Reliable uncertainty quantification is crucial for trustworthy decision-making and the deployment of AI models in medical imaging. While prior work has explored the ability of neural networks to quantify predictive, epistemic, and aleatoric uncertainties using an information-theoretical approach in synthetic or well defined data settings like natural image classification, its applicability to real life medical diagnosis tasks remains underexplored. In this study, we provide an extensive uncertainty quantification benchmark for multi-label chest X-ray classification using the MIMIC-CXR-JPG dataset. We evaluate 13 uncertainty quantification methods for convolutional (ResNet) and transformer-based (Vision Transformer) architectures across a wide range of tasks. Additionally, we extend Evidential Deep Learning, HetClass NNs, and Deep Deterministic Uncertainty to the multi-label setting. Our analysis provides insights into uncertainty estimation effectiveness and the ability to disentangle epistemic and aleatoric uncertainties, revealing method- and architecture-specific strengths and limitations.

Computer-aided medical image analysis is crucial for disease diagnosis and treatment planning, yet limited annotated datasets restrict medical-specific model development. While vision-language models (VLMs) like CLIP offer strong generalization capabilities, their direct application to medical imaging analysis is impeded by a significant domain gap. Existing approaches to bridge this gap, including prompt learning and one-way modality interaction techniques, typically focus on introducing domain knowledge to a single modality. Although this may offer performance gains, it often causes modality misalignment, thereby failing to unlock the full potential of VLMs. In this paper, we propose \textbf{NEARL-CLIP} (i\underline{N}teracted qu\underline{E}ry \underline{A}daptation with o\underline{R}thogona\underline{L} Regularization), a novel cross-modality interaction VLM-based framework that contains two contributions: (1) Unified Synergy Embedding Transformer (USEformer), which dynamically generates cross-modality queries to promote interaction between modalities, thus fostering the mutual enrichment and enhancement of multi-modal medical domain knowledge; (2) Orthogonal Cross-Attention Adapter (OCA). OCA introduces an orthogonality technique to decouple the new knowledge from USEformer into two distinct components: the truly novel information and the incremental knowledge. By isolating the learning process from the interference of incremental knowledge, OCA enables a more focused acquisition of new information, thereby further facilitating modality interaction and unleashing the capability of VLMs. Notably, NEARL-CLIP achieves these two contributions in a parameter-efficient style, which only introduces \textbf{1.46M} learnable parameters.

Angle estimation is an important step in the Doppler ultrasound clinical workflow to measure blood velocity. It is widely recognized that incorrect angle estimation is a leading cause of error in Doppler-based blood velocity measurements. In this paper, we propose a deep learning-based approach for automated Doppler angle estimation. The approach was developed using 2100 human carotid ultrasound images including image augmentation. Five pre-trained models were used to extract images features, and these features were passed to a custom shallow network for Doppler angle estimation. Independently, measurements were obtained by a human observer reviewing the images for comparison. The mean absolute error (MAE) between the automated and manual angle estimates ranged from 3.9{\deg} to 9.4{\deg} for the models evaluated. Furthermore, the MAE for the best performing model was less than the acceptable clinical Doppler angle error threshold thus avoiding misclassification of normal velocity values as a stenosis. The results demonstrate potential for applying a deep-learning based technique for automated ultrasound Doppler angle estimation. Such a technique could potentially be implemented within the imaging software on commercial ultrasound scanners.

In recent years, transformer-based methods have achieved remarkable progress in medical image segmentation due to their superior ability to capture long-range dependencies. However, these methods typically suffer from two major limitations. First, their computational complexity scales quadratically with the input sequences. Second, the feed-forward network (FFN) modules in vanilla Transformers typically rely on fully connected layers, which limits models' ability to capture local contextual information and multiscale features critical for precise semantic segmentation. To address these issues, we propose an efficient medical image segmentation network, named TCSAFormer. The proposed TCSAFormer adopts two key ideas. First, it incorporates a Compressed Attention (CA) module, which combines token compression and pixel-level sparse attention to dynamically focus on the most relevant key-value pairs for each query. This is achieved by pruning globally irrelevant tokens and merging redundant ones, significantly reducing computational complexity while enhancing the model's ability to capture relationships between tokens. Second, it introduces a Dual-Branch Feed-Forward Network (DBFFN) module as a replacement for the standard FFN to capture local contextual features and multiscale information, thereby strengthening the model's feature representation capability. We conduct extensive experiments on three publicly available medical image segmentation datasets: ISIC-2018, CVC-ClinicDB, and Synapse, to evaluate the segmentation performance of TCSAFormer. Experimental results demonstrate that TCSAFormer achieves superior performance compared to existing state-of-the-art (SOTA) methods, while maintaining lower computational overhead, thus achieving an optimal trade-off between efficiency and accuracy.

