigned to six capability tiers before analysis. The primary outcome was the area under the receiver-operating-characteristic curve (AUROC). Secondary outcomes included calibration, language effects and comparisons with published results and generative models. Across 179,111 cases, the mean AUROC of the best model per task was 0.94 for medical knowledge, 0.93 for text understanding, 0.83 for clinical judgement from text, 0.89 for image perception, 0.84 for 2D image findings and 0.69 for volumetric findings. Imaging tasks scored 0.09 (95% CI, 0.06 to 0.11) lower than text tasks (P < 0.001). The best model matched or exceeded the best published result on 13 of 26 text tasks and on 8 of 35 imaging tasks. Triage nurses outperformed the best model (accuracy, 0.85 vs. 0.56). One decision took 0.34 to 0.44 seconds and 24 to 81 joules on open-weight models run locally. Accuracy on identical questions fell by up to 39 percentage points from English to Yoruba. Generative models of the same family were 11 to 76 times slower, with 3 to 5 percentage points higher balanced accuracy. System 1 models decide quickly, cheaply and with usable confidence when the answer is stated in a text or recalled from medical knowledge, but not when it must be judged or read from an image.
Oct 21, 2026

Two for One: Privileged Learning for Resource-Constrained Screening

Machine learning systems are often developed under data-rich conditions but deployed when only incomplete information is available. A natural example arises in colour fundus photography (CFP), where both eyes are commonly imaged because they can provide complementary information for ocular disease assessment. Prior approaches that exploit bilateral information either fuse both eyes in a single model or use knowledge distillation to compress a dual-eye teacher into a lightweight student that still requires both images at test time. However, bilateral acquisition may not always be feasible. We thus propose a privileged learning framework in which a teacher model has access to bilateral CFP images during training and learns representations informed by information from both eyes, while the student is trained to transfer this knowledge from the bilateral teacher using only a single-eye image as input. At inference, the student operates exclusively on this single-eye input and does not require the contralateral image. Experimental results show consistent improvements in area under the receiver operating characteristic curve over conventional monocular training, with the largest gains observed for left-eye images for the detection of age-related macular degeneration (0.931 vs. 0.966) and diabetic retinopathy (0.789 vs. 0.836).%, while requiring only a single-eye image at inference. By enabling models to benefit from bilateral information during training while requiring only a single-eye image at inference, our approach may support more reliable deployment in resource-constrained clinical and screening environments.
Oct 21, 2026

From Canon to Nikon: Camera Domain Shift and Potential Equity Implications in Retinal AI

Diabetic retinopathy is a leading cause of preventable blindness, yet access to regular screening remains limited worldwide. AI offers a scalable approach, but its performance can be sensitive to image quality and acquisition conditions. In a Brazilian cohort of 8,524 patients, retinal images acquired with the Nikon NF5050 were more than twice as likely to have inadequate image quality compared with those acquired with the Canon CR-2, despite refurbished Canon systems costing more than twice as much as Nikon systems. This potential cost-quality trade-off highlights an important consideration for the reliability of AI systems deployed across heterogeneous imaging systems. Using a frozen pretrained retinal foundation model, we systematically evaluate whether camera-associated performance differences persist across 12 retinal disease classification tasks, four input resolutions, and five downstream classifiers. We find that camera-associated performance differences vary substantially across tasks, and that neither higher input resolution nor classifier choice consistently mitigates cross-camera variation. These findings identify camera platform as a potentially overlooked source of performance variability, with implications for the safety and reliability of retinal AI when deployed across heterogeneous imaging systems. We release all source code, evaluation scripts, and per-task results on GitHub to support reproducible benchmarking.
Oct 20, 2026

DuoPIC: Uncertainty-Aware Voxel-Level Digital Twins of Cerebral Perfusion with Transport PDEs in Acute Stroke

Digital twins of cerebral perfusion relate tracer dynamics to cerebral blood flow (CBF), blood volume (CBV) and mean transit time (MTT), but must be data-driven, physically consistent and uncertainty-aware. We introduce DuoPIC, a digital twin that needs no reference maps, in which a Windkessel flow model drives a tracer transport partial differential equation (PDE). A population-level neural closure divides each region into fast and slow pathways with tracer mass and MTT conserved by construction, and a tempered posterior quantifies the uncertainty of every estimate. Its voxel maps set the state of the art for lesion detection on ISLES'17 MR (AUC 0.745) and ISLES'24 CT (0.743). On held-out patients, the neural PDE reduced the compartment's reconstruction error by 19% and lifted 95% MTT interval coverage from 88% to 94%. Code: https://github.com/detinatran/duopic
Oct 20, 2026

