Caption AI can help sonographers produce quality diagnostic cardiac ultrasound scans regardless of their experience

Caption AI allows inexperienced sonographers to obtain quality images: study

February 24, 2021
by John R. Fischer, Senior Reporter
A multicenter study is touting Caption Health’s Caption AI platform for its ability to help inexperienced healthcare workers capture quality diagnostic cardiac ultrasound scans.

Caption AI is the first and only AI-guided medical imaging acquisition software to obtain FDA clearance for cardiac ultrasound exams at the point of care, according to the company.

“There is a shortage of professionals with this level of training. That shortage, in turn, limits access to imaging at the point of care. The intent of Caption AI is not to replace skilled sonographers, but to expand the number of healthcare professionals who can perform focused cardiac ultrasound at the point of care to enable timely diagnosis and management,” Charles Cadieu, CEO and co-founder of Caption Health, told HCB News.

Caption AI is integrated onto a full-service ultrasound system and utilizes Caption Guidance, which provides real-time guidance on how to position and manipulate the transducer on a patient’s body; and Caption Interpretation, which automatically calculates ejection fraction from any combination of parasternal long-axis (PLAX), apical 4-chamber (AP4), and apical 2-chamber (AP2) at the point of care.

The study, titled Utility of a Deep-Learning Algorithm to Guide Novices to Acquire Echocardiograms for Limited Diagnostic Use, was the basis for the February 2020 Authorization of Caption Guidance through the FDA’s De Novo pathway, and included 240 patients between 20 and 91, with 42% female; 17.6% black or African American; and 33% with a BMI of 30 or greater. Each underwent an ultrasound performed by a nurse with no prior experience and one from an experienced registered diagnostic cardiac sonographer. Experienced cardiologists assessed their diagnostic quality and made diagnostic assessments.

The platform produced quality diagnostic scans to assess left ventricular size and function in 98.8% of patients, right ventricular size and function in 92.5%, and the presence of pericardial effusion in 98.8%. Results showed at least 92.5% agreement between the nurse and sonographer scans in diagnostic assessment.

They also indicated that Caption Guidance performed well for patients with various cardiac pathologies that may be seen in clinical practice, as more than 90% of participants had cardiac abnormalities found in full echocardiograms performed within two weeks of the study. Findings were consistent across BMI, sex and race.

The platform may allow providers to better evaluate different pathologies in critical care, emergency departments and other settings, and could possibly identify them earlier with AI.

“We are currently focused on cardiac ultrasound, but we plan to expand our AI capabilities in the future — in November 2020, Caption Health received a $4.95M grant from the Bill & Melinda Gates Foundation to develop AI technology for lung ultrasound,” said Cadieu.

Caption AI is already being used in hospitals to assess and treat the cardiac-related effects of COVID-19, as the study shows that technologies like Caption Guidance may reduce the exposure of sonographers to the disease.

The findings were published in JAMA Cardiology.