Advanced esophageal squamous cell carcinoma (ESCC) has an extremely poor prognosis. Preoperative chemoradiotherapy (CRT) can significantly prolong survival, especially in those who achieve pathological complete response (pCR). However, the pretherapeutic prediction of pCR remains challenging. To predict pCR and survival in ESCC patients undergoing CRT using an artificial intelligence (AI)-based diffusion-weighted magnetic resonance imaging (DWI-MRI) radiomics model. We retrospectively analyzed 70 patients with ESCC who underwent curative surgery following CRT. For each patient, pre-treatment tumors were semi-automatically segmented in three dimensions from DWI-MRI images (<i>b</i> = 0, 1000 second/mm²), and a total of 76 radiomics features were extracted from each segmented tumor. Using these features as explanatory variables and pCR as the objective variable, machine learning models for predicting pCR were developed using AutoGluon, an automated machine learning library, and validated by stratified double cross-validation. pCR was achieved in 15 patients (21.4%). Apparent diffusion coefficient skewness demonstrated the highest predictive performance [area under the curve (AUC) = 0.77]. Gray-level co-occurrence matrix (GLCM) entropy (<i>b</i> = 1000 second/mm²) was an independent prognostic factor for relapse-free survival (RFS) (hazard ratio = 0.32, <i>P</i> = 0.009). In Kaplan-Meier analysis, patients with high GLCM entropy showed significantly better RFS (<i>P</i> < 0.001, log-rank). The best-performing machine learning model achieved an AUC of 0.85. The predicted pCR-positive group showed significantly better RFS than the predicted pCR-negative group (<i>P</i> = 0.007, log-rank). AI-based radiomics analysis of DWI-MRI images in ESCC has the potential to accurately predict the effect of CRT before treatment and contribute to constructing optimal treatment strategies.

Federated Learning (FL) trains models locally at each research center or clinic and aggregates only model updates, making it a natural fit for medical imaging, where strict privacy laws forbid raw data sharing. A major obstacle is scanner-induced domain shift: non-biological variations in hardware or acquisition protocols can cause models to fail on external sites. Most harmonization methods correct this shift by directly comparing data across sites, conflicting with FL's privacy constraints. Domain Generalization (DG) offers a privacy-friendly alternative - learning site-invariant representations without sharing raw data - but standard DG pipelines still assume centralized access to multi-site data, again violating FL's guarantees. This paper meets these difficulties with a straightforward integration of a Domain-Adversarial Neural Network (DANN) within the FL process. After demonstrating that a naive federated DANN fails to converge, we propose a proximal regularization method that stabilizes adversarial training among clients. Experiments on T1-weighted 3-D brain MRIs from the OpenBHB dataset, performing brain-age prediction on participants aged 6-64 y (mean 22+/-6 y; 45 percent male) in training and 6-79 y (mean 19+/-13 y; 55 percent male) in validation, show that training on 15 sites and testing on 19 unseen sites yields superior cross-site generalization over FedAvg and ERM while preserving data privacy.

Structural MRI of the brain is routinely performed on patients referred to memory clinics; however, resulting radiology reports, including volumetric assessments, are conventionally stored as unstructured free text. We sought to use natural language processing (NLP) to extract text relating to intracranial volumetric assessment from brain MRI text reports to enhance routine data availability for research purposes. Electronic records from a large mental healthcare provider serving a geographic catchment of 1.3 million residents in four boroughs of south London, UK. A corpus of 4007 de-identified brain MRI reports from patients referred to memory assessment services. An NLP algorithm was developed, using a span categorisation approach, to extract six binary (presence/absence) categories from the text reports: (i) global volume loss, (ii) hippocampal/medial temporal lobe volume loss and (iii) other lobar/regional volume loss. Distributions of these categories were evaluated. The overall F1 score for the six categories was 0.89 (precision 0.92, recall 0.86), with the following precision/recall for each category: presence of global volume loss 0.95/0.95, absence of global volume loss 0.94/0.77, presence of regional volume loss 0.80/0.58, absence of regional volume loss 0.91/0.93, presence of hippocampal volume loss 0.90/0.88, and absence of hippocampal volume loss 0.94/0.92. These results support the feasibility and accuracy of using NLP techniques to extract volumetric assessments from radiology reports, and the potential for automated generation of novel meta-data from dementia assessments in electronic health records.

