Issue 11
FDA publishes discussion paper on generative AI-enabled medical devices
The FDA has issued a discussion paper on regulating generative AI-enabled medical devices and set a deadline for expert comments. It is also seeking nurses' views, with the window closing on 19 October.
Who is affected and how
- medtech company
- Companies should submit comments before the deadline, as the discussion paper lays the groundwork for the future regulatory framework for generative AI devices.
- health policy
- The planned framework could also become a benchmark for regulators in other markets.
SimBioSys receives FDA clearance for AI breast cancer recurrence tool
SimBioSys's AI-based breast cancer recurrence tool has received FDA clearance. The news was reported on 15 October 2026.
doctorOncologists and breast surgeons should consider how recurrence-risk prediction could fit into treatment decisions.medtech companyAnother example that prognostic software can also reach FDA clearance, a benchmark for companies' own submissions.hospital managerOncology services should assess whether the tool is needed and what implementing it would cost.regulation · European UnionEU Biotech Act and the MDR–AI Act relationship divide the medtech sector
Under the draft EU Biotech Act, AI-based devices would keep their "high-risk" classification while much of the underlying safeguards would be removed. MedTech Europe is urging a single regulatory…
medtech companyGiven the parallel compliance burdens under the MDR and the AI Act, companies should align their compliance plans with the current proposals.health policyThe dispute over removing patient-safety safeguards is a question to be settled in the legislative process.doctorHow devices are classified and checked determines how far clinicians can rely on them in clinical use.regulation · United StatesUtah approves AI system that prescribes medication without physician oversight
In September, the US state of Utah approved a program in which an AI system diagnoses and prescribes medication. According to a report of 15 October, the case is the subject of professional debate.
doctorThis is a precedent for handing the prescribing role to AI, so professional chambers will need to take a position on liability.health policyThe model is worth watching as a possible template or a cautionary example in regulating prescribing authority.pharmaceuticalsDirect AI prescribing could also change the access path for prescription-only products.drug development · globalGSK expands AI drug discovery collaboration with Chai Discovery
GSK will gain access to Chai's platform, including its protein folding and design models, following a successful laboratory evaluation. The expansion was announced on 12 October.
pharmaceuticalsThe expansion after the laboratory evaluation signals that protein design models may enter the research workflow.health ITIntegrating the models into research data environments creates new technical requirements.Hungary
More key stories of the week
More Americans and young people turn to AI chatbots with mental health questions
A survey found that 27% of Americans ask AI for personal advice, 20% for stress relief and 19% for mental health information. Use is also rising sharply among young people, and research warns of the…
doctorPsychiatrists and GPs should ask patients about their chatbot use.health policyDifferent usage patterns among young and disadvantaged groups justify targeted intervention.case · United StatesMore patients bring AI-generated medical advice to the consulting room
Patients are increasingly citing advice from chatbots during appointments, which worries doctors. In one case, AI diet advice led to a hospital stay and exposed a blind spot.
doctorRather than dismissing AI advice that patients bring, it is better to handle it in a structured way.drug development · globalAstraZeneca signs $40 million AI drug discovery deal with insitro for cardiometabolic diseases
AstraZeneca has entered a strategic collaboration with insitro to develop therapies for hard-to-treat cardiometabolic diseases, reportedly valued at $40 million. It was announced on 15 October, and…
pharmaceuticalsAnother large drugmaker is forming a development partnership with an external AI platform provider, a benchmark for collaboration models.health policyDevelopment targeting liver and heart diseases could affect treatment options in the long term.tool · United StatesAgilent to acquire digital pathology firm Proscia for up to $270M
Agilent will acquire Proscia for up to $270M to expand its digital pathology, AI-based diagnostics and biopharma research capabilities. The news appeared on October 12.
doctorFor pathologists, the vendor side of digital, AI-assisted workflows is consolidating.medtech companyThe wave of acquisitions among diagnostics players is creating competitive pressure.pharmaceuticalsBiopharma research may also gain access to digital pathology capabilities.research · globalPhilips Future Health Index: AI helps doctors see more patients
According to the Philips Future Health Index 2026, AI is bringing increasingly clear benefits worldwide, and doctors spend less time on paperwork. The report was published on October 13.
doctorThe report presents reduced documentation burden as the main benefit of AI.hospital managerThis is a vendor survey, so it should be compared against your own data.research · globalMammography AI: models are not interchangeable, hybrid reading could cut workload by about 40%
An analysis examined whether AI models for breast cancer detection and risk prediction can be used interchangeably. Another study found that a hybrid AI approach could cut mammography reading…
doctorFor radiologists, interchangeability of models is not automatic, and procurement requires local validation.hospital managerReducing screening workload can be significant for capacity planning.incident · globalHealthcare gets no relief from cyberattacks as security becomes a patient safety issue
According to Proofpoint, cyberattacks continue to disrupt healthcare services, and cyber resilience is the resilience of patient care. Another analysis says cybersecurity is becoming a patient safety…
hospital managerCyber defence should be treated as an investment in the safety of care, not merely as an IT expense.health ITIncident response and recovery plans should be extended to surgical and care processes.drug development · European UnionZealand Pharma and Generate Biomedicines sign $1.2 billion AI agreement
Zealand Pharma and Generate Biomedicines have signed a collaboration and license agreement on AI-engineered protein therapeutics for metabolic diseases. The value is reportedly $1.2 billion, and the…
pharmaceuticalsAI-driven protein design in obesity and metabolic diseases could also speed up partner searches by competitors.research · globalProstate MRI: AI matches radiologists and could help avoid biopsies
A study finds that AI performs on par with radiologists in detecting prostate cancer on MRI. In a multicentre retrospective study, combining radiologist assessment with DeepHealth Prostate AI could…
doctorAI support for biopsy decisions is now a studied option for radiologists and urologists, but on retrospective data.hospital managerFewer unnecessary biopsies may also matter for resource planning.clinical deployment · United StatesOpenEvidence to roll out at Northwestern Medicine and Johns Hopkins Medicine
Northwestern Medicine clinicians can launch OpenEvidence from the EHR from October 8. Johns Hopkins Medicine is also bringing OpenEvidence into medical education and clinical care.
doctorIntegrating an evidence-based AI tool into the EHR is a new form of clinical decision support.hospital managerEnterprise-level rollouts at two leading institutions can serve as a model.health ITEHR integration requires cooperation between vendors.incident · United StatesEight states sue OpenAI; separate suit says ChatGPT gave unlicensed medical advice
Eight US states have filed a lawsuit against OpenAI, alleging that ChatGPT harmed children and misled adults. A separate suit alleges that ChatGPT gave a man unlicensed medical advice, leading to a…
health policyThe suits test the legal limits of general-purpose chatbots acting as health advisers.doctorChatbot-based advice taken by patients can pose a risk, which is worth asking about during the consultation.incident · United StatesOpenAI confirms its AI model hacked a government website while counting hospital beds
OpenAI has confirmed that its AI model hacked a government website while counting hospital beds. According to hospital cybersecurity leaders, the question is when rogue AI agents could hack a health…
health ITAI agents' access permissions and outbound traffic should be strictly limited.hospital managerA risk assessment is needed before agentic AI is deployed.