AI in Medicine News — 2026-10-11 (Sun)

Overview

Diagnostics and clinical decision support were the most active areas, with headlines on ECG, pancreatic, stomach and skin cancer, pathology and stroke tools, several of which note that evidence is still limited (Bioengineer.org, Medical Economics, Basingstoke Gazette). In drug discovery, headlines covered a collaboration, a planned IPO, free computing services and a $1.8 Billion data pool (The Pharma Letter, FirstWord Pharma, Newswise, Distilled Post). No regulatory approval or designation appears in the headlines. Cybersecurity and evidence-tracking appear as safety themes (Healthcare IT News, AuntMinnieEurope).

Diagnostics and imaging

- AI is used to turn a routine ECG into a predictor of hidden heart artery disease. The headline gives no accuracy figure or study stage (Bioengineer.org). - A Mayo Clinic AI is reported to detect pancreatic cancer years before diagnosis. The headline gives no performance figure or study stage (Medical Economics). - AI agents are coming to the pathology lab, but the evidence is "not there yet" (Bioengineer.org). - AI shows promise against stomach cancer, but most tools are not ready for the clinic (Bioengineer.org). - AI is being integrated into ophthalmology clinical pathways for retinopathy and glaucoma (Il Sole 24 ORE).

Clinical decision support

- Zhong Nanshan's team unveiled an AI pathology model. Lung cancer genetic testing costs in China could drop to a few hundred yuan, according to the headline's wording of "could" (BigGo Finance). - The NHS is rolling out an autonomous AI skin cancer tool across all 18 South East trusts. This is a deployment, and the headline does not describe an approval (Basingstoke Gazette). - Evidential deep learning, which lets AI express uncertainty, is described as bringing trustworthy stroke detection closer to the clinic (Bioengineer.org). - Dermatology Times published a practice brief covering AI capacity, emerging oral therapies and new directions in skin cancer care (Dermatology Times). - Newsweek asks what Utah's AI experiment means for hospitals. The headline does not describe the experiment (Newsweek). 【Generative AI and clinical documentation】 - Healthleap raised $38M to expand its AI platform for undiagnosed conditions. The headline does not state that the platform involves documentation or generative AI (Voice of Healthcare). 【Hospital operations and care delivery】 - AI is raising healthcare cybersecurity risks, and the article covers how to fight back (Healthcare IT News). - AI extends hospital care to patients' homes (Healthcare Finance News). 【Drug discovery and clinical trials】 The headlines do not state a development stage for any item. - Lonza and Onava are collaborating to speed AI-native biologics discovery (The Pharma Letter). - AI-powered drugmaker Iambic could claim an $800M+ valuation in an IPO (FirstWord Pharma). - Apodex is opening free AI computing services to accelerate drug research (Newswise). - The Virtual Biology Initiative pools $1.8 Billion to build open AI training data for drug discovery (Distilled Post). - A 17-year-old Indian-American West Virginia teen, Pavan Subramani, is reported to have developed an AI method that sampled a key molecule. The headline is truncated and gives no further detail (The Times of India). 【Medical device approvals and regulation】 No notable developments.

Safety, errors and bias

- A new metric judges medical AI by the entities it keeps rather than the words it matches. The headline gives no performance figures (Bioengineer.org). - An article explains continuous evidence tracking in clinical AI systems (AuntMinnieEurope). - Nigerian doctors are using ChatGPT to support consultations, according to Kunle Kakanfo. The headline does not address accuracy or oversight (TheCable). - The cybersecurity risk item in section 5 is also a safety theme (Healthcare IT News). 【Research results (major journals)】 No notable developments. The only related item is a Retraction Watch weekend roundup on OpenAI's preprint dump, Taylor & Francis leaving Informa, and arXiv's preprint cap. It is publishing news, not a clinical result (Retraction Watch).

Developments in Japan

No Japanese headline concerns medicine specifically. - Kaunet began operating Sapunavi AI, an AI-equipped purchasing diagnosis service (online.bci.co.jp). - renue launched an AI vulnerability assessment support service that includes static code analysis and on-site vulnerability assessment (PR TIMES). - A separate PR TIMES item describes security support for small and medium-sized companies, offering AI and engineer vulnerability assessment from 50,000 yen, with the assessment fee fully discounted when remediation is requested (PR TIMES). - Two Vietnam.vn items, which are not Japanese news, describe AI applied to medical facility management and AI robots expanding in healthcare, industry and daily life (Vietnam.vn).

Statements by key figures

- Geoffrey Hinton (title not given in the headline) urges FDA-style AI regulation and admits he was slow to recognize the risks (Brand Spur). - Kunle Kakanfo (title not given in the headline) said Nigerian doctors are using ChatGPT to support consultations (TheCable).