Back to Results

Screening for Aortic Stenosis Using a Machine Learning Algorithm in Primary Care

Title Screening for Aortic Stenosis Using a Machine Learning Algorithm in Primary Care
Therapeutic Area Aortic Stenosis
Principal Investigator Benjamin Wessler, MD
Min Age 65 Years
Max Age 100 Years
Gender Any
Contact Mazhar Kadwalwala, MD, MBA


Aortic stenosis (AS) affects over 12.6 million adults and causes an estimated 102,700 deaths annually. When severe AS is untreated, the stenotic valve obstructs blood flow, ultimately leading to pressure overload on the left ventricle, congestive heart failure, and death. Our research is motivated by the observation that undiagnosed severe AS can be deadly. Without treatment, 50% of patients with symptomatic AS will die in 2 years. We previously used machine learning (ML) to develop a system for fully automated AS screening.8 This system offers a novel approach to identifying AS using limited echocardiogram datasets and point-of-care ultrasound device (POCUS). The current knowledge gap and challenge in using ML programs into clinical practice is the lack of clarity in the delivery science behind their implementation. There is a need to address how to integrate ML into clinical practice by designing, implementing and evaluating a system where these novel technologies can be used effectively and the system itself reliably replicated. Our study hopes to address this issue. 

Study Details

Inclusion Criteria

All comers between the age of 65 and 100 years of age to the Tufts Medical Center General Medical Associates Primary Care Clinic. 

Exclusion Criteria

1. Adults unable to consent (Cognitively impaired adults)

Study Requirements

1-2 echocardiography images taken by a trained medical professional at the start or the end of a primary care visit.