Artificial intelligence (AI) is making significant strides in the field of healthcare, with two groundbreaking studies set to be presented at the American Heart Association’s Scientific Sessions 2023. These studies showcase the potential of AI in detecting heart valvular disease and predicting cardiovascular disease (CVD) risk. Led by prominent researchers, these studies offer promising results that could revolutionize the way cardiovascular diseases are diagnosed and managed.
A pioneering study conducted at three primary care clinics in the United States has demonstrated the power of AI in identifying undiagnosed valvular heart disease. Traditional methods of detecting heart valve issues often rely on the skills of healthcare professionals using standard stethoscopes. However, this study aimed to compare the performance of human clinicians with that of an AI program utilizing sound data from a digital stethoscope.
The study: Key findings
369 adults aged 50 and older, with no prior heart valve disease diagnosis, participated in the study. Both human clinicians and AI-based digital stethoscopes examined the participants.
None of the healthcare professionals were aware of the AI results or echocardiogram outcomes, ensuring a blind study.
The AI method using the digital stethoscope detected an impressive 94.1% of valvular heart disease cases, compared to only 41.2% detected by human clinicians with standard stethoscopes. AI identified 22 previously undiagnosed cases of moderate-or-greater heart valve disease, while human clinicians identified only 8.
Dr. Moshe Rancier, the senior medical director of Mass General Brigham Community Physicians in Lawrence, Massachusetts, emphasized the significance of this study’s results. He noted that undiagnosed or late-diagnosed valvular heart disease can have dire consequences and impose a substantial burden on the healthcare system. The AI-enabled digital stethoscope offers a more efficient and effective screening tool for valvular heart disease, potentially leading to earlier diagnoses and improved patient outcomes.
Study limitations and future directions
While the study demonstrated the AI method’s high sensitivity in detecting heart valve disease, human clinicians using standard stethoscopes were more specific in their diagnoses, reducing the potential for false positives. Further research will be needed to assess clinical outcomes and additional diagnostic tests and treatments in a larger and more diverse patient population.
The second study, based on data from the UK Biobank, explored the use of AI in predicting the risk of cardiovascular disease events through the analysis of retinal images. This innovative approach relies on deep learning algorithms to assess the back-of-the-eye images and determine individuals’ risk for cardiovascular events, such as heart attacks and strokes.
The study included 1,101 individuals with prediabetes or type 2 diabetes.
Researchers categorized participants into low-risk, moderate-risk, and high-risk groups based on AI analysis of retinal images. Over a median period of 11 years, 8.2% of those in the low-risk group, 15.2% in the moderate-risk group, and 18.5% in the high-risk group experienced cardiovascular disease events.
After accounting for various risk factors, including age, gender, medication use, and smoking history, individuals in the moderate-risk group were 57% more likely to experience a cardiovascular event than those in the low-risk group. High-risk individuals were 88% more likely.
Dr. Chan Joo Lee, the study’s lead author and an associate professor at Yonsei University in Seoul, Korea, highlighted the potential of AI analysis of retinal imaging in early detection and risk management of heart disease in high-risk groups. This approach could lead to timely interventions and better outcomes for patients with prediabetes and type 2 diabetes.
While the study leveraged a large dataset from the UK Biobank, the population was primarily of European ancestry, raising questions about the applicability of these findings to more diverse racial and ethnic groups. Further research will be needed to validate the effectiveness of retinal imaging compared to other cardiovascular risk assessment tools.
Implications and future directions
The promising results of these two studies underscore the transformative potential of AI in the field of cardiovascular disease diagnosis and risk assessment. Dr. Dan Roden, a leading expert in personalized medicine at Vanderbilt University Medical Center, expressed his optimism regarding the future of AI in cardiovascular care. He suggested that combining AI-enabled stethoscopes with other imaging modalities could further enhance the detection of valvular and other heart diseases, revolutionizing cardiovascular care.
However, it’s crucial to address the limitations of these AI tools, including their accuracy and applicability to diverse populations. Future research and validation studies will be essential to ensure the reliability and effectiveness of these AI-based approaches in clinical practice.