BackgroundMyocarditis is a life-threatening disease with significant hemodynamic risks during the acute phase. Although histopathological examination of myocardial biopsy specimens remains the gold standard for diagnosis, there is no established method for objectively quantifying cardiomyocyte damage. We aimed to develop an AI model to evaluate clinical myocarditis severity using comprehensive pathology data. MethodsWe retrospectively analyzed 314 patients (1076 samples) who underwent myocardial biopsy from 2002 to 2021 at the National Cerebrovascular Center. Among these patients, 158 were diagnosed with myocarditis based on the Dallas criteria. A Multiple Instance Learning (MIL) model served as a pre-trained classifier to detect myocarditis across whole-slide images. We then constructed two clinical severity-prediction models: (1) a logistic regression model (Model 1) using the density of inflammatory cells per unit area, and (2) a Transformer-based model (Model 2), which processed the top-ranked patches identified by the MIL model to predict clinical severe outcomes. ResultsModel 1 achieved an AUROC of 0.809, indicating a robust association between inflammatory cell density and severe myocarditis. In contrast, Model 2, the Transformer-based approach, yielded an AUROC of 0.993 and demonstrated higher accuracy and precision for severity prediction. Attention score visualizations showed that Model 2 captured both inflammatory cell infiltration and additional morphological features. These findings suggest that combining MIL with Transformer architectures enables more comprehensive identification of key histological markers associated with clinical severe disease. ConclusionsOur results highlight that a Transformer-based AI model analyzing whole-slide pathology images can accurately assess clinical myocarditis severity. Moreover, simply quantifying the extent of inflammatory cell infiltration also correlates strongly with clinical outcomes. These methods offer a promising avenue for improving diagnostic precision, guiding treatment decisions, and ultimately enhancing patient management. Future prospective studies are warranted to validate these models in broader clinical settings and facilitate their integration into routine pathological workflows. What is new?- This is the first study to apply an AI model for the diagnosis and severity assessment of myocarditis. - New evidence shows that inflammatory cell infiltration is related to the severity of myocarditis. - Using information from the entire tissue, not just inflammatory cells, allows for a more accurate assessment of myocarditis severity. What are the clinical implications?- The use of the AI model allows for an unprecedented histological evaluation of myocarditis severity, which can enhance early diagnosis and intervention strategies. - Rapid and precise assessments of myocarditis severity by the AI model can support clinicians in making timely and appropriate treatment decisions, potentially improving patient outcomes. - The incorporation of this AI model into clinical practice may streamline diagnostic workflows and optimize the allocation of medical resources, enhancing overall patient care.

Clinical decision-making relies heavily on understanding relative positions of anatomical structures and anomalies. Therefore, for Vision-Language Models (VLMs) to be applicable in clinical practice, the ability to accurately determine relative positions on medical images is a fundamental prerequisite. Despite its importance, this capability remains highly underexplored. To address this gap, we evaluate the ability of state-of-the-art VLMs, GPT-4o, Llama3.2, Pixtral, and JanusPro, and find that all models fail at this fundamental task. Inspired by successful approaches in computer vision, we investigate whether visual prompts, such as alphanumeric or colored markers placed on anatomical structures, can enhance performance. While these markers provide moderate improvements, results remain significantly lower on medical images compared to observations made on natural images. Our evaluations suggest that, in medical imaging, VLMs rely more on prior anatomical knowledge than on actual image content for answering relative position questions, often leading to incorrect conclusions. To facilitate further research in this area, we introduce the MIRP , Medical Imaging Relative Positioning, benchmark dataset, designed to systematically evaluate the capability to identify relative positions in medical images.

Classification models that provide human-interpretable explanations enhance clinicians' trust and usability in medical image diagnosis. One research focus is the integration and prediction of pathology-related visual attributes used by radiologists alongside the diagnosis, aligning AI decision-making with clinical reasoning. Radiologists use attributes like shape and texture as established diagnostic criteria and mirroring these in AI decision-making both enhances transparency and enables explicit validation of model outputs. However, the adoption of such models is limited by the scarcity of large-scale medical image datasets annotated with these attributes. To address this challenge, we propose synthesizing attribute-annotated data using a generative model. We enhance the Diffusion Model with attribute conditioning and train it using only 20 attribute-labeled lung nodule samples from the LIDC-IDRI dataset. Incorporating its generated images into the training of an explainable model boosts performance, increasing attribute prediction accuracy by 13.4% and target prediction accuracy by 1.8% compared to training with only the small real attribute-annotated dataset. This work highlights the potential of synthetic data to overcome dataset limitations, enhancing the applicability of explainable models in medical image analysis.

