AI in Medicine News — 2026-10-07 (Wed)

Overview

- What moved the most was regulation. The UK government has accepted all 44 recommendations of the committee and has indicated that it will continue to monitor AI medical devices after approval. FDA has laid out its AI guidance plan for 2027. - In diagnosis, FDA cleared MRI body composition evaluation AI. In operations, Vitalize raised $31 million for hospital staffing and scheduling. - Reports on drug discovery mainly focus on funding and partnerships, and we cannot confirm any cases that have progressed from the headlines to the clinical stage.

Diagnosis/Image

- The FDA has granted clearance for body composition evaluation AI using MRI (Diagnostic Imaging, Radiology Business). The item name is not in the heading. The regulatory status is "clearance", not approval. - Regarding the distribution of FDA approvals, a study was reported that three-quarters of the approvals for AI medical devices are in the radiation field (PR Newswire, led by GGHz founder). - Hemophilia shows "early promise" in diagnosis and care (Hemophilia News Today). The evidence stage is early. Measurement targets and performance values are unknown. - In heart disease, self-adjusting AI models reportedly boosted early diagnosis (Bioengineer.org). Evaluation environment and performance values are unknown.

Clinical decision support

- Reviews show that AI medical tools are reaching patients before the evidence is ready (Bioengineer.org). - Taiwan's Cheng Kung University (NCKU) has developed an AI wound healing imaging system (Taiwan News). At the development stage, it is unclear whether there will be any clinical evaluation. - AI is expected to reduce wait times for skin cancer tests in Sussex, UK (SussexWorld). The headline is the expectation that the period will be shortened, not the actual results. - At COGC 2026, there was a talk that AI requires improved workflows, not additional tools (Oncodaily). - AI Health Bus will reportedly test Berkshire's productivity goals (Distilled Post). The content cannot be identified by the heading alone.

Generation AI and medical documents/medical records

- There is one report stating that ``Doctor AI will examine patients'' (WSB-TV). It is not possible to judge from the heading whether it applies to document creation and recording. - In Japan, introduction letter/FAX reading services have appeared (Section 10).

Hospital management/medical provision system

- Vitalize raises $31 million to expand its AI operating system for hospital staffing and scheduling (Fierce Healthcare). - In the UK, the government accepted all 44 recommendations from the NHS AI panel (Digital Health, pharmaphorum). - NHS has opened applications for the 6th Clinical AI Fellowship (Distilled Post). The targets and methods of the investigation are unknown.

Drug discovery/clinical trials

- In clinical trial screening, Ochsner uses AI to expand the scope of candidate selection for trials (Healthcare IT News). At the operational stage, the phase of testing is unknown. - In target discovery (pre-clinical stage), BullFrog AI identified multiple candidate drug targets in a depression project of one of the top five pharmaceutical companies (Stock Titan). It is "potential" and the verification stage is unknown. - In funding/partnerships, Baio Labs received funding to accelerate generative AI drug discovery (Startupticker). PacBio strengthens AI drug discovery with acquisition of TychoBio (Yahoo Finance). In both cases, the development stage is not mentioned in the headline. - Commentary: "AI biotech" is no longer a useful classification for investors (European Biotechnology Magazine).

Medical device approval/regulation

| Category | Contents | ｜ーー｜ーー｜ | Guidance (plan) | FDA presents AI guidance plan for 2027 (STAT) | | Organization | FDA has established an AI office. Commentary on the difficult part of practice from now on (The Clinical Trial Vanguard) | | Policy (acceptance) | The UK government has accepted all 44 recommendations. Continued monitoring of post-approval AI medical devices (Fierce Biotech, Medical Device Network, Technology Org, Innovation News Network) | | Commentary | Responding to the evolution of devices will become an issue due to review of UK AI regulations (MLex) | | Clearance | MRI body composition evaluation AI (item name is not in the heading). See Section 2 |

The UK case is about policy acceptance, not approval or designation of individual devices.

Safety/Error/Bias

- Nolla Health's AI app is reportedly dispensing acne prescriptions in Utah without going through a doctor (Quartz). The regulatory treatment is unclear from the headline. - WEF discussed the construction of a ``human-in-command'' system in which humans are ultimately responsible for medical AI. - Delays in the development of evidence (paragraph 3) and the burden of regulations on equipment evolution (paragraph 7) are both issues in terms of safety. - Spring Health's AI mental health guide is a provider-sourced document and is not an independent evaluation.

Research results (major journals)

- There is no headline that clearly states that the article will be published in a major magazine. - Cameras and AI have been shown to have the potential to help diagnose movement disorders in newborns (Medical Xpress). “Can be useful” stage. - A Romanian researcher was awarded an award in Spain for his work on using AI to recognize signs of depression on social media (Romania Insider). No clinical validation has been shown. It is unclear whether the content is medically specific.

Japanese movements

- Regarding the performance evaluation of medical AI, Niigata University of Health and Welfare announced that ``what data was used for evaluation'' is important (PR TIMES). - In joint research, five institutions including NEC, Taiho Pharmaceutical, and Waseda University have started joint research on next-generation cancer immunotherapy using quantum technology and AI (Business + IT). At the beginning stage of research, there are no results. - HiMaWaSa Sync has started offering ``Temasui'', a generation AI for small and medium-sized hospitals. Read letters of introduction and faxes and create a list (Medical Tech News). - Excluded from the scope are the AI Utilization Diagnosis Survey (PR published in the Mainichi Shimbun) and Kaunet's purchasing optimization tool, which cannot be evaluated based on the headline because the relationship to medical care cannot be confirmed.

Remarks from VIPs

- The list of important people being tracked is blank and there is no relevant information. - Reference: The headline includes Fred Ashbury's talk at COGC 2026 (Oncodaily). The title is not in the heading. The gist is that ``AI requires improved workflows, not additional tools.''