Artificial Intelligence at Mayo Clinic

Mayo Clinic is a nonprofit educational medical heart headquartered in Rochester, Minnesota, with further campuses in Arizona and Florida and a regional well being system spanning three Upper Midwest states. The group employed almost 85,000 individuals in 2025 and posted $473 million in working revenue. Mayo committed $9 billion in capital funding by its “Bold. Forward. Unbound.” enlargement. Independent rankings have named it Newsweek’s No. 1 hospital on this planet for the eighth consecutive 12 months.

That scale extends to synthetic intelligence. Mayo describes greater than 200 AI initiatives underway throughout levels of maturity, from early feasibility work to full medical deployment.

• In 2025, Mayo integrated 22 AI-enabled Mayo Clinic Platform options into medical observe and closed almost 200 new AI, biopharma, and diagnostics agreements.
• Since launching the Mayo Clinic Platform in 2019, it has assembled a analysis knowledge infrastructure spanning greater than 15 million affected person information and billions of radiology photos, lab outcomes, and medical notes.

This article examines two AI use instances that present how Mayo applies that knowledge and medical experience inside its personal operations.

• AI-Enabled ECG Screening for Early Disease Detection — Detecting asymptomatic coronary heart illness earlier than signs seem, utilizing knowledge generated by a routine, low-cost cardiac check.
• AI-Powered Chart Review with Record Time — Cutting the hours physicians spend manually reviewing fragmented, unsorted affected person information from different well being programs, liberating up extra time for direct affected person care.

We start by analyzing how Mayo Clinic applies AI-enabled ECG evaluation to deal with early, asymptomatic detection of coronary heart illness.

AI-Enabled ECG Screening

Asymptomatic left ventricular dysfunction is a precursor to coronary heart failure that customary screening typically misses, since diagnosing it has historically required an echocardiogram. This check wants specialised tools and a educated sonographer.

Heart failure now affects almost 6.7 million Americans, a determine projected to achieve 8.7 million by 2030, and cost the U.S. well being system an estimated $32 billion in direct medical spending in 2020 alone. Because the underlying dysfunction normally goes undiagnosed till signs seem, fewer than one in 4 eligible sufferers receive guideline-recommended remedy. When a situation is that this pricey to deal with late and this tough to catch early, the highest-yield AI funding is usually earlier detection constructed from knowledge a corporation is already amassing — which is what Mayo’s cardiology researchers got down to do.

Mayo screened greater than 625,000 paired ECG and echocardiogram information from its personal sufferers to assemble a research inhabitants, then trained a neural community on almost 98,000 of these pairs to acknowledge electrical patterns in a typical EKG tied to a weakened coronary heart pump. The method has since expanded effectively past that unique use:

• It now additionally flags atrial fibrillation, cardiac amyloidosis, aortic stenosis, hypertrophic cardiomyopathy, and organic age.
• It works throughout each conventional 12-lead ECGs and single-lead readings from smartwatches and a digital stethoscope.
• The amyloidosis model was validated throughout 25,525 sufferers at 4 U.S. well being programs, attaining 78.9% sensitivity and 91.2% specificity.

Model high quality right here tracks the depth of Mayo’s personal historic knowledge relatively than any specific algorithmic novelty.

The screening step provides no new work for the clinician or the affected person: no new check is ordered, because the AI reads the ECG already collected throughout a routine go to. A optimistic sign prompts a confirmatory echocardiogram or referral that may not in any other case have been ordered. Portable variations lengthen that screening past the clinic—an AI-enabled digital stethoscope flagged twice as many instances of peripartum cardiomyopathy as customary care in a Nigerian obstetric research, illustrating how AI adoption tends to stay when it rides an present workflow step relatively than including a brand new one to recollect.

Mayo tested the device prospectively within the EAGLE trial, enrolling 22,641 sufferers throughout 348 major care clinicians and 45 medical facilities in Minnesota and Wisconsin over eight months. The outcomes, published in Nature Medicine:

• AI-guided screening elevated diagnoses of low ejection fraction by 32% general in contrast with regular care — about 5 further diagnoses per 1,000 sufferers screened.
• The 12-lead algorithm is now FDA-cleared and licensed to Anumana, an organization Mayo co-founded with nference, a healthcare knowledge analytics agency. At the identical time, Mayo partnered with Eko Health to develop and commercialize a single-lead model for handheld and wearable units.
• Newer purposes, equivalent to amyloidosis detection (cleared by the FDA in April 2026), are nonetheless early in industrial rollout.

Screenshot of:  Illustration for Anumana’s ECG-AI LEF synthetic intelligence (AI) mannequin. (Source: Cardiovascular.com)

Taken collectively — randomized trial proof, regulatory clearance, and out of doors licensing — that is one in all Mayo’s most evidence-backed AI purposes so far, not an experimental pilot.

AI-Powered Chart Review with Record Time

Ahead of a affected person go to, Mayo Clinic physicians typically face dozens and even tons of of pages of medical information to overview — a burden compounded by the truth that many sufferers come to Mayo in search of a 3rd or fourth opinion, arriving with unsorted paperwork from different well being programs. According to Dr. Alexander Ryu, an inside medication doctor and vice chair of innovation for Mayo’s Department of Medicine, the hospital receives tens of tens of millions of pages of information annually, and wanted a solution to floor the essential data buried in that quantity.

That quantity downside isn’t distinctive to Mayo. According to a 2025 survey revealed within the Journal of the American Medical Informatics Association, 77% of the well being system leaders surveyed cited immature AI instruments as one of many greatest obstacles to AI adoption — underscoring why a device fixing a well-defined administrative bottleneck, relatively than a broad diagnostic declare, was the extra tractable place to begin.

Record Time, developed with Scale AI utilizing the Scale Generative AI Platform, ingests fragmented exterior affected person information and organizes them chronologically, producing concise summaries and making the fabric searchable. According to Scale AI, the collaboration’s broader scope additionally includes automating detection of security occasions — equivalent to wrong-site surgical procedures or falls —hidden inside routine reporting noise, and serving to workers spend much less time on administrative duties. Data used within the collaboration remains inside Mayo Clinic’s HIPAA-compliant atmosphere.

Physicians spend much less time manually assembling a affected person’s historical past earlier than a go to and extra time on direct interplay, illustrated by two knowledge factors:

• Physicians spend much less time manually assembling a affected person’s historical past earlier than a go to and extra time on direct interplay.
• According to Scale AI, physicians have gained a median of 11 extra minutes with every affected person because the instruments launched, whereas sustaining an knowledgeable customary of care.
• Dr. Ryu individually described time financial savings of 5 to thirty minutes of prep per go to, relying on case complexity, and mentioned the device additionally helps him keep away from lacking particulars buried in a file that might have an effect on remedy or testing selections.

This is a deployed, in-use device, not a pilot — one in all roughly 150 AI fashions Mayo now has running throughout the system, giving it a observe report past a single division or trial.