The diagnosis of medical diseases faces challenges such as the misdiagnosis of small lesions. Deep learning, particularly multimodal approaches, has shown great potential in the field of medical disease diagnosis. However, the differences in dimensionality between medical imaging and electronic health record data present challenges for effective alignment and fusion. To address these issues, we propose the Multimodal Multiscale Cross-Attention Fusion Network (MMCAF-Net). This model employs a feature pyramid structure combined with an efficient 3D multi-scale convolutional attention module to extract lesion-specific features from 3D medical images. To further enhance multimodal data integration, MMCAF-Net incorporates a multi-scale cross-attention module, which resolves dimensional inconsistencies, enabling more effective feature fusion. We evaluated MMCAF-Net on the Lung-PET-CT-Dx dataset, and the results showed a significant improvement in diagnostic accuracy, surpassing current state-of-the-art methods. The code is available at https://github.com/yjx1234/MMCAF-Net

Interpolating missing data in k-space is essential for accelerating imaging. However, existing methods, including convolutional neural network-based deep learning, primarily exploit local predictability while overlooking the inherent global dependencies in k-space. Recently, Transformers have demonstrated remarkable success in natural language processing and image analysis due to their ability to capture long-range dependencies. This inspires the use of Transformers for k-space interpolation to better exploit its global structure. However, their lack of interpretability raises concerns regarding the reliability of interpolated data. To address this limitation, we propose GPI-WT, a white-box Transformer framework based on Globally Predictable Interpolation (GPI) for k-space. Specifically, we formulate GPI from the perspective of annihilation as a novel k-space structured low-rank (SLR) model. The global annihilation filters in the SLR model are treated as learnable parameters, and the subgradients of the SLR model naturally induce a learnable attention mechanism. By unfolding the subgradient-based optimization algorithm of SLR into a cascaded network, we construct the first white-box Transformer specifically designed for accelerated MRI. Experimental results demonstrate that the proposed method significantly outperforms state-of-the-art approaches in k-space interpolation accuracy while providing superior interpretability.

Automated segmentation of multiple sclerosis (MS) lesions using multicontrast magnetic resonance (MR) images improves efficiency and reproducibility compared to manual delineation, with deep learning (DL) methods achieving state-of-the-art performance. However, these DL-based methods have yet to simultaneously optimize in-domain accuracy and out-of-domain generalization when trained on a single source with limited data, or their performance has been unsatisfactory. To fill this gap, we propose a method called UNISELF, which achieves high accuracy within a single training domain while demonstrating strong generalizability across multiple out-of-domain test datasets. UNISELF employs a novel test-time self-ensembled lesion fusion to improve segmentation accuracy, and leverages test-time instance normalization (TTIN) of latent features to address domain shifts and missing input contrasts. Trained on the ISBI 2015 longitudinal MS segmentation challenge training dataset, UNISELF ranks among the best-performing methods on the challenge test dataset. Additionally, UNISELF outperforms all benchmark methods trained on the same ISBI training data across diverse out-of-domain test datasets with domain shifts and missing contrasts, including the public MICCAI 2016 and UMCL datasets, as well as a private multisite dataset. These test datasets exhibit domain shifts and/or missing contrasts caused by variations in acquisition protocols, scanner types, and imaging artifacts arising from imperfect acquisition. Our code is available at https://github.com/uponacceptance.

