Anumana Pulmonary Hypertension AI Tool Wins Clearance

Share

FDA has cleared an Anumana 510(k) for an artificial intelligence–based tool designed to detect early signs of pulmonary hypertension, a serious and often underdiagnosed cardiovascular condition. The company’s software-as-a-medical-device algorithm is intended to analyze standard 12-lead electrocardiograms (ECGs) to identify subtle indicators of pulmonary hypertension.

According to the company, the algorithm is the first of its kind cleared for use with routine ECGs, potentially enabling broader adoption across clinical settings without requiring specialized equipment or workflow changes.

Pulmonary hypertension is a progressive and potentially life-threatening disease affecting the lungs and the right side of the heart. Early diagnosis remains challenging due to nonspecific symptoms such as shortness of breath, often leading to delays of more than two years — delays that can worsen patient outcomes, according to the company.

Anumana’s tool is designed to augment traditional ECG interpretation by detecting patterns not readily visible to clinicians, helping flag patients who may benefit from further testing, such as echocardiography. “Pulmonary hypertension is often difficult to recognize until it has progressed significantly,” says Paul Friedman of the Mayo Clinic, who is also an advisor to the company. “This provides a practical way to identify patients earlier and guide next steps in care.”

The algorithm integrates with electronic health record systems and operates within a health system’s infrastructure, avoiding the need to transfer patient data externally, the company says.

Clinical validation data cited by Anumana show the tool achieved 73% sensitivity and 74.4% specificity in detecting pulmonary hypertension in a multi-center study of more than 21,000 patients presenting with shortness of breath. Additional analyses suggest the technology may identify a high proportion of patients with certain treatable subtypes of the disease before formal diagnosis.

Read more