Top Things to Know: Use of Artificial Intelligence in Improving Outcomes in Heart Disease
Published: February 28, 2024
- Artificial intelligence (AI) and machine learning (ML) are rapidly being adapted to improve quality of care in all areas of medicine. Currently, there are few widespread applications of AI to cardiovascular and stroke care. AI has the potential to improve patient outcomes with cardiovascular disease in a wide range of areas.
- Imaging is necessary to detect cardiovascular disease and stroke. Becoming an expert at image interpretation requires years of experience, a process that could be streamlined by AI. AI already has several applications in the field of imaging, including:
- Echocardiography (ECG).
- Cardiac Computed Tomography (CT).
- Cardiac Magnetic Resonance Imaging (MRI).
- Cardiac treatment planning – AI can be used to facilitate structural interventions.
- Stroke diagnosis and treatment planning – AI can be used to detect stroke by detecting intracranial hemorrhage on contrast CT.
- AI has advanced in electrocardiography, potentially streamlining ECG interpretation, while improving the detection of ECG results that humans cannot detect and predict phenotyping results. Challenges in applying AI to electrocardiography include limited availability of ECG data for algorithm training.
- AI applications for in-hospital monitoring can improve the information provided by bedside monitors. AI can reduce the rate of false alarms, detect clinical deterioration and conditions such as sepsis, hypotension, atrial fibrillation, and drug related pro-arrhythmia, anticipate cardiac arrest, and provide information about patient risks pre-surgery.
- Implantable and wearable devices have the potential to harness the power of AI. These can be used for the detection of atrial fibrillation and monitoring measures of heart health such as blood pressure.
- DNA sequencing technologies allow for AI applications. AI can use genome wide association studies (GWAS) to identify variants that may carry risk for CV disease. AI also explains more variance in phenotypes than a standard polygenic risk score and can be used to assess the detrimental nature of genetic variants. AI may also expand upon ancestry characterization as well as predict genetic conditions from phenotypes.
- AI has application for Electronic Health Records (EHR) analysis. AI can be used to predict many things using EHR, including in-hospital mortality, cardiovascular outcomes, and specific cardiovascular disease. Additionally, AI can be used for disease classification.
- There are several challenges to the broad adoption of AI in cardiovascular medicine, including the need for larger datasets, presence of built in disparities in algorithms which may reflect bias, accountability and reliability, cybersecurity, and the ethics of using patient data.
- To implement AI in cardiovascular medicine, AI should be used to augment clinical decisions rather than replacing them altogether. Additionally, there are considerations with the development of AI in cardiovascular medicine, including accounting for potential bias within AI technologies and equitable distribution of benefits.
Citation
Armoundas AA, Narayan SM, Arnett DK, Spector-Bagdady K, Bennett DA, Celi LA, Friedman PA, Gollob MH, Hall JL, Kwitek AE, Lett E, Menon BK, Sheehan KA, Al-Zaiti SS; on behalf of the American Heart Association Institute for Precision Cardiovascular Medicine; Council on Cardiovascular and Stroke Nursing; Council on Lifelong Congenital Heart Disease and Heart Health in the Young; Council on Cardiovascular Radiology and Intervention; Council on Hypertension; Council on the Kidney in Cardiovascular Disease; and Stroke Council. Use of artificialintelligence in improving outcomes in heart disease: a scientific statement from the American Heart Association. Circulation. Published online February 28, 2024. doi: 10.1161/CIR.0000000000001201