We explore biases present in publicly available fetal ultrasound (US) imaging datasets, currently at the disposal of researchers to train deep learning (DL) algorithms for prenatal diagnostics. As DL increasingly permeates the field of medical imaging, the urgency to critically evaluate the fairness of benchmark public datasets used to train them grows. Our thorough investigation reveals a multifaceted bias problem, encompassing issues such as lack of demographic representativeness, limited diversity in clinical conditions depicted, and variability in US technology used across datasets. We argue that these biases may significantly influence DL model performance, which may lead to inequities in healthcare outcomes. To address these challenges, we recommend a multilayered approach. This includes promoting practices that ensure data inclusivity, such as diversifying data sources and populations, and refining model strategies to better account for population variances. These steps will enhance the trustworthiness of DL algorithms in fetal US analysis.

This study aimed to develop and evaluate models for classifying the severity of neurological impairment in acute ischemic stroke (AIS) patients using multimodal MRI data. A retrospective cohort of 1227 AIS patients was collected and categorized into mild (NIHSS<5) and moderate-to-severe (NIHSS≥5) stroke groups based on NIHSS scores. Eight baseline models were constructed for performance comparison, including a clinical model, radiomics models using DWI or multiple MRI sequences, and deep learning (DL) models with varying fusion strategies (early fusion, later fusion, full cross-fusion, and DWI-centered cross-fusion). All DL models were based on the Vision Transformer (ViT) framework. Model performance was evaluated using metrics such as AUC and ACC, and robustness was assessed through subgroup analyses and visualization using Grad-CAM. Among the eight models, the DL model using DWI as the primary sequence with cross-fusion of other MRI sequences (Model 8) achieved the best performance. In the test cohort, Model 8 demonstrated an AUC of 0.914, ACC of 0.830, and high specificity (0.818) and sensitivity (0.853). Subgroup analysis shows that model 8 is robust in most subgroups with no significant prediction difference (p > 0.05), and the AUC value consistently exceeds 0.900. A significant predictive difference was observed in the BMI group (p < 0.001). The results of external validation showed that the AUC values of the model 8 in center 2 and center 3 reached 0.910 and 0.912, respectively. Visualization using Grad-CAM emphasized the infarct core as the most critical region contributing to predictions, with consistent feature attention across DWI, T1WI, T2WI, and FLAIR sequences, further validating the interpretability of the model. A ViT-based DL model with cross-modal fusion strategies provides a non-invasive and efficient tool for classifying AIS severity. Its robust performance across subgroups and interpretability make it a promising tool for personalized management and decision-making in clinical practice.

Automatic segmentation of anatomical structures is critical in medical image analysis, aiding diagnostics and treatment planning. Skin segmentation plays a key role in registering and visualising multimodal imaging data. 3D skin segmentation enables applications in personalised medicine, surgical planning, and remote monitoring, offering realistic patient models for treatment simulation, procedural visualisation, and continuous condition tracking. This paper analyses and compares algorithmic and AI-driven skin segmentation approaches, emphasising key factors to consider when selecting a strategy based on data availability and application requirements. We evaluate an iterative region-growing algorithm and the TotalSegmentator, a deep learning-based approach, across different imaging modalities and anatomical regions. Our tests show that AI segmentation excels in automation but struggles with MRI due to its CT-based training, while the graphics-based method performs better for MRIs but introduces more noise. AI-driven segmentation also automates patient bed removal in CT, whereas the graphics-based method requires manual intervention.

High-resolution computed tomography (CT) imaging is essential for medical diagnosis but requires increased radiation exposure, creating a critical trade-off between image quality and patient safety. While deep learning methods have shown promise in CT super-resolution, they face challenges with complex degradations and limited medical training data. Meanwhile, large-scale pre-trained diffusion models, particularly Stable Diffusion, have demonstrated remarkable capabilities in synthesizing fine details across various vision tasks. Motivated by this, we propose a novel framework that adapts Stable Diffusion for CT blind super-resolution. We employ a practical degradation model to synthesize realistic low-quality images and leverage a pre-trained vision-language model to generate corresponding descriptions. Subsequently, we perform super-resolution using Stable Diffusion with a specialized controlling strategy, conditioned on both low-resolution inputs and the generated text descriptions. Extensive experiments show that our method outperforms existing approaches, demonstrating its potential for achieving high-quality CT imaging at reduced radiation doses. Our code will be made publicly available.