Accurate segmentation of small lesions in Breast Dynamic Contrast-Enhanced MRI (DCE-MRI) is critical for early cancer detection, especially in high-risk patients. While recent deep learning methods have advanced lesion segmentation, they primarily target large lesions and neglect valuable longitudinal and clinical information routinely used by radiologists. In real-world screening, detecting subtle or emerging lesions requires radiologists to compare across timepoints and consider previous radiology assessments, such as the BI-RADS score. We propose LesiOnTime, a novel 3D segmentation approach that mimics clinical diagnostic workflows by jointly leveraging longitudinal imaging and BIRADS scores. The key components are: (1) a Temporal Prior Attention (TPA) block that dynamically integrates information from previous and current scans; and (2) a BI-RADS Consistency Regularization (BCR) loss that enforces latent space alignment for scans with similar radiological assessments, thus embedding domain knowledge into the training process. Evaluated on a curated in-house longitudinal dataset of high-risk patients with DCE-MRI, our approach outperforms state-of-the-art single-timepoint and longitudinal baselines by 5% in terms of Dice. Ablation studies demonstrate that both TPA and BCR contribute complementary performance gains. These results highlight the importance of incorporating temporal and clinical context for reliable early lesion segmentation in real-world breast cancer screening. Our code is publicly available at https://github.com/cirmuw/LesiOnTime

Intracranial aneurysms (IAs) are abnormal dilations of cerebral blood vessels that, if ruptured, can lead to life-threatening consequences. However, their small size and soft contrast in radiological scans often make it difficult to perform accurate and efficient detection and morphological analyses, which are critical in the clinical care of the disorder. Furthermore, the lack of large public datasets with voxel-wise expert annotations pose challenges for developing deep learning algorithms to address the issues. Therefore, we proposed a novel weakly supervised 3D multi-task UNet that integrates vesselness priors to jointly perform aneurysm detection and segmentation in time-of-flight MR angiography (TOF-MRA). Specifically, to robustly guide IA detection and segmentation, we employ the popular Frangi's vesselness filter to derive soft cerebrovascular priors for both network input and an attention block to conduct segmentation from the decoder and detection from an auxiliary branch. We train our model on the Lausanne dataset with coarse ground truth segmentation, and evaluate it on the test set with refined labels from the same database. To further assess our model's generalizability, we also validate it externally on the ADAM dataset. Our results demonstrate the superior performance of the proposed technique over the SOTA techniques for aneurysm segmentation (Dice = 0.614, 95%HD =1.38mm) and detection (false positive rate = 1.47, sensitivity = 92.9%).

In clinical practice, medical image analysis often requires efficient execution on resource-constrained mobile devices. However, existing mobile models-primarily optimized for natural images-tend to perform poorly on medical tasks due to the significant information density gap between natural and medical domains. Combining computational efficiency with medical imaging-specific architectural advantages remains a challenge when developing lightweight, universal, and high-performing networks. To address this, we propose a mobile model called Mobile U-shaped Vision Transformer (Mobile U-ViT) tailored for medical image segmentation. Specifically, we employ the newly purposed ConvUtr as a hierarchical patch embedding, featuring a parameter-efficient large-kernel CNN with inverted bottleneck fusion. This design exhibits transformer-like representation learning capacity while being lighter and faster. To enable efficient local-global information exchange, we introduce a novel Large-kernel Local-Global-Local (LGL) block that effectively balances the low information density and high-level semantic discrepancy of medical images. Finally, we incorporate a shallow and lightweight transformer bottleneck for long-range modeling and employ a cascaded decoder with downsample skip connections for dense prediction. Despite its reduced computational demands, our medical-optimized architecture achieves state-of-the-art performance across eight public 2D and 3D datasets covering diverse imaging modalities, including zero-shot testing on four unseen datasets. These results establish it as an efficient yet powerful and generalization solution for mobile medical image analysis. Code is available at https://github.com/FengheTan9/Mobile-U-ViT.

Accurate segmentation of small lesions in Breast Dynamic Contrast-Enhanced MRI (DCE-MRI) is critical for early cancer detection, especially in high-risk patients. While recent deep learning methods have advanced lesion segmentation, they primarily target large lesions and neglect valuable longitudinal and clinical information routinely used by radiologists. In real-world screening, detecting subtle or emerging lesions requires radiologists to compare across timepoints and consider previous radiology assessments, such as the BI-RADS score. We propose LesiOnTime, a novel 3D segmentation approach that mimics clinical diagnostic workflows by jointly leveraging longitudinal imaging and BIRADS scores. The key components are: (1) a Temporal Prior Attention (TPA) block that dynamically integrates information from previous and current scans; and (2) a BI-RADS Consistency Regularization (BCR) loss that enforces latent space alignment for scans with similar radiological assessments, thus embedding domain knowledge into the training process. Evaluated on a curated in-house longitudinal dataset of high-risk patients with DCE-MRI, our approach outperforms state-of-the-art single-timepoint and longitudinal baselines by 5% in terms of Dice. Ablation studies demonstrate that both TPA and BCR contribute complementary performance gains. These results highlight the importance of incorporating temporal and clinical context for reliable early lesion segmentation in real-world breast cancer screening. Our code is publicly available at https://github.com/cirmuw/LesiOnTime