Sudden cardiac death (SCD) from ventricular arrhythmias (VAs) remains a leading cause of mortality worldwide. Traditional risk stratification, primarily based on left ventricular ejection fraction (LVEF) and other coarse metrics, often fails to identify a large subset of patients at risk and frequently leads to unnecessary device implantations. Advances in artificial intelligence (AI) offer new strategies to improve both long-term SCD risk prediction and near-term VAs forecasting. In this review, we discuss how AI algorithms applied to the 12-lead electrocardiogram (ECG) can identify subtle risk markers in conditions such as hypertrophic cardiomyopathy (HCM), arrhythmogenic right ventricular cardiomyopathy (ARVC), and coronary artery disease (CAD), often outperforming conventional risk models. We also explore the integration of AI with cardiac imaging, such as scar quantification on cardiac magnetic resonance (CMR) and fibrosis mapping, to enhance the identification of the arrhythmogenic substrate. Furthermore, we investigate the application of data from implantable cardioverter-defibrillators (ICDs) and wearable devices to predict ventricular tachycardia (VT) or ventricular fibrillation (VF) events before they occur, thereby advancing care toward real-time prevention. Amid these innovations, we address the medicolegal and ethical implications of AI-driven automated alerts in arrhythmia care, highlighting when clinicians can trust AI predictions. Future directions include multimodal AI fusion to personalise SCD risk assessment, as well as AI-guided VT ablation planning through imaging-based digital heart models. This review provides a comprehensive overview for general medical readers, focusing on peer-reviewed advances globally in the emerging intersection of AI, VAs, and SCD prevention.

Traditional brain connectivity studies have focused mainly on structural connectivity, often relying on tractography with diffusion MRI (dMRI) to reconstruct white matter pathways. In parallel, studies of functional connectivity have examined correlations in brain activity using fMRI. However, emerging methodologies are advancing our understanding of brain networks. Here we explore advanced connectivity approaches beyond conventional tractography, focusing on dMRI morphometry and the integration of structural and functional connectivity analysis. dMRI morphometry enables quantitative assessment of white matter pathway volumes through statistical comparison with normative populations, while functional connectivity reveals network organization that is not restricted to direct anatomical connections. More recently, approaches that combine diffusion tensor imaging (DTI) with functional correlation tensor (FCT) analysis have been introduced, and these complementary methods provide new perspectives into brain structure-function relationships. Together, such approaches have important implications for neurodevelopmental and neurological disorders as well as brain plasticity. The integration of these methods with artificial intelligence techniques have the potential to support both basic neuroscience research and clinical applications.

Ultra-low-field magnetic resonance imaging (ULF-MRI), operating below 0.2 Tesla, is gaining renewed interest as a re-emerging diagnostic modality in a field dominated by high- and ultra-high-field systems. Recent advances in magnet design, RF coils, pulse sequences, and AI-based reconstruction have significantly enhanced image quality, mitigating traditional limitations such as low signal- and contrast-to-noise ratio and reduced spatial resolution. ULF-MRI offers distinct advantages: reduced susceptibility artifacts, safer imaging in patients with metallic implants, low power consumption, and true portability for point-of-care use. This narrative review synthesizes the physical foundations, technological advances, and emerging clinical applications of ULF-MRI. A focused literature search across PubMed, Scopus, IEEE Xplore, and Google Scholar was conducted up to August 11, 2025, using combined keywords targeting hardware, software, and clinical domains. Inclusion emphasized scientific rigor and thematic relevance. A comparative analysis with other imaging modalities highlights the specific niche ULF-MRI occupies within the broader diagnostic landscape. Future directions and challenges for clinical translation are explored. In a world increasingly polarized between the push for ultra-high-field excellence and the need for accessible imaging, ULF-MRI embodies a modern "David versus Goliath" theme, offering a sustainable, democratizing force capable of expanding MRI access to anyone, anywhere.

To evaluate the image quality and clinical utility of ultra-fast T2-weighted imaging (UF-T2WI), which acquires all slice data in 7 s using a single-shot turbo spin-echo technique combined with dual-type deep learning (DL) reconstruction, incorporating DL-based image denoising and super-resolution processing, by comparing UF-T2WI with conventional T2WI. We analyzed data from 38 patients who underwent both conventional T2WI and UF-T2WI with the dual-type DL-based image reconstruction. Two board-certified radiologists independently performed blinded qualitative assessments of the patients' images obtained with UF-T2WI with DL and conventional T2WI, evaluating the overall image quality, anatomical structure visibility, and levels of noise and artifacts. In cases that included central nervous system diseases, the lesions' delineation was also assessed. The quantitative analysis included measurements of signal-to-noise ratios in white and gray matter and the contrast-to-noise ratio between gray and white matter. Compared to conventional T2WI, UF-T2WI with DL received significantly higher ratings for overall image quality and lower noise and artifact levels (p < 0.001 for both readers). The anatomical visibility was significantly better in UF-T2WI for one reader, with no significant difference for the other reader. The lesion visibility in UF-T2WI was comparable to that in conventional T2WI. Quantitatively, the SNRs and CNRs were all significantly higher in UF-T2WI than conventional T2WI (p < 0.001). The combination of SSTSE with dual-type DL reconstruction allows for the acquisition of clinically acceptable T2WI images in just 7 s. This technique shows strong potential to reduce MRI scan times and improve clinical workflow efficiency.