Anorexia nervosa (AN), a severe eating disorder marked by extreme weight loss and malnutrition, leads to significant alterations in brain structure. This study used machine learning (ML) to estimate brain age from structural MRI scans and investigated brain-predicted age difference (brain-PAD) as a potential biomarker in AN. Structural MRI scans were collected from female participants aged 10-40 years across two institutions (Boston, USA, and Jena, Germany), including acute AN (acAN; n=113), weight-restored AN (wrAN; n=35), and age-matched healthy controls (HC; n=90). The ML model was trained on 3487 healthy female participants (ages 5-45 years) from ten datasets, using 377 neuroanatomical features extracted from T1-weighted MRI scans. The model achieved strong performance with a mean absolute error (MAE) of 1.93 years and a correlation of r = 0.88 in HCs. In acAN patients, brain age was overestimated by an average of +2.25 years, suggesting advanced brain aging. In contrast, wrAN participants showed significantly lower brain-PAD than acAN (+0.26 years, p=0.0026) and did not differ from HC (p=0.98), suggesting normalization of brain age estimates following weight restoration. A significant group-by-age interaction effect on predicted brain age (p<0.001) indicated that brain age deviations were most pronounced in younger acAN participants. Brain-PAD in acAN was significantly negatively associated with BMI (r = -0.291, p_fdr = 0.005), but not in wrAN or HC groups. Importantly, no significant associations were found between brain-PAD and clinical symptom severity. These findings suggest that acute AN is linked to advanced brain aging during the acute stage, and that may partially normalize following weight recovery.

Purpose: Inversion recovery prepared ultra-short echo time (IR-UTE)-based MRI enables radiation-free visualization of osseous tissue. However, sufficient signal-to-noise ratio (SNR) can only be obtained with long acquisition times. This study proposes a data-driven approach to reconstruct undersampled IR-UTE knee data, thereby accelerating MR-based 3D imaging of bones. Methods: Data were acquired with a 3D radial IR-UTE pulse sequence, implemented using the open-source framework Pulseq. A denoising convolutional neural network (DnCNN) was trained in a supervised fashion using data from eight healthy subjects. Conjugate gradient sensitivity encoding (CG-SENSE) reconstructions of different retrospectively undersampled subsets (corresponding to 2.5-min, 5-min and 10-min acquisition times) were paired with the respective reference dataset reconstruction (30-min acquisition time). The DnCNN was then integrated into a Landweber-based reconstruction algorithm, enabling physics-based iterative reconstruction. Quantitative evaluations of the approach were performed using one prospectively accelerated scan as well as retrospectively undersampled datasets from four additional healthy subjects, by assessing the structural similarity index measure (SSIM), the peak signal-to-noise ratio (PSNR), the normalized root mean squared error (NRMSE), and the perceptual sharpness index (PSI). Results: Both the reconstructions of prospective and retrospective acquisitions showed good agreement with the reference dataset, indicating high image quality, particularly for an acquisition time of 5 min. The proposed method effectively preserves contrast and structural details while suppressing noise, albeit with a slight reduction in sharpness. Conclusion: The proposed method is poised to enable MR-based bone assessment in the knee within clinically feasible scan times.

The adoption of neural network models in medical imaging has been constrained by strict privacy regulations, limited data availability, high acquisition costs, and demographic biases. Deep generative models offer a promising solution by generating synthetic data that bypasses privacy concerns and addresses fairness by producing samples for under-represented groups. However, unlike natural images, medical imaging requires validation not only for fidelity (e.g., Fr\'echet Inception Score) but also for morphological and clinical accuracy. This is particularly true for colour fundus retinal imaging, which requires precise replication of the retinal vascular network, including vessel topology, continuity, and thickness. In this study, we in-vestigated whether a distance-based loss function based on deep activation layers of a large foundational model trained on large corpus of domain data, colour fundus imaging, offers advantages over a perceptual loss and edge-detection based loss functions. Our extensive validation pipeline, based on both domain-free and domain specific tasks, suggests that domain-specific deep features do not improve autoen-coder image generation. Conversely, our findings highlight the effectiveness of con-ventional edge detection filters in improving the sharpness of vascular structures in synthetic samples.