With the increasing number of CT scan examinations, there is a need for automated methods such as organ segmentation, anomaly detection and report generation to assist radiologists in managing their increasing workload. Multi-label classification of 3D CT scans remains a critical yet challenging task due to the complex spatial relationships within volumetric data and the variety of observed anomalies. Existing approaches based on 3D convolutional networks have limited abilities to model long-range dependencies while Vision Transformers suffer from high computational costs and often require extensive pre-training on large-scale datasets from the same domain to achieve competitive performance. In this work, we propose an alternative by introducing a new graph-based approach that models CT scans as structured graphs, leveraging axial slice triplets nodes processed through spectral domain convolution to enhance multi-label anomaly classification performance. Our method exhibits strong cross-dataset generalization, and competitive performance while achieving robustness to z-axis translation. An ablation study evaluates the contribution of each proposed component.

Estimating the elasticity of soft tissue can provide useful information for various diagnostic applications. Ultrasound shear wave elastography offers a non-invasive approach. However, its generalizability and standardization across different systems and processing pipelines remain limited. Considering the influence of image processing on ultrasound based diagnostics, recent literature has discussed the impact of different image processing steps on reliable and reproducible elasticity analysis. In this work, we investigate the need of ultrasound pre-processing steps for deep learning-based ultrasound shear wave elastography. We evaluate the performance of a 3D convolutional neural network in predicting shear wave velocities from spatio-temporal ultrasound images, studying different degrees of pre-processing on the input images, ranging from fully beamformed and filtered ultrasound images to raw radiofrequency data. We compare the predictions from our deep learning approach to a conventional time-of-flight method across four gelatin phantoms with different elasticity levels. Our results demonstrate statistically significant differences in the predicted shear wave velocity among all elasticity groups, regardless of the degree of pre-processing. Although pre-processing slightly improves performance metrics, our results show that the deep learning approach can reliably differentiate between elasticity groups using raw, unprocessed radiofrequency data. These results show that deep learning-based approaches could reduce the need for and the bias of traditional ultrasound pre-processing steps in ultrasound shear wave elastography, enabling faster and more reliable clinical elasticity assessments.

PurposeTo develop and validate an artificial intelligence (AI)-based model that automatically measures choroidal mass dimensions on B{square}scan ophthalmic ultrasound still images and cine loops. DesignRetrospective diagnostic accuracy study with internal and external validation. ParticipantsThe dataset included 1,822 still images and 283 cine loops of choroidal masses for model development and testing. An additional 182 still images were used for external validation, and 302 control images with other diagnoses were included to assess specificity MethodsA deep convolutional neural network (CNN) based on the U-Net architecture was developed to automatically measure the apical height and basal diameter of choroidal masses on B-scan ultrasound. All still images were manually annotated by expert graders and reviewed by a senior ocular oncologist. Cine loops were analyzed frame by frame and the frame with the largest detected mass dimensions was selected for evaluation. Outcome MeasuresThe primary outcome was the models measurement accuracy, defined by the mean absolute error (MAE) in millimeters, compared to expert manual annotations, for both apical height and basal diameter. Secondary metrics included the Dice coefficient, coefficient of determination (R2), and mean pixel distance between predicted and reference measurements. ResultsOn the internal test set of still images, the model successfully detected the tumor in 99.7% of cases. The mean absolute error (MAE) was 0.38 {+/-} 0.55 mm for apical height (95.1% of measurements <1 mm of the expert annotation) and was 0.99 {+/-} 1.15 mm for basal diameter (64.4% of measurements <1 mm). Linear agreement between predicted and reference measurements was strong, with R2 values of 0.74 for apical height and 0.89 for basal diameter. When applied to the control set of 302 control images, the model demonstrated a moderate false positive rate. On the external validation set, the model maintained comparable accuracy. Among the cine loops, the model detected tumors in 89.4% of cases with comparable accuracy. ConclusionDeep learning can deliver fast, reproducible, millimeter{square}level measurements of choroidal mass dimensions with robust performance across different mass types and imaging sources. These findings support the potential clinical utility of AI-assisted measurement tools in ocular oncology workflows.