AI-Driven Heart Disease Detection Software to Identify Hidden Cardiovascular Conditions

Anumana, Inc. (Cambridge, MA, USA) and Novartis Pharmaceuticals (Basel, Switzerland) have launched a multi-year strategic collaboration to drive the development and delivery of ECG AI algorithms to help physicians accelerate detection and intervention for patients with previously undetected life-threatening heart disease. The ECG AI algorithm is still in development and not yet FDA authorized for commercial clinical use. The collaboration will support Anumana’s efforts to implement AI-enabled diagnostic software that can detect signals from ECGs that humans cannot interpret. The collaboration will initially focus on cardiovascular diseases.

Working with its expert partners from Mayo Clinic (Rochester, MN, USA), Anumana will deploy novel solutions that use AI to analyze an ECG, a widely available, painless test that records the heart’s electrical signals to identify undiagnosed left ventricular dysfunction, or a weak heart pump, which can lead to heart failure. The AI will also screen for atherosclerotic cardiovascular disease, which can lead to heart attack and stroke. In addition, an evidence-based, digital point-of-care solution will be developed to guide in optimizing guideline-directed medical therapies with the aim to lower risk for potentially avoidable hospitalizations and cardiovascular death.

The Mayo Clinic Cardiology team pioneered the application of AI in cardiology and developed several algorithms based on millions of ECGs, including a low ejection fraction algorithm that received FDA Breakthrough Device Designation in 2019 and Emergency Use Authorization for COVID-19 in 2020. Further validating the technology, a recent study presented by Mayo Clinic used a modified version of Anumana’s 12-lead ECG algorithm to detect left ventricular dysfunction with single-lead ECGs in smartwatches. These algorithms are licensed to Anumana for development of clinical solutions and have been validated by over 30 peer-reviewed publications, including a first of its kind prospective clinical impact study on low ventricular ejection fraction that was published in Nature Medicine in 2021. These software solutions are currently in development with each algorithm as a candidate for marketing authorization through an FDA DeNovo request.

“Anumana technology is designed to help physicians identify patients who are at maximum risk of heart failure, long before they develop symptoms,” said Murali Aravamudan, CEO of Anumana. “Bringing together premier global organizations will allow us to expand access to best-in-class, AI-powered digital tools to benefit patients through earlier detection and intervention, when and where health care providers need it most.”

“Many heart diseases develop for years before signs and symptoms appear, but the first event may be life threatening,” said Paul Friedman, M.D., Chair of the Department of Cardiovascular Medicine at Mayo Clinic and Chair of Anumana’s Mayo Clinic Board of Advisors. “AI enables us to uncover hidden signals our bodies transmit to detect otherwise occult heart diseases, potentially years before symptoms appear. This collaboration has the potential to transform the use of an ubiquitous inexpensive test, the ECG, with the aim of democratizing disease detection and helping medical care teams to proactively manage heart disease ahead of time and prevent some clinical events from ever happening.”

“Cardiovascular disease is a widespread and multi-factorial disease and, in order to mitigate its impact, we must look beyond therapeutic innovation and re-imagine how we approach cardiovascular care,” said Victor Bulto, President, Novartis Innovative Medicines US. "Novartis is proud to collaborate with Anumana on innovative and data-driven solutions to better predict the risk of life-threatening heart disease, further driving forward our commitment to improving patient experiences and population health outcomes in this patient population.”

Related Links:
Anumana, Inc. 
Novartis Pharmaceuticals 
Mayo Clinic