Healthcare AI
Latest stories covering Healthcare AI, with summaries and links to related coverage.
6 stories · 6 reports · page 1 of 1
5 stories · 5 reports in the last 7 days
Latest report · Counts as of (UTC)
Models in this coverage
No tracked models are linked to this topic’s reports in the last 30 days.
Connected topics
Topics connected through shared stories across the archive.
Latest stories
A 10-year-old CNN outperforms frontier LLMs for skin cancer diagnosis
opinion
5 min read
A dermatologist tested GPT-6 Astra, Claude Fable 5.1 and a fine-tuned Inception v3 convolutional neural network on 200 randomly selected, biopsy-confirmed skin lesions from the public MILK10k dataset. The CNN correctly identified 125 of 200 lesions across six diagnoses, or 62%, while GPT-6 Astra and Claude Fable 5.1 identified 84 and 85, or 42% each. Inception v3 was introduced by Google in 2015 and was fine-tuned on 68,011 ISIC Archive clinical and dermoscopic images for about 20 minutes on one H100 GPU. The test set contained 100 lesions requiring further care, including 30 melanomas, and 100 benign lesions considered safe to discharge. The CNN identified 13 melanomas as its first choice, versus eight for GPT-6 Astra and 11 for Claude Fable 5.1. With a second diagnostic choice allowed, the CNN reached 173 correct cases, compared with 134 and 138 for the LLMs. Both LLMs incorrectly labeled a pictured melanoma as a benign seborrhoeic keratosis and said the patient could be discharged. The sample size was small, each LLM response was generated once, and all lesions had been biopsied, making the cases unusually difficult. At its preset discharge threshold, the CNN discharged 18 melanomas, though matching its unsafe-lesion retention rate to the LLMs reduced that figure to two or one. The browser-based CNN runs locally with WebGPU and does not upload images.
New AI Is More Than Twice as Good at Diagnosing Diseases as Doctors
research
3 min read
Microsoft’s AI Diagnostic Orchestrator, or MAI-DxO, achieved 83.2% diagnostic accuracy on complex cases from The New England Journal of Medicine, compared with 36.1% for generalist physicians using reference materials. Physicians without those materials scored about 22%. In a separate emergency-department admissions test, MAI-DxO scored 80.3%, versus 46.8% for generalist physicians. Microsoft reported its strongest results with OpenAI’s GPT-5 and additional improvements with GPT-5.5. MAI-DxO begins with a patient’s chief complaint, then sequentially selects follow-up questions and diagnostic tests before reaching a conclusion. It assigns AI agents roles including proposing diagnoses, choosing tests, challenging assumptions, and weighing costs. Microsoft said simulated-case estimates indicated the system reached diagnoses using fewer tests and procedures at roughly half to one-third of the cost, though those estimates were not actual medical expenses. The findings, first posted as an arXiv preprint last year, have been published in Science. Microsoft AI chief Mustafa Suleyman said real-world use would require clinical testing. Dr. Dominic King, Microsoft’s AI health lead, said testing was expected to begin at Beth Israel Deaconess Medical Center in Boston within months. The researchers said the performance gap would likely narrow against specialist physicians rather than generalists, and medical experts have said diagnostic benchmark performance does not establish that AI should replace doctors.
AI analysis of CT scans finds 42.9% of patients have coronary inflammation
research
3 min read
Biograph completed a U.S. study of Caristo Diagnostics’ AI-powered CaRi-Heart coronary-inflammation analysis in asymptomatic adults considered low risk and without known heart disease. The study paired high-resolution coronary CT angiography, cardiometabolic profiling, CaRi-Heart inflammation measurements and CaRi-Plaque plaque and stenosis analysis. Biograph said the study was available to clinically eligible members of its Black tier and that its manuscript has been published as a preprint. Biograph reported elevated coronary inflammation in 42.9% of participants and very high inflammation in at least one major coronary artery in 17.9%. About 47% of participants with no detectable coronary plaque had elevated perivascular inflammation, according to the findings. Cholesterol measures including ApoB and non-HDL-C, HbA1c, hs-CRP, and the AHA PREVENT-ASCVD risk calculator showed no correlation with the detected inflammation in this study. Biograph said the results may identify an earlier phase of coronary disease, but said further research is needed to validate the findings and determine the prognostic value of inflammation measurements and the benefits of care directed at them. CaRi-Heart has FDA De Novo marketing authorization to quantify coronary inflammation from routine coronary CT angiography, while CaRi-Plaque has FDA clearance for AI-assisted plaque and stenosis analysis.
