AI Reshapes Emergency Cardiology, Drug Discovery, and Cancer Screening

This week's health roundup covers a cluster of AI breakthroughs spanning emergency cardiology, drug discovery, cancer screening, and environmental health — with one cautionary note on the hidden costs of AI infrastructure.

FDA Clears First AI Tool for Paramedics to Detect Heart Attacks in the Field

The US Food and Drug Administration has granted De Novo clearance — a relatively rare regulatory pathway — to Powerful Medical's Queen of Hearts algorithm, an AI model that analyses ECG readings to identify ST-elevation myocardial infarctions (STEMI) before a patient reaches hospital. Fewer than ten De Novo designations are issued in any given year, underscoring the significance of this approval.

STEMI, caused by a complete blockage of a coronary artery, is a medical emergency where minutes determine outcomes. Currently, paramedics must transmit ECG data to a hospital cardiologist for remote interpretation — a process that can delay activation of a cardiac catheterisation team. Queen of Hearts is designed to flag the diagnosis autonomously in the ambulance, giving emergency services the confidence to mobilise that team earlier. Independent evaluations published in TCTMD suggest the model's performance is comparable to expert cardiologist review. The device is designated for paramedic-supervised use and does not replace physician oversight at the hospital.

Rentosertib: The World's First Fully AI-Designed Drug Reaches Phase III, with Early Evidence of Biological Age Reversal

Two significant milestones arrived this week for rentosertib, developed by Insilico Medicine for idiopathic pulmonary fibrosis (IPF) — a progressive scarring of the lungs with no cure and a median survival of three to five years after diagnosis.

First, Phase 2a data from 42 IPF patients, published in Nature Biotechnology in collaboration with researchers at Harvard Medical School and Stanford University, showed that patients treated with rentosertib demonstrated reductions in predicted biological age of up to three years, measured across six independent proteomic aging clocks. This is early-stage clinical evidence — 42 patients, Phase 2a — and biological age biomarkers are surrogate endpoints, not hard clinical outcomes such as survival or lung function. The consistency across six separate measurement methods is notable, but the finding should not be read as proof that the drug extends lifespan.

Second, Insilico Medicine announced that the first patient has been dosed in GENESIS-IPF-3, a 52-week, 320-patient, placebo-controlled Phase III trial launched at Peking Union Medical College Hospital in Beijing. What makes rentosertib historically distinctive is that both its biological target (the TNIK enzyme, implicated in lung-fibrosis pathways) and its molecular structure were identified and designed by generative AI — a first in pharmaceutical history at the Phase III level.

NYU Langone AI Predicts Breast Cancer Risk Over Five Years with Greater Accuracy

Researchers at NYU Langone Health have published results for NYU-DRP, a deep-learning model trained on 313,531 longitudinal three-dimensional mammograms from 161,165 women. When tested against held-out data, NYU-DRP predicted five-year breast cancer risk with 72% accuracy, compared with 70% for single-scan 3D AI and 68% for 2D AI models.

The practical implication is that a model capable of incorporating imaging data over time — rather than reading each scan in isolation — may eventually allow clinicians to offer individualised screening intervals rather than the current age-based population schedules. Women at lower predicted risk might safely extend the interval between mammograms; those at higher risk could be screened more frequently or referred for additional imaging sooner. This research is peer-reviewed, but the NYU Langone announcement notes that a prospective clinical trial is the next step before the tool could be adopted in practice.

ARPA-H Funds $62.7 Million Push for Autonomous AI in Heart Failure Care

The US Advanced Research Projects Agency for Health (ARPA-H) has awarded $62.7 million across six research teams to develop fully autonomous AI systems for managing heart failure — the leading cause of hospitalisation in adults over 65 in the United States. The first tranche of $33.7 million is already disbursed.

Recipients include Kaiser Permanente, Duke University, Stanford University, Tempus AI, and Atman Health. The goal is to build AI-driven treatment decision tools that carry FDA authorisation to act without requiring a clinician to approve each individual step — an ambition that goes substantially further than current AI diagnostic aids. Heart failure management typically involves frequent medication titration, fluid monitoring, and device adjustments; the intent is for AI to handle these autonomously for patients who lack regular access to specialist review. This is the most significant federal investment to date in clinical-grade autonomous AI, and it will raise important questions about liability, informed consent, and equitable access as the systems are developed.

AI Data Centre Expansion Linked to 1,300 Projected Premature Deaths in the US by 2028

A report authored by former US Environmental Protection Agency (EPA) officials warns that the rapid expansion of AI data centres — accelerated by rollbacks in environmental regulation — could generate air pollution sufficient to cause approximately 1,300 premature deaths and 600,000 asthma symptom episodes annually in the United States by 2028, at an estimated public health cost exceeding $20 billion per year.

The mechanism is direct: data centres require large quantities of electricity; much of that electricity is still generated by gas and coal power plants that emit fine particulate matter and nitrogen oxides; those pollutants cause and exacerbate respiratory and cardiovascular disease. The report raises a structural tension that policymakers will need to address: the same AI systems now being praised for their potential to accelerate drug discovery and improve diagnostics depend on infrastructure that, under current energy conditions, carries measurable public health costs. For communities in Cyprus and the broader eastern Mediterranean — already experiencing intensifying heat and deteriorating summer air quality — the relationship between energy infrastructure, air quality, and health is not abstract.

---

This article is a journalistic summary of published research and public-health developments. It is not medical advice. Readers should consult a qualified clinician before making any change to their treatment, medications, or health-screening schedule.