EndoSeg-RT targets a practical bottleneck in surgical AI: accurate spinal endoscopy segmentation that can run in real time on limited hardware.

In this prospective cohort, novice-performed AI-guided focused cardiac ultrasound was feasible, fast, and highly accurate for selected structural heart diseases.

A machine learning model using routine clinical and contextual data predicted 7-day hospitalisation or death in young children with pneumonia in Malawi with strong external performance.

Fed-GPL offers a communication-efficient way to adapt vision foundation models for multi-center medical imaging without sharing patient data.

Cardiology AI is shifting from specialized imaging to opportunistic screening, extracting risk from routine tests. The bottleneck? Standardization across diverse populations. This batch shows the field knows what…

A multicenter MRI study proposes quantitative biomarkers for lumbar aging that may help distinguish expected age-related change from accelerated disc degeneration risk.

AI-assisted wrist-hand ultrasound improved acquisition quality for novice musculoskeletal sonologists in an initial multicenter clinical evaluation.

Brief GPT-4o chatbot conversations increased parents’ short-term intentions to vaccinate their children against HPV, but they did not outperform country-matched government public health materials.

New registry-based models can estimate 12-month disability after lumbar spinal stenosis surgery with reasonable accuracy across Scandinavian populations.

Transcript INTRO Welcome to Medical AI Digest Podcast. Today, a paper from Radiology. MAIN The study we’re discussing is titled “Artificial Intelligence Detection Scores in Screening Mammography for…