A Multi-Agent Framework for Automated Machine Learning with Evidence-based Planning for Medical Image Analysis

Deep learning has become the default tool for medical image analysis, driven by the open-source movement. However, achieving the best performance on a given task remains difficult, with architecture selection and hyperparameter optimization (HPO) still considered "more of an art than a science". Automated machine learning (AutoML) automates parts of model training, yet helps little in medical imaging: general-purpose approaches lack medical architectures and preprocessing, and medical-focused ones demand engineering expertise and remain confined to narrow tasks such as 3D segmentation. We introduce MedicalMLOmni, a multi-agent AutoML framework that automates model training for diverse single- and multi-modal medical tasks through a natural language interface. Specifically, MedicalMLOmni performs evidence-based planning by retrieving published results to guide model selection and configuration, and employs parallel HPO where multiple agents explore adjustment strategies concurrently. MedicalMLOmni achieves the best score on 19 of 20 benchmarks; PMC-VQA is the exception, where a vision-language model with broader pre-training leads. Backbone-matched controls attribute the gains to the framework, not the backbone. By removing coding requirements, MedicalMLOmni lowers the barrier for domain experts with rich datasets but limited engineering expertise.
Oct 19, 2026

XGLASS: X-Ray, Gaze and Language Assisted Slice Sequencing for CT Volumes

Vision-language models (VLMs) achieve remarkable performance on 2D medical images, such as X-rays or pathology images, while 3D VLMs that directly process CT/MRI volumes lag behind due to limited 3D training data and the predominance of 2D foundation architectures. To leverage powerful 2D VLMs for analyzing medical volumes, which contain hundreds of slices, it is therefore necessary to focus only on those few slices, which are diagnostically relevant. To enable such a prioritization, we propose XGLASS, a method that transfers medical expertise from 2D X-rays to 3D CT for task-aware slice prioritization. XGLASS uses disease-conditioned importance maps derived from language-aligned radiologist gaze on X-rays and spatially aligns them with coronal CT projections, bringing gaze priors and CT content into a shared 2D representation. The fused gaze-prior/CT projection is then used to predict a disease-specific ordering of axial slices, prioritizing diagnostically relevant slices. XGLASS significantly improves budgeted retrieval over 2D and 3D-native baselines, achieving a top-10 hit rate up to 75.9% for the best-performing disease. In practice, this means that after inspecting only the top 10 slices (4.17% of a 240-slice scan), XGLASS identifies a diagnostically relevant slice in more than 75% of cases. For downstream slice-wise finding classification, an XGLASS-selected 30-slice candidate set reduces model calls by 87.5% while improving F1 by 11.3 percentage points over all-slice inference, 4.2 points over the strongest 2D selector, and remains competitive with 3D-native selectors while requiring up to 130.8 times less time and 10.9 times less GPU memory per finding, supporting its use as a lightweight preselector for volumetric CT analysis. The code and pretrained checkpoint are available on https://github.com/michaelgloeckler/XGLASS.
Oct 15, 2026

Spurious Attention Heads in Vision Transformers: Mechanistic Identification of Annotation-Driven Bias in Medical Chest X-Ray Analysis

Vision Transformers deployed in medical imaging exploit spurious correlations, yet the internal mechanisms remain unidentified and impossible to fix. We apply mechanistic interpretability to locate one such mechanism in ViT-Base 16: a single attention head (L8H10) that functions as a dedicated annotation-reader on chest X-rays, attending to corner text rather than lung pathology in 62% of 200 COVID-19 images. Causal validation via inpainting confirms annotation content rather than luminance or edge density drives the behaviour, increasing lung attention 1.52 times upon removal in affected images (p<0.0001p{<}0.0001). The finding replicates across NIH ChestX-ray14 (79%), DeiT-Base (88%), and ViT-Large 16 (80%), establishing spurious annotation-attending heads as a consistent property of ViT-family architectures across two datasets and three architectures. Full fine-tuning does not suppress the behaviour and may entrench it: 98.4% of spurious images remain spurious and the spurious rate rises marginally, as the shortcut is reinforced rather than removed. Critically, single-head ablation produces p=0.340p{=}0.340 and multi-head ablation of all three identified carriers produces p=0.532p{=}0.532, meaning a causally verified annotation-reading circuit is completely invisible to output-level evaluation. Mechanistic auditing finds what behavioural testing cannot.