Accurate classification of Human epidermal growth factor receptor 2 (HER2) expression is crucial for guiding treatment in breast cancer, especially with emerging therapies like trastuzumab deruxtecan (T-DXd) for HER2-low patients. Current gold-standard methods relying on invasive biopsy and immunohistochemistry suffer from sampling bias and interobserver variability, highlighting the need for reliable non-invasive alternatives. We developed an artificial intelligence framework that integrates a pretrained foundation model with a task-specific classifier to predict HER2 expression categories (HER2-zero, HER2-low, HER2-positive) directly from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). The model was trained and validated using multicenter datasets. Model interpretability was assessed through feature visualization using t-SNE and UMAP dimensionality reduction techniques, complemented by SHAP analysis for post-hoc interpretation of critical predictive imaging features. The developed model demonstrated robust performance across datasets, achieving micro-average AUCs of 0.821 (95% CI 0.795–0.846) and 0.835 (95% CI 0.797–0.864), and macro-average AUCs of 0.833 (95% CI 0.818–0.847) and 0.857 (95% CI 0.837–0.872) in external validation. Subgroup analysis demonstrated strong discriminative power in distinguishing HER2 categories, particularly HER2-zero and HER2-low cases. Visualization techniques revealed distinct, biologically plausible clustering patterns corresponding to HER2 expression categories. This study presents a reproducible, non-invasive solution for comprehensive HER2 phenotyping using DCE-MRI, addressing fundamental limitations of biopsy-dependent assessment. Our approach enables accurate identification of HER2-low patients who may benefit from novel therapies like T-DXd. This framework represents a significant advancement in precision oncology, with potential to transform diagnostic workflows and guide targeted therapy selection in breast cancer care. The online version contains supplementary material available at 10.1186/s13058-025-02118-2.

Recent advances in vision-language models (VLMs) have achieved remarkable performance on standard medical benchmarks, yet their true clinical reasoning ability remains unclear. Existing datasets predominantly emphasize classification accuracy, creating an evaluation illusion in which models appear proficient while still failing at high-stakes diagnostic reasoning. We introduce Neural-MedBench, a compact yet reasoning-intensive benchmark specifically designed to probe the limits of multimodal clinical reasoning in neurology. Neural-MedBench integrates multi-sequence MRI scans, structured electronic health records, and clinical notes, and encompasses three core task families: differential diagnosis, lesion recognition, and rationale generation. To ensure reliable evaluation, we develop a hybrid scoring pipeline that combines LLM-based graders, clinician validation, and semantic similarity metrics. Through systematic evaluation of state-of-the-art VLMs, including GPT-4o, Claude-4, and MedGemma, we observe a sharp performance drop compared to conventional datasets. Error analysis shows that reasoning failures, rather than perceptual errors, dominate model shortcomings. Our findings highlight the necessity of a Two-Axis Evaluation Framework: breadth-oriented large datasets for statistical generalization, and depth-oriented, compact benchmarks such as Neural-MedBench for reasoning fidelity. We release Neural-MedBench at https://neuromedbench.github.io/ as an open and extensible diagnostic testbed, which guides the expansion of future benchmarks and enables rigorous yet cost-effective assessment of clinically trustworthy AI.

Semantic segmentation is a fundamental task in medical image analysis, aiding medical decision-making by helping radiologists distinguish objects in an image. Research in this field has been driven by deep learning applications, which have the potential to scale these systems even in the presence of noise and artifacts. However, these systems are not yet perfected. We argue that performance can be improved by incorporating common medical knowledge into the segmentation model's loss function. To this end, we introduce Logic Tensor Networks (LTNs) to encode medical background knowledge using first-order logic (FOL) rules. The encoded rules span from constraints on the shape of the produced segmentation, to relationships between different segmented areas. We apply LTNs in an end-to-end framework with a SwinUNETR for semantic segmentation. We evaluate our method on the task of segmenting the hippocampus in brain MRI scans. Our experiments show that LTNs improve the baseline segmentation performance, especially when training data is scarce. Despite being in its preliminary stages, we argue that neurosymbolic methods are general enough to be adapted and applied to other medical semantic segmentation tasks.