Artificial intelligence (AI) is increasingly being used for many healthcare applications, including the diagnosis and classification of heart disease. Various applications of AI offer promising tools for evolving healthcare by providing reliable, specific, and efficient methods for delivering high-quality care.
Cardiovascular disease (CVD) is a major medical issue and the leading cause of death in the US and around the globe. Accurately identifying and classifying heart-related pathology can often be challenging for doctors. Artificial intelligence is now being applied in various forms to help mitigate this struggle and improve the quality of care. AI breakthroughs in heart disease are capitalizing on the use of complex algorithms that lay out relevant steps for successful diagnosis and treatment.
For example, emerging AI programs can accurately identify and classify heart disease utilizing information from digital chest radiographs. This technology provides promise for expanding access to quality care that is delivered in a fraction of the time of manual methods.
Background on Heart Disease
CVD includes a range of pathologies and conditions that impact the heart and blood vessels. These conditions result in one in every five deaths in the US, making them the leading cause of death in the country. Globally, around 31% of all deaths are due to CVD.
Various types of heart disease fall under this umbrella. Some of the major risk factors for developing heart disease include high blood pressure (hypertension), high cholesterol, diabetes, smoking, obesity, and physical inactivity. These can contribute to the development of atherosclerosis, where fatty plaque builds up in the arteries, limiting blood flow.
When this happens, it increases your risk of a heart attack from coronary artery disease. The most common type of CVD impacting over 20 million adults over age 19 is coronary artery disease, which involves the arteries that supply blood to your heart becoming narrow or blocked, putting you at risk for chest pain and heart attack.
If the heart becomes damaged or weak and cannot pump enough blood to meet the body's needs, heart failure may develop. Congestive heart failure results in fatigue, shortness of breath, and swelling (edema) in the legs and feet. Another condition that impacts the heart occurs when your heartbeat is irregular (arrhythmia). This can cause palpitations, dizziness, and an increased risk of stroke.
Many challenges in heart disease diagnosis remain. Effective prevention and mitigation of the complications of CVD requires the identification of risk factors and early recognition and accurate diagnosis of disease. Physicians are overwhelmed with demands on their time and have to deal with an increasingly large array of data. Cardiovascular diseases often involve immense complexity, with the heart being a truly dynamic organ, making accurate diagnosis challenging and requiring specialized expertise to interpret testing.
Development of the AI Program
One solution that is being developed to address these challenges and improve the quality of care for patients is AI. AI includes systems that can learn and make decisions, mimicking human intelligence.
Developing AI for cardiology involves engaging in machine learning that can be programmed to detect patterns in complex data. Machine learning is able to develop systems that learn from retrospective data and can create resulting classifications, clustering, or regression models. This allows for the development of automated clinical decision systems that help with the interpretation of results and the implementation of personalized clinical decisions.
For example, working together, data scientists, cardiologists, and machine learning experts collaborated to develop a deep-learning model that can simultaneously detect valvular disease and cardiac functions from chest radiographs in order to estimate transthoracic echocardiography results.
Training AI in heart disease recognition of this type involved using datasets from multiple institutions to develop and evaluate this model. Chest radiographs were taken during the same time period as two-dimensional echocardiographic examinations and labeled using the echocardiography reports for valvular disease (mitral regurgitation, aortic stenosis, aortic regurgitation, mitral stenosis, tricuspid regurgitation, and pulmonary regurgitation) and severity of disease (none, mild, moderate, or severe).
In order to be as accurate and generalizable as possible and prevent overfitting, the data was collected from several different facilities. Using five-fold cross-validation, these data sets were used to train and validate a multi-label deep-learning model.
How the AI Program Works
Heart disease can be recognized, diagnosed, and classified by assessing various cardiac functions utilizing a variety of technologies and imaging techniques. Traditionally, echocardiograms are commonly used for evaluating valvular function and blood flow, which helps to diagnose heart failure. More recently, cardiac MRI imaging has also been used.
In this AI program, chest radiography images are used to identify and classify heart disease. They found that AI can recognize information from these radiological images that is difficult for humans to identify. Using this technology, they trained, validated, and externally tested a deep learning-based model to develop algorithms that can estimate echocardiography results, such as left ventricular ejection fraction and the presence of valvular heart disease, to classify heart disease based on the interpretation of chest radiographs. Choosing a deep learning method allowed them to quantify outcomes with complex features since it can be developed based on features from training data without requiring the manual definition of features.
With AI, the identification and classification of heart failure can be accomplished much faster and without the need for specialized technicians or equipment required to obtain this data from echocardiograms. In addition, the model only requires a small size of the input image (512 × 512 pixels), so it does not require a large capacity system to run the model and can be done using a regular office computer in daily clinical practice.
Accuracy and Efficacy
Within the field of cardiology, AI has shown promise in improving the diagnosis, treatment management, and risk prediction of cardiac diseases. It has also been successfully utilized to support more accurate and efficient analysis of medical images, such as echocardiograms, chest radiography, and cardiac MRI scans. AI is able to process vast amounts of patient data in short periods of time, allowing for the identification of patterns that predict the likelihood of developing certain heart diseases.
Machine learning algorithms are able to interpret clinical data from electrocardiograms (ECGs), echocardiograms, and other medical imaging to diagnose cardiovascular diseases, accurately predict outcomes, classify risk, and identify patterns in data. In many cases, AI can detect abnormalities that are too subtle for the human eye to detect, such as minute changes in the heart’s electrical activity that may predict future problems.
Such AI-based image analysis algorithms offer quick and accurate methods for identifying various heart diseases. Utilizing predictive models with AI allows for analyzing images and detecting abnormalities in an accurate and efficient manner.
