New FDA approvals in epilepsy, fertility, and heart failure
Three medical device clearances, new GLP-1 side-effect signals, and a UK-France biomedical alliance mark this week's health landmark.
By Dr. Asher Knippel
A cluster of regulatory decisions and a major European research partnership defined the closing days of May 2026 for global medicine.
Friday, 30 May: ECoG Assistant Brings AI to Epilepsy Brain Monitoring
The US Food and Drug Administration approved NeuroPace's ECoG Assistant™ on 29 May 2026 — a clinician-facing AI feature built on the world's only long-term intracranial electroencephalography (EEG) dataset. The tool automatically analyses years of neural-signal recordings from patients living with NeuroPace's implanted neurostimulation device, surfacing patterns and anomalies that would otherwise require hours of manual review. The aim is to help neurologists adjust stimulation settings and treatment decisions with greater speed and confidence. NeuroPace will formally showcase ECoG Assistant at the American Society for Stereotactic and Functional Neurosurgery (ASSFN) Annual Meeting on 30 May 2026. This is the first in a planned AI-tool suite for the platform; the company's next-generation patient data management system is separately under FDA review. Epilepsy affects approximately 50 million people globally; effective management depends critically on interpreting long, complex brain-signal records.
Thursday, 29 May: Non-Invasive Wearable Sensor Estimates Heart-Failure Pressure
Cardiosense received FDA clearance on 28 May for its PCWP Analysis Software — the first AI-enabled tool to estimate pulmonary capillary wedge pressure (PCWP), the key marker of dangerous fluid accumulation in heart failure, without a catheter. Patients wear a small, non-invasive chest sensor; the AI interprets the acoustic signal to estimate PCWP in near real time. Today's gold standard requires threading a catheter through the heart into the pulmonary artery — an invasive procedure limited to hospital settings. If validated in larger real-world studies, this approach could allow earlier hospital discharge and home-based monitoring for high-risk patients, a particular benefit for older adults across Cyprus and the broader Mediterranean region.
Thursday, 29 May: IVF Embryo-Selection AI Cleared on Randomised Trial Evidence
Alife Health's Embryo Predict received FDA clearance on 28 May 2026, backed by a prospective randomised multi-centre trial conducted across seven US fertility clinics. The tool applies AI to evaluate and rank embryos for transfer during in vitro fertilisation (IVF), a decision currently dependent on embryologist visual scoring with documented high inter-observer variability. The prospective randomised design — uncommon for AI fertility tools, which more often rely on retrospective validation — provides a stronger evidence base for the clearance. IVF use is rising across Europe, including in Cyprus, where the national health system covers a limited number of cycles per patient.
Thursday, 29 May: Penn Study Surfaces Overlooked GLP-1 Side Effects Via Social Media
University of Pennsylvania researchers analysed more than 410,000 Reddit posts from approximately 67,000 users discussing GLP-1 receptor agonists — semaglutide (Ozempic, Wegovy) and tirzepatide (Mounjaro, Zepbound) — and identified a set of symptoms not prominently flagged in current prescribing information: irregular or heavy menstrual bleeding, intermenstrual spotting, and temperature-related complaints including chills, hot flashes, and fever-like sensations. The AI-assisted text analysis, updated 28 May 2026, confirmed well-documented gastrointestinal effects but found reproductive and thermoregulatory complaints at frequencies suggesting they are systematically underreported in formal clinical trials, which have historically enrolled fewer women of reproductive age. The researchers propose that large-scale social media analysis can serve as a pharmacovigilance early-warning system. GLP-1 receptor agonists are now among the most-prescribed drugs globally; European Medicines Agency safety reviews of this drug class are ongoing.
Thursday, 29 May: AstraZeneca and Nucs AI Target Prostate Cancer Response Prediction
Nucs AI announced a collaboration with AstraZeneca on 29 May to develop AI models that predict individual patient response to a PSMA-targeted radioconjugate in metastatic prostate cancer. The collaboration builds on Nucs AI's existing platform for predicting response to lutetium-based radiopharmaceutical therapies and will extend those models to the characteristics of a next-generation agent. The intent is to shift patient selection from population-level eligibility criteria to patient-level response prediction: which individual is most likely to benefit. PSMA-targeted radiotherapy has become one of the most active areas in advanced prostate cancer research; this is pre-clinical AI modelling work and results will require prospective validation before influencing prescribing practice.
Thursday, 29 May: Oxford and Institut Pasteur Launch Cross-Border Biomedical Alliance
The University of Oxford, Université Paris Cité, and the Institut Pasteur joined with the UK's Diamond Light Source and France's Synchrotron SOLEIL on 29 May 2026 to launch a biomedical and AI health research alliance. The partnership's initial focus spans women's health, infectious diseases, and pandemic preparedness — three domains where both institutions have long-standing expertise. The alliance will combine advanced synchrotron imaging, which resolves protein and cellular structures at resolutions otherwise inaccessible to most labs, with large-scale AI analysis of biomedical datasets. Direct clinical applications lie years ahead, but the infrastructure strengthens European capacity in areas directly relevant to Mediterranean and global public health.
Friday, 30 May: Colorado Restricts AI in Psychotherapy and Mandates Crisis Referral
Colorado Governor Jared Polis signed two AI mental-health bills on 29 May 2026. House Bill 1195 restricts the autonomous delivery of psychotherapy by AI systems — the first US state law to do so — while permitting AI tools that assist licensed therapists. House Bill 1263 requires AI chatbot developers to detect signs of suicidal ideation in users and automatically refer them to crisis services, including the 988 Suicide and Crisis Lifeline. Colorado becomes the first US state to enact both types of restriction simultaneously. Mental health access is a recognised public-health priority in Cyprus and across the EU; the European AI Act's provisions for high-risk AI systems in healthcare are expected to require similar safeguards at the Union level.
The content of this article is journalistic reporting and is not medical advice. Readers should consult a qualified clinician before making any change to their treatment, medication, or healthcare decisions.