AI outperforms medical professionals in detecting cardiovascular disease events
Two preliminary research studies, presented at the 2023 American Heart Association’s Scientific Sessions (11–12 November, PA, USA), have showcased AI’s incredible ability to predict and detect cardiovascular disease events.
With computational methods for health and disease analysis and prediction growing in complexity, these studies dive into what such innovative tools offer for real-world applications.
Study number one: a stethoscopic symphony
The first study investigated the performance of an AI-driven digital stethoscope in identifying potential heart valve diseases compared to the performance of a medical professional using a traditional stethoscope. The medical professionals were blinded to both the echocardiogram and AI results.
Involving 369 participants aged 50 and above from three primary care clinics in the US, the AI-driven stethoscope far outshone the traditional stethoscopes used by primary care professionals, detecting a total of 94.1% of valvular heart disease cases compared to 41.2% detected by primary care professionals. Furthermore, the AI-driven stethoscope identified 22 people with previously undiagnosed moderate heart disease, while medical professionals using traditional stethoscopes identified only 8 of these participants.
”The implications of undiagnosed or late diagnosis of valvular heart disease are dire and pose a significant cost to our health care system,” explained Moshe Rancier, Senior Medical Director of Mass General Brigham Community Physicians in Lawrence (MA, USA).
“This study demonstrates that health care professionals can screen patients for valvular heart disease more effectively and quickly using a digital stethoscope paired with high-performing AI that could detect cardiac murmurs associated with significant valvular heart disease.”
The authors do note some study limitations. These include the small sample size and risk of inflated false positives due to the reduced specificity of AI-driven heart disease diagnoses: 95.5% specificity for medical professionals using traditional stethoscopes compared to 84.5% for AI-driven digital stethoscopes.
Despite these constraints, Dan Roden, Senior VP for personalized medicine at Vanderbilt University Medical Center (TN, USA), envisions a transformative future for this emerging technology: “using an AI-enabled stethoscope and perhaps combining it with other imaging modalities, like an AI-enabled echocardiogram built into your stethoscope, will likely help to transform CVD care.”
Study number two: a retinal revolution
A second study presented at the Scientific Sessions used data from the UK Biobank to evaluate the effectiveness of deep learning-based retinal imaging for predicting cardiovascular disease events, including heart attack, ischemic stroke and transient ischemic attack.
The deep learning algorithm was used to categorize retinal images of 1,101 individuals with prediabetes or Type 2 diabetes into three groups varying in cardiovascular disease risk: low, moderate and high. The number of cardiovascular disease events among the participants was then recorded over a median duration of 11 years.
The findings unveiled a correlation between retinal images and cardiovascular events, with participants classified as high risk being 88% more likely to experience a cardiovascular event compared to participants classified as low risk. A total of 8.2% of individuals in the low-risk group, 15.2% in the moderate-risk group and 18.5% in the high-risk group experienced a cardiovascular disease event during the 11-year period.
Chan Joo Lee, associate professor at Yonsei University (Seoul, South Korea), stated that these results “could lead to early interventions and better management of these patient groups, ultimately reducing the incidence of heart disease-related complications.”
93% of the study participants in the UK Biobank are of European ancestry, which raises concerns over any meaningful generalizability for the wider population. Although both these studies present promising early detection tools for cardiovascular disease events, it is vital to obtain results from diverse datasets.
As AI proceeds to pave its way in cardiac care, these studies open the first doors to a future where predictive diagnostics and early interventions are directed by the algorithms of innovation.
