On April 8, 2026, the FDA cleared an algorithm that claims to spot a disease most doctors never catch until it is advanced: cardiac amyloidosis, a condition where misfolded proteins build up inside heart muscle until the organ can no longer pump properly. The clearance went to Anumana, a company co-founded by nference and the Mayo Clinic, for what regulators and the company both describe as the first and only FDA-cleared algorithm that can flag the condition using nothing more than a standard 12-lead electrocardiogram — the same test performed millions of times a year in routine cardiology visits.
A disease that hides in plain sight
Cardiac amyloidosis is notoriously underdiagnosed because its symptoms — shortness of breath, fatigue, swelling — mimic far more common heart conditions like hypertension or ordinary heart failure. Patients often see multiple cardiologists over years before a biopsy or specialized imaging test finally identifies amyloid deposits, by which point the disease has frequently caused irreversible damage. Cardiologists have long suspected that ECG waveforms contain subtle signatures of the condition, but those patterns are too faint and complex for a human reader to reliably detect by eye.
How the algorithm works and how well it performed
Anumana’s ECG-AI software runs on ECGs already collected during routine care, meaning it requires no new test, no additional cost to the patient, and no change to existing clinical workflow — it simply reprocesses data clinicians already have. In a validation study spanning 25,525 patients across four U.S. health systems, the algorithm achieved 78.9% sensitivity and 91.2% specificity for detecting both major forms of the disease, ATTR and AL cardiac amyloidosis, among patients with signs, symptoms, or comorbidities suggestive of the condition. Those numbers mean the tool will still miss roughly one in five true cases and will falsely flag a small share of healthy patients, but proponents argue that even an imperfect screening signal is valuable for a disease currently caught mostly by chance or by a specialist’s hunch.
The Mayo Clinic pedigree
The algorithm was originally developed inside Mayo Clinic and later validated independently before Anumana brought it to market. Dr. Martha Grogan of Mayo Clinic, who worked on the underlying research, said a tool that helps clinicians recognize suspicion of amyloidosis from a routine ECG could support earlier diagnosis — a modest claim by design, since the FDA clearance positions the software as a decision-support aid rather than a diagnostic replacement for biopsy or cardiac MRI. Dr. Angela Dispenzieri, also of Mayo Clinic, emphasized that the multicenter validation size gives the tool credibility beyond a single-institution proof of concept. Anumana CEO Maulik Nanavaty framed the clearance as part of a broader cardiovascular AI portfolio the company has been building, following earlier FDA clearances for algorithms detecting low ejection fraction and pulmonary hypertension from ECGs.
Skepticism about screening at scale
Not everyone in cardiology is convinced that ECG-based screening for a rare condition will translate into better outcomes at a population level. Cardiac amyloidosis, while underdiagnosed, is still relatively uncommon, and screening tools with a sub-80% sensitivity risk generating false reassurance in patients the algorithm clears, while a nontrivial false-positive rate could send otherwise healthy patients toward invasive confirmatory testing, adding cost and anxiety without benefit. Anumana’s answer is that clinicians retain full discretion over what to do with a flag — the tool doesn’t diagnose amyloidosis, it nominates patients for further workup, leaving the ultimate decision, and liability, with the treating physician.
Where this fits in the wider AI-cardiology boom
The clearance adds to a rapidly filling category: cardiology now has more than 200 FDA-cleared AI algorithms, spanning arrhythmia detection, structural heart disease screening, and now rare protein-deposition disorders. The amyloidosis algorithm previously received FDA Breakthrough Device Designation and was selected for the agency’s Total Product Life Cycle Advisory Program, both signals that regulators saw enough clinical promise to fast-track its review. For patients, the practical upshot is that a test many of them already get during a routine cardiology visit could soon come with an invisible second read attached, one that costs nothing extra and simply flags a small subset of people for a conversation their doctor might not otherwise have thought to have.
What happens next
Anumana has not disclosed pricing or rollout timelines beyond the clearance itself, but the company’s pattern with its previous ECG-AI products suggests health systems will be able to license the software for integration into existing ECG machines and electronic health record systems rather than needing new hardware. The bigger question industry watchers are asking is whether payers will reimburse for AI-flagged amyloidosis workups at scale, and whether a 78.9% sensitivity rate holds up once the tool leaves the four health systems that validated it and moves into more diverse, real-world patient populations.
