Radiology AI is shifting from image interpretation toward accountable clinical decisions: faster reporting, actionable triage and planning, and personalized risk models—all tempered by the need for robust validation…

Global reported mpox did not follow a single epidemiological pattern between 2022 and 2025. Across four distinct phases, its model-associated correlates shifted from behavioral, sociocultural, and mobility factors…

A high app-store rating is not a proxy for clinical effectiveness. A cross-market analysis of 286 mental health apps found a persistent divide between developer-side evidence and the…

TEMPO-HCC integrates MRI, digital pathology, and clinical variables to estimate individualized mortality risk at 1, 2, 3, and 5 years after hepatocellular carcinoma resection. In external validation, it…

PULSE, a knowledge-enhanced large language model agent, significantly improved clinician-rated responses to real-world atrial fibrillation questions across four general-purpose models. The findings show how curated evidence, consistency checking,…

A multimodal framework combining structured clinical data, brain MRI, and retrieval-augmented language-model reasoning improved diagnostic performance for cognitive impairment, particularly when patient information was incomplete. Crucially, the language…

A three-tier, guideline-grounded large language model generated accurate and safe recommendations for diabetic kidney disease management, with nearly 88% of cases handled locally. Its most notable gains involved…

A nationally deployed, multistage AI workflow increased cancer detection among general radiologists by nearly one-third, bringing their adjusted performance in line with fellowship-trained breast imaging specialists. The gain,…

Longitudinal sleep data collected passively at home distinguished dementia-related patterns from those of typical ageing. Crucially, the signal reflected deviation from normal age-associated sleep—not simply accelerated biological ageing.

A CT-based multi-task deep learning model predicted EGFR mutation status in non-small cell lung cancer across internal and external validation cohorts. Combining its imaging score with clinical and…