Using thousands of AI agents to uncover new leads in biological data
research
5 min read
OpenAI, PerturbAI, and Scripps Research used thousands of GPT-Rosalind-powered AI agents to analyze a whole-brain CRISPR atlas containing data from millions of mouse brain cells. The agents examined individual gene disruptions, measured cellular responses, searched scientific literature, and combined their findings into a reasoning atlas linking genes, cell effects, publications, and possible biological mechanisms. The analysis took 32 minutes; the research team estimated comparable sequential human analysis and literature review would take months to a year. The distributed-agent system reproduced findings from human analyses and generated additional research leads. Its novelty and scientific-value assessments aligned with blinded human experts, according to the team. At the largest tested workload, a single agent cited 11 references, while distributed agents cited more than 700, and human experts preferred the distributed interpretations. The agents found that disrupting Tsc2, but not Tsc1, altered dozens of neuronal-communication genes in cerebellar cells, despite both genes regulating mTOR. They also identified a potential link between DCTN1 disruption, increased expression of cholesterol-production genes, and altered cholesterol transport in neurons. DCTN1 mutations cause Perry syndrome, a neurodegenerative disorder with Parkinson-like symptoms. The publicly available reasoning atlas is intended to help researchers prioritize laboratory experiments that can test these proposed mechanisms.
Anthropic's Claude autonomously designs protein binders targeting scorpion neurotoxins
research
2 min read
Anthropic and Citrini Research used an internal Claude model to design protein minibinders targeting five neurotoxins from four scorpion species. Following one prompt, Claude generated 318,484 candidate designs during a 24-hour autonomous run in Anthropic’s Claude Science research environment. Researchers tested 211 designs—180 selected by Claude and 31 added by the team—and found that 17 met their criteria for functional binders. The two strongest binders attached with high affinity to Cn2, a toxin produced by a Mexican scorpion. Laboratory evidence suggests the binders attach at or near the site targeted by an antibody that neutralizes Cn2 in mice. The binders have not been tested as antivenom in animals or humans, so their ability to treat scorpion stings remains unverified. Scorpion neurotoxins kill more than 3,250 people annually. An earlier Claude study produced minibinders for 14 of 15 target proteins, with reported hit rates of 22% to 35%, compared with a typical 10% to 15% baseline. Susana Vázquez Torres and Timothy P. Jenkins described the scorpion-toxin results as a promising first step, and two independent laboratories confirmed the findings.
Alternatives to Animal Testing Are Finally Taking Off
news
14 min read
Nonanimal methods for drug development and toxicity testing now include human organoids, organs-on-chips, computational models, and AI systems that analyze experimental data. Harvard’s 2010 lung-on-a-chip study used human cells in microfluidic channels and simulated breathing motions; it showed that tissue movement affected uptake of silica nanoparticles. Multi-organ chip systems can link as many as 10 organs, although they do not reproduce human physiology perfectly. U.S. law explicitly authorized nonanimal methods in preclinical drug studies through the FDA Modernization Act 2.0 in late 2022. The FDA said in 2025 that it would seek to make animal studies the exception rather than the norm in drug-safety testing. In September 2026, the agency issued a proposed rule replacing “animal tests” with “nonclinical tests” in drug-development regulations, recognizing validated human-cell systems, organs-on-chips, and computer models when appropriate. NIH announced in 2025 that applicants studying animal models would also need to incorporate nonanimal research. Validation, standardization, cost, data access, and scientific training remain barriers. Emulate reported that its liver-on-a-chip identified about seven-eighths of drugs that passed animal testing but later proved toxic to human livers. Oxford University and Janssen researchers reported that computational heart-cell simulations identified compounds causing a dangerous arrhythmia with 89 percent accuracy, versus 75 percent for animal studies. Many methods still lack rigorous head-to-head evidence across intended uses.