AccurKardia receives patent for AI-ECG cardiac amyloidosis detection
AccurKardia has been granted a U.S. patent for a machine learning system that identifies cardiac amyloidosis from a standard 12-lead electrocardiogram, adding a new investigational tool to its AI-ECG pipeline. The company says the technology could help detect a disease that often goes undiagnosed until significant damage has occurred.
Patent covers major amyloidosis subtypes
The patent, numbered 12,620,488, covers detection of AL amyloidosis, wild-type transthyretin amyloidosis, and hereditary ATTR — the three main subtypes of the disease. Cardiac amyloidosis is a serious condition in which abnormal proteins build up in heart tissue, stiffening the muscle and eventually leading to heart failure. Currently, the algorithm is for research use only and has not been cleared by the FDA for clinical use in the United States.
Why an ECG-based approach matters
Juan C. Jimenez, cofounder and CEO of AccurKardia, said cardiac amyloidosis is one of the most underdiagnosed diseases in cardiology. Studies estimate that 13-15% of heart failure patients actually have amyloidosis, which often leads to ineffective initial treatment and late referrals once the correct diagnosis is made. Newer therapies can halt disease progression but do not reverse existing cardiac damage, so earlier detection is clinically important.
Related: 7 Surprising Facts About Healthy Food
“What makes an ECG-based approach compelling is that it works inside care that is already happening,” Jimenez said. Most patients eventually diagnosed with amyloidosis have had numerous ECGs over the years before anyone suspected the condition. An ECG is low-cost and non-invasive, so the screening signal could help flag which patients need further workup without adding a new test or procedure. The data, in many cases, already exists.
Explainable machine learning, not a black box
AccurKardia’s approach uses feature-based machine learning built on annotated ECG parameters, rather than a “black box” model. The company says this design supports clinical interpretability, regulatory transparency, and future deployment. The patent adds to a broader AI-ECG portfolio that already includes the FDA-cleared AccurECG 2.0 platform and two investigational applications — for aortic stenosis and hyperkalemia — that have received FDA Breakthrough Device designations. Jimenez said the cardiac amyloidosis patent is important because it establishes the company’s claim across all major subtypes, rather than just one form of the disease.
Validation work still needed before clinical use
The company has already completed large single-site validation studies. Next, it plans retrospective multi-site validation using large, diverse datasets that are representative by demographics, geography, and site type. “Before clinical deployment, we would want to see strong, consistent performance across those patient subgroups, not just in aggregate,” Jimenez said. He added that beyond clearance, the firm is committed to evidence development needed to demonstrate clinical efficacy and support reimbursement.
Related: The Future of Health Care: What to Expect in a Rapidly Evolving Landscape
Outside expert raises practical questions
Dr. Michelle Kittleson, a professor of medicine at the Smidt Heart Institute at Cedars-Sinai in Los Angeles, said several questions would need to be answered before the FDA could approve an AI-ECG tool for amyloidosis screening. “Before use in clinical practice, we would need to know: Is it accurate in real-world populations and which ones — all older adults, heart failure with preserved ejection fraction, etc.? Is it cost-effective?”
Kittleson also noted the importance of considering the downstream effect of initiating an amyloidosis diagnostic algorithm — calculating how many tests are completed versus how many cases are confirmed, and how the model compares to other available AI models. Clinicians need evidence that screening improves diagnosis and treatment timing without creating excessive false positives or unnecessary testing.
“Ultimately, what I would want to see as a clinician is that among patients with HFpEF or unexplained LVH, AI-ECG screening can identify cardiac amyloidosis earlier than usual care, with high sensitivity and an acceptable false-positive rate, leading to faster diagnosis and treatment initiation in a cost-effective fashion,” she said. The tool remains investigational, and AccurKardia has not announced a timeline for FDA submission.
