AI in Medicine News — 2026-10-25 (Sun) Daily News

Overview

Diagnostics and clinical decision support moved most: Microsoft's diagnostic AI drew wide coverage, with a reported 83% accuracy on complex cases against 36% for physicians, in a simulated setting (Business+IT, Around Prague). Regulation saw FDA clearances of an aortic CT AI tool and AI-enabled stethoscopes (Ortho Spine News, Cardiovascular Business). Drug discovery saw large financings, including $215 million for Warp Sciences and a $237M Series E for Pathos AI (Kingy AI, Wowtale).

Diagnostics and imaging

- AI reads ultrasound to predict breast cancer spread before surgery. The headline gives no accuracy figure or evidence stage (Bioengineer.org). - AI fails to compensate for night-shift fatigue among residents (AuntMinnie). - AI technology detects high-risk coronary inflammation in many asymptomatic patients. The headline gives no detection rate or evidence stage (Cardiovascular Business). - Microsoft's AI tool beat generalist doctors at step-by-step diagnosis, reported as a study in Science (The Next Web, Unite.AI). Details are in section 9. - A Korea IT Times series installment says Korean imaging AI must turn overseas clearance into sustained clinical use (Korea IT Times).

Clinical decision support

- AI breast cancer risk models guide triage and nurse-led patient counseling (Oncology Nursing News). - Chronic kidney disease (CKD) risk stratification, payment models and large language models (LLMs) aim to close care gaps (The American Journal of Managed Care). - Microsoft has developed an AI tool that diagnoses patients step by step (Mezha.net). - FDA clearance of AI-enabled stethoscopes is covered in section 7. 【Generative AI and clinical documentation】 - Zoom expands AI clinical notes across video, phone and in-person visits (Digital Health News). - A market-research press release projects the healthcare data center market to reach USD 37.11 Billion by 2035, citing AI genomics, electronic health record resilience and hosted capacity (PR Newswire UK). This is a vendor forecast, not clinical evidence. 【Hospital operations and care delivery】 - Black Book's new 2027 accountability report examines the full cost of hospital revenue cycle management (RCM), including outsourcing, AI savings claims and vendor consolidation (Rutland Herald). - Five AI tools to spot prostate cancer are backed for NHS use (pharmaphorum). The headline does not say which body backed them or what evidence supports them. - A piece on how AI is enhancing home healthcare (PEI Private Equity). - A piece on who is absorbing the liability for AI (Hospitalogy). - A list article ranks 10 healthcare AI development companies (Yonkers Times). 【Drug discovery and clinical trials】 No headline states a development stage (preclinical, Phase I and so on). - Warp Sciences launches with $215 million to bring AI drug discovery into the lab (Kingy AI). - Pathos AI raises $237M in a Series E at about a $3B valuation, led by Bristol Myers Squibb (Wowtale). - Bristol Myers Squibb is betting on AI drug discovery and wider trial access (Yahoo Finance). - Miguel Bronchud says AI can guide cancer drug discovery, but clinical validation still matters (Oncodaily). - The US Treasury proposes expanding investment curbs on Chinese AI and biotech firms. This is a proposal, not a final rule (ANI News). 【Medical device approvals and regulation】 - Coreline Soft received FDA 510(k) clearance for AVIEW Aorta Metrics, expanding its chest CT AI portfolio in the US. This is a clearance (Ortho Spine News). - The FDA cleared new AI-enabled stethoscopes that combine analog and digital listening. This is a clearance, and the headline does not name the products (Cardiovascular Business). - Five AI tools for prostate cancer are backed for NHS use. The type of decision is not stated (pharmaphorum). - The US Treasury proposal on investment curbs is an investment-policy proposal, not a device approval (ANI News).

Safety, errors and bias

- A piece titled "Algorithmic gaslighting" discusses hidden harm of AI in healthcare (TRT Afrika). - Who decides when medical AI may be updated? (Medical Xpress) - Seven years on, Japanese doctors and the public are still hesitant about AI-led care (Bioengineer.org). - Job displacement and the economic realities of AI in healthcare (Hospitalogy). - HIMSS AI Forum attendees discuss healthcare AI, Part 2 (MobiHealthNews). - Nelson Advisors' weekly roundup of European HealthTech, MedTech and Health AI is dated October 23, 2026 (healthcare.digital). - AI fails to compensate for night-shift fatigue among residents (AuntMinnie). 【Research results (major journals)】 - Microsoft's AI diagnostic system MAI-DxO is reported at 83% accuracy on complex cases (KuCoin). Business+IT adds that the result was published in Science and that physicians scored 36%. Unite.AI says it outperformed generalist physicians. - Around Prague reports that the AI beat generalist doctors on hard cases, but only in simulation. This indicates a simulated evaluation, not clinical use. - Brief IA describes Microsoft testing a medical AI orchestrator. Human Resources Director asks whose job is next. - Hindustan Times argues AI is neither poison nor panacea for doctors.

Developments in Japan

- Japan's Ministry of Health, Labour and Welfare will verify the effectiveness of AI in suicide prevention, in the context of social media consultations. 47NEWS notes challenges remain (Yahoo! News Japan, 47NEWS). - Business+IT reports the MAI-DxO result described in section 9. - A Vietnam.vn piece says that however well AI detects lesions, physician verification is still needed. - ThinkIT carries an article on moving from embedded systems to image AI to medical AI, with Kaggle as a help (ThinkIT). - Bioengineer.org reports that Japanese doctors and the public remain hesitant about AI-led care (see section 8).

Statements by key figures

No notable developments.