AI in Medicine News — 2026-11-08 (Sun)

Overview

Activity centered on diagnostics and imaging, clinical decision support, and drug discovery partnerships. In regulation, the FDA granted a Breakthrough Device Designation for a skin pathology AI tool, and the FDA's request for information on AI in early phase clinical trials drew legal commentary. Accountability for healthcare AI failures was also a recurring theme.

Diagnostics and imaging

- A new study reports benefits of AI analysis for ultrasound reports and incidental findings of hepatic steatosis. The headline gives no accuracy figures (Diagnostic Imaging). - Ultrasound radiomics with machine learning shows modest accuracy in assessing fetal ventricular disproportion (Bioengineer.org). - An AI ECG model is reported as validated as an adjunct for heart attack triage. The headline does not state the performance level (Bioengineer.org). - Machine learning is reported to sharpen leukemia diagnosis, but it still awaits rigorous validation, so the evidence is at an early stage (Bioengineer.org).

Clinical decision support

- Mayo Clinic Platform launched an AI-driven clinical decision support tool (Fierce Healthcare). A separate outlet describes it as an AI chatbot for doctors (Pain News Network). - A report says AI may step in to fill healthcare workforce gaps. This is the outlet's framing, not a demonstrated result (Medscape). - A machine learning model predicts respiratory distress risk in patients with acute pancreatitis (Bioengineer.org). - A machine learning model predicts kidney injury in children with sepsis (Bioengineer.org). - The headlines for the two prediction models do not state performance figures or whether they were validated in clinical use. 【Generative AI and medical documentation】 - Dimension NXG launched OJAS Forward Deployed Engineering to take hospital AI from pilot to patient, six months ahead of India's data protection deadline (ANI News). 【Hospital operations and care delivery】 - HIMSSCast discusses what drives real return on investment from AI (Healthcare Finance News). - A new trial of an AI phone assistant will monitor patients with long-term illness. It is a trial, so no outcomes are reported yet (Medical Xpress). - CureMeAbroad is making global healthcare more accessible through AI (BioSpectrum India). - A Q3 CY2026 deep dive on DOCS describes AI-driven product expansion meeting margin compression (TradingView). - Home-based patient care is among the AI applications being tested at a next-generation 5G and 6G research centre (The Straits Times). 【Drug discovery and clinical trials】 The headlines do not state development stages (preclinical, Phase I, etc.). - QIAGEN expanded its AI push with new tools for biological research and drug discovery (BioPharma APAC). - Chai Discovery and Roche will collaborate on AI drug discovery (Contract Pharma). - Big Tech's AI ambitions could be a boon for biotech startups (Axios). - A review finds that AI drug discovery tools dominated early-stage pharma research after COVID-19 (Bioengineer.org). - SK Biopharmaceuticals is building a massive AI dataset to speed up drug discovery (The Korea Times). 【Device approvals and regulation】 - Designation: the FDA granted Breakthrough Device Designation to the Cleo Skin AI pathology tool. This is a designation, not an approval or clearance (Clinical Lab Products). - Guidance process: a law firm commentary discusses the FDA's request for information (RFI) on AI in early phase clinical trials as a chance to help set the rules. This is an RFI, not final guidance (Crowell & Moring LLP). - Other: a Drug Topics piece takes an inside look at what it calls the first AI-prescribing technology. The headline does not state its regulatory status (Drug Topics). - A report says Roche Holding stock gains AI and diagnostics catalysts (ad-hoc-news.de). No market view is offered here.

Safety, errors, and bias

- A piece titled "Who Gets to Be Wrong? The Double Standard Shaping AI Accountability in Healthcare" examines accountability (The National Law Review). - A piece titled "Transparency is Not Accountability: What Hospitals Can Learn from the Challenges of Open AI Models" addresses hospital lessons (Digital Health News). - Medscape asks who is accountable when healthcare AI fails (Medscape). 【Research results (major journals)】 - A K&L Gates–CMU conference examined the opportunities and challenges of AI in medicine (Carnegie Mellon University). - AI assesses heart function from MRI better than clinicians, according to the headline. The study design and evidence level are not stated (Medical Xpress).

Developments in Japan

- NTT, NTT Docomo Business, and Igaku-Shoin have started a joint demonstration with Ubie toward implementing a medical AI information platform (Mix Online). - A Vietnamese outlet reports that simulation and AI cannot replace practice in clinical settings, in the context of medical training (Vietnam.vn). - Yakuodo renewed its AI skin diagnosis function (Diamond Chain Store Online). - An event in Oita, held for the first time there, aims to promote exchange between Kyushu, Taiwan, and Silicon Valley, with researchers discussing AI (Oita Godo Shimbun). - A horse racing AI item (SPAIA Keiba) is unrelated to medicine and is not covered.

Statements by key figures

No notable developments.