Incomplete multi-modal medical image segmentation faces critical challenges from modality imbalance, including imbalanced modality missing rates and heterogeneous modality contributions. Due to their reliance on idealized assumptions of complete modality availability, existing methods fail to dynamically balance contributions and neglect the structural relationships between modalities, resulting in suboptimal performance in real-world clinical scenarios. To address these limitations, we propose a novel model, named Dynamic Modality-Aware Fusion Network (DMAF-Net). The DMAF-Net adopts three key ideas. First, it introduces a Dynamic Modality-Aware Fusion (DMAF) module to suppress missing-modality interference by combining transformer attention with adaptive masking and weight modality contributions dynamically through attention maps. Second, it designs a synergistic Relation Distillation and Prototype Distillation framework to enforce global-local feature alignment via covariance consistency and masked graph attention, while ensuring semantic consistency through cross-modal class-specific prototype alignment. Third, it presents a Dynamic Training Monitoring (DTM) strategy to stabilize optimization under imbalanced missing rates by tracking distillation gaps in real-time, and to balance convergence speeds across modalities by adaptively reweighting losses and scaling gradients. Extensive experiments on BraTS2020 and MyoPS2020 demonstrate that DMAF-Net outperforms existing methods for incomplete multi-modal medical image segmentation. Extensive experiments on BraTS2020 and MyoPS2020 demonstrate that DMAF-Net outperforms existing methods for incomplete multi-modal medical image segmentation. Our code is available at https://github.com/violet-42/DMAF-Net.

To synthesize current evidence on prognostic factors, tools, and strategies influencing functional outcomes in patients with traumatic brain injury (TBI), with a focus on the acute and postacute phases of care. Key early predictors such as Glasgow Coma Scale (GCS) scores, pupillary reactivity, and computed tomography (CT) imaging findings remain fundamental in guiding clinical decision-making. Prognostic models like IMPACT and CRASH enhance early risk stratification, while outcome measures such as the Glasgow Outcome Scale-Extended (GOS-E) provide structured long-term assessments. Despite their utility, heterogeneity in assessment approaches and treatment protocols continues to limit consistency in outcome predictions. Recent advancements highlight the value of fluid biomarkers like neurofilament light chain (NFL) and glial fibrillary acidic protein (GFAP), which offer promising avenues for improved accuracy. Additionally, artificial intelligence models are emerging as powerful tools to integrate complex datasets and refine individualized outcome forecasting. Neurological prognostication after TBI is evolving through the integration of clinical, radiological, molecular, and computational data. Although standardized models and scales remain foundational, emerging technologies and therapies - such as biomarkers, machine learning, and neurostimulants - represent a shift toward more personalized and actionable strategies to optimize recovery and long-term function.

Differences in CT-based body composition (BC) have been observed by race. We sought to investigate whether indices reporting census block group-level disadvantage, area deprivation index (ADI) and social vulnerability index (SVI), age, sex, and/or clinical factors could explain race-based differences in body composition. The first abdominal CT exams for patients in Durham County at a single institution in 2020 were analyzed using a fully automated and open-source deep learning BC analysis workflow to generate cross-sectional areas for skeletal muscle (SMA), subcutaneous fat (SFA), and visceral fat (VFA). Patient level demographic and clinical data were gathered from the electronic health record. State ADI ranking and SVI values were linked to each patient. Univariable and multivariable models were created to assess the association of demographics, ADI, SVI, and other relevant clinical factors with SMA, SFA, and VFA. 5,311 patients (mean age, 57.4 years; 55.5% female, 46.5% Black; 39.5% White 10.3% Hispanic) were included. At univariable analysis, race, ADI, SVI, sex, BMI, weight, and height were significantly associated with all body compartments (SMA, SFA, and VFA, all p<0.05). At multivariable analyses adjusted for patient characteristics and clinical comorbidities, race remained a significant predictor, whereas ADI did not. SVI was significant in a multivariable model with SMA.