Every story in this issue with a short summary, so you can decide what to read before leaving the site. Summaries are written by AI (Claude) only from each publisher's own description and checked by Jev for anything not in that description. They are not full translations. Follow "Read the original" for the full story.
The FFR-IQ software recently demonstrated 92.2 percent sensitivity for non-invasive estimation of fractional flow reserve (FFRC) based on coronary CT angiography (CCTA) scans.
Read the original →The British Institute of Radiology has launched the BIR AI Academy, a growing program of live, interactive AI education for radiologists and imaging professionals. DeepHealth and Harrison.ai are its first industry collaborators.
Read the original →AI can improve human experts' ability to diagnose and treat diseases like breast cancer. Well-publicized errors across the broader AI field are cause for concern.
Read the original →A critical narrative review finds that AI improves detection in mammography and colonoscopy but does not improve complex diagnostic reasoning. It also notes that experimental evidence raises concerns.
Read the original →New trials are testing AI at the first point of care. Their results show why answering a medical question, directing a patient and improving treatment are different achievements.
Read the original →M42 announced the launch of Ossi, its next-generation multi-agent AI Clinical Platform, during the Abu Dhabi Future Health Summit.
Read the original →Three global healthcare leaders met in Seoul on Monday on the sidelines of the International Hospital Federation (IHF) World Hospital Congress. Among them was Deborah J. Bowen, president and CEO of the American College of Healthcare Executives (ACHE).
Read the original →SickKids and Microsoft are teaming up to provide individualized pediatric care through enhanced cloud solutions and artificial intelligence.
Read the original →Alcidion has launched an AI-enabled hospital admin system designed to speed up referrals and bookings, manage waiting lists and connect administrative teams with clinical care.
Read the original →The alert covers five core challenges of AI-generated evidence in life sciences: hallucinations, skewed predictions from unrepresentative training data, opaque AI reasoning, input sensitivity that undermines reproducibility, and adversarial manipulation of training data. A regulatory submission relying on AI-generated data without independent human validation may not meet existing FDA data integrity standards. AI-generated expert analysis offered in litigation may be open to challenge on reproducibility grounds under Federal Rule of Evidence 702.
Read the original →AI drug discovery is moving from a software story to a physical-capacity story, and a large share of that capacity sits in China. Before AI transforms drug discovery, it may strain the infrastructure used to test molecules worldwide. China's biotech services sector represents one of the largest pools of that infrastructure.
Read the original →NVIDIA announced collaborations with leading global healthcare and research organizations to advance protein design, including through a newly released open model, RFD4-Proteina. The collaborations also aim to accelerate drug discovery and bring greater precision to cancer care.
Read the original →Following publication of an FDA discussion paper and draft guidance, medical device companies should prepare a cohesive patent and pre-market submission strategy. This involves considering FDA recommendations when developing AI-enabled features and coordinating statements made to the USPTO and the FDA.
Read the original →Rick Abramson, director of the FDA's Digital Health Center of Excellence, said to expect draft guidance in the coming months.
Read the original →Lee Wang-jun urges a nationwide system akin to the 1989 health insurance program. The aim is to ensure equitable access and maintain medical competitiveness.
Read the original →An artificial intelligence model combined electronic health records and electrocardiograms. It was used to select and screen suitable patients for structural heart diseases.
Read the original →The author has been involved in medical AI research and development for 11 years, working from industry, government and academic positions on Japan's medical AI and healthcare digital transformation. Overseas, administrative efforts range from policy to guidelines, and they differ in form as well as in the national background behind them.
Read the original →Marso, based in Minato Ward, Tokyo, announced on October 21 the launch of "MRSO Security," a security support service for medical institutions. It uses AI to run vulnerability diagnostics on institutions' public websites. It provides an individual report covering confirmed findings, expected impact and recommended actions.
Read the original →TRIAD proposes a three-layer medical AI framework that evaluates clinicians and algorithms together rather than the model alone, in contrast to long-standing accuracy-centered evaluation. Published on February 10, 2026, it links trustworthy governance, real clinical value, and integrated, adaptive deployment. The authors, affiliated with Beijing Friendship Hospital and Capital Medical University, argue that clinical intelligence belongs to human-AI teams rather than to algorithms operating in isolation.
Read the original →