Accurate prediction of post-stroke language outcomes using machine learning offers the potential to enhance clinical treatment and rehabilitation for aphasic patients. This study of 758 English speaking stroke patients from the PLORAS project explores the impact of sample size on the performance of logistic regression and a deep learning (ResNet-18) model in predicting language outcomes from neuroimaging and impairment-relevant tabular data. We assessed the performance of both models on two key language tasks from the Comprehensive Aphasia Test: Spoken Picture Description and Naming, using a learning curve approach. Contrary to expectations, the simpler logistic regression model performed comparably or better than the deep learning model (with overlapping confidence intervals), with both models showing an accuracy plateau around 80% for sample sizes larger than 300 patients. Principal Component Analysis revealed that the dimensionality of the neuroimaging data could be reduced to as few as 20 (or even 2) dominant components without significant loss in accuracy, suggesting that classification may be driven by simple patterns such as lesion size. The study highlights both the potential limitations of current dataset size in achieving further accuracy gains and the need for larger datasets to capture more complex patterns, as some of our results indicate that we might not have reached an absolute classification performance ceiling. Overall, these findings provide insights into the practical use of machine learning for predicting aphasia outcomes and the potential benefits of much larger datasets in enhancing model performance.

This study aimed to develop a simple, interpretable model using routinely available data for predicting COVID-19 mortality at admission, addressing limitations of complex models, and to provide a statistically robust framework for controlled clinical use, managing model uncertainty for responsible healthcare application. Data from Belgium's first COVID-19 wave (UZ Brussel, n = 252) were used for model development. External validation utilized data from unvaccinated patients during the late second and early third waves (AZ Delta, n = 175). Various machine learning methods were trained and compared for diagnostic performance after data preprocessing and feature selection. The final model, the M3-score, incorporated three features: age, white blood cell (WBC) count, and AI-derived total lung involvement (TOTAL_AI) quantified from CT scans using Icolung software. The M3-score demonstrated strong classification performance in the training cohort (AUC 0.903) and clinically useful performance in the external validation dataset (AUC 0.826), indicating generalizability potential. To enhance clinical utility and interpretability, predicted probabilities were categorized into actionable likelihood ratio (LR) intervals: highly unlikely (LR 0.0), unlikely (LR 0.13), gray zone (LR 0.85), more likely (LR 2.14), and likely (LR 8.19) based on the training cohort. External validation suggested temporal and geographical robustness, though some variability in AUC and LR performance was observed, as anticipated in real-world settings. The parsimonious M3-score, integrating AI-based CT quantification with clinical and laboratory data, offers an interpretable tool for predicting in-hospital COVID-19 mortality, showing robust training performance. Observed performance variations in external validation underscore the need for careful interpretation and further extensive validation across international cohorts to confirm wider applicability and robustness before widespread clinical adoption.

To develop and validate a hybrid radiomics model to predict the overall survival in pancreatic cancer patients and identify risk factors that affect patient prognosis. We conducted a retrospective analysis of 272 pancreatic cancer patients diagnosed at the First Affiliated Hospital of Soochow University from January 2013 to December 2023, and divided them into a training set and a test set at a ratio of 7:3. Pre-treatment contrast-enhanced computed tomography (CT), magnetic resonance imaging (MRI) images, and clinical features were collected. Dimensionality reduction was performed on the radiomics features using principal component analysis (PCA), and important features with non-zero coefficients were selected using the least absolute shrinkage and selection operator (LASSO) with 10-fold cross-validation. In the training set, we built clinical prediction models using both random survival forests (RSF) and traditional Cox regression analysis. These models included a radiomics model based on contrast-enhanced CT, a radiomics model based on MRI, a clinical model, 3 bimodal models combining two types of features, and a multimodal model combining radiomics features with clinical features. Model performance evaluation in the test set was based on two dimensions: discrimination and calibration. In addition, risk stratification was performed in the test set based on predicted risk scores to evaluate the model's prognostic utility. The RSF-based hybrid model performed best with a C-index of 0.807 and a Brier score of 0.101, outperforming the COX hybrid model (C-index of 0.726 and a Brier score of 0.145) and other unimodal and bimodal models. The SurvSHAP(t) plot highlighted CA125 as the most important variable. In the test set, patients were stratified into high- and low-risk groups based on the predicted risk scores, and Kaplan-Meier analysis demonstrated a significant survival difference between the two groups (p < 0.0001). A multi-modal model using radiomics based on clinical tabular data and contrast-enhanced CT and MRI was developed by RSF, presenting strengths in predicting prognosis in pancreatic cancer patients.