Logistic regression algorithms are another way to utilize machine learning in cardiology. This technology uses inputs of patient risk factors such as age, gender, and blood pressure to predict the likelihood of developing a cardiovascular event such as a heart attack or heart failure.
AI also offers support when it comes to cardiology interventions and the management of heart disease. AI is facilitating robotic procedures such as percutaneous coronary intervention and electrophysiology for catheter ablation of cardiac arrhythmias. These AI models improve precision and outcomes by utilizing techniques like auto maneuvering to improve patient safety and reduce radiation exposure.
On an even more widely accessible level, AI algorithms are being used in conjunction with wearable technology like smartwatches. This approach can check your heartbeat rate and pattern to detect arrhythmias like atrial fibrillation and provide recommendations for management. It has also been shown to benefit patients with hypertension. AI-based systems can continuously monitor blood pressure via wearable technologies. In addition, machine learning algorithms are able to identify new hypertension genes for early diagnosis and the prevention of complications and can be integrated with omics-based technologies to help select the most appropriate drug for an individual patient.
Implications for Healthcare Providers and Patients
AI can support improved outcomes for both providers and patients. Physicians are stretched with many tasks, administrative burdens, and time constraints. With healthcare data expanding rapidly at almost 48% each year, physicians and healthcare systems are overwhelmed. There is frequent burnout in the field, resulting in serious medical errors, inefficient care, and higher costs.
AI technologies that can help to analyze this extensive data and facilitate the clinical decision process for physicians can relieve some of this burden on providers while offering improved length and quality of life for patients via early recognition and personalized treatment. There is immense potential for improving patient outcomes with AI. AI-supported technologies advance early and precise diagnosis.
This technology can ease the burden on physicians and reduce errors. AI can provide guidance and alerts regarding safe and effective care, such as possible drug interactions and allergies. AI can also be used to detect subtle changes and abnormalities that may go unnoticed for some time by traditional means of detection. Especially in a field like cardiology, precision, accuracy, and quick response times are critical for patients' quality of life and even survival.
To utilize this expanding technology in the most effective and ethical manner, conscious and considered approaches are needed. As the research and available AI technologies expand, cardiologists must become aware of the best ways to evaluate and reassess the AI-acquired information so that they can integrate this knowledge into their clinical decision-making. The best care comes when AI-provided information is thoughtfully integrated with physicians’ knowledge and experience to support clinical decision-making.
Challenges and Ethical Considerations
Despite the promise of these AI-supported technologies, various challenges and ethical considerations remain. Such rapidly expanding and evolving technologies are not yet at full maturity and require a substantial investment to build.
The use of such technology raises concerns about data privacy and security concerns. For example, how data is acquired, used, and stored must be established and protected. Solid security systems are needed for storing vulnerable private data with methods in place to prevent hacking. Data used to input into AI algorithms must also be of high quality and maintain standards of robustness, transparency, trust, and verifiability.
In addition, there must also be attention to the potential for exacerbating bias and discrimination. AI-driven technologies are built upon existing features and dynamics of the populations they analyze, which can lead to the continuation and even amplification of patterns of marginalization, inequalities, and discrimination.
There are also regulatory and legal issues that continue to be evaluated and threshed out. Currently, the World Health Organization (WHO) has established ethical guidance for the implementation of artificial intelligence in healthcare, and further official international legal, ethical, and methodical requirements for applying AI models to medicine are still in discussion and need to be developed.
In order to provide the most ethical, compassionate, and high-quality clinical care, effective communication between AI systems and healthcare providers is needed. Keeping in mind the irreplaceable role of humans in caregiving and the delicate physician-patient relationship will help mitigate the dehumanization of medicine.
Future Prospects and Integration into Healthcare Systems
AI is a new and rapidly evolving innovative field with many applications in the healthcare arena. As technology advances, integrating AI into healthcare systems has the potential to revolutionize diagnostic and treatment methods and patient care.
As we continue to build upon the increase in knowledge obtained from sequencing of the human genome and expanding molecular biology research, AI can complement the growth of precision medicine. This opens up innovative possibilities and facilitates the physician’s delivery of care that is personalized for each patient.
Many cardiovascular diseases are complex, involving interactions between gender, genetic, lifestyle, and environmental factors. AI can work in conjunction with precision medicine to more efficiently develop personalized management for heart disease patients by managing and utilizing this wealth of data.
Deep learning is facilitating our ability to translate the vastly expanding genomic knowledge obtained from DNA sequencing to demonstrate features like 3D protein configurations and transcription start sites and predict gene expression. Having this advanced interpretation of this vast pool of data allows for novel applications based on newly identified connections between genomic variation and disease presentation. This allows for advances in prognosis and therapeutic strategies.
For example, AI-based applications have provided a deeper understanding of varying presentations of heart failure and congenital heart disease that have allowed for precision phenotyping. This understanding is facilitating the development of innovative treatment strategies targeting the features of these individual phenotypes.
Integrating data about genomic factors, social determinants of health, behavioral risk factors, and clinical considerations can allow for improved personalized care. Studies show that combining information from electronic health records with genetic data can successfully predict cardiovascular disease.
As AI and deep learning technologies continue to advance, it is supporting innovation within the field of cardiology. More accurate and efficient means of diagnosing and managing cardiovascular diseases using AI will ultimately lead to more accessible care and better patient outcomes. AI has been shown to offer improved diagnostic accuracy, innovative risk prediction, and improved treatment outcomes to facilitate cardiac care.
Embracing AI in cardiac healthcare through the advancement of ongoing research and development can expand access to this type of specialized but often limited care. Continued collaboration between medical professionals, data scientists, and machine learning experts will allow future innovation in the field.
Lab Tests in This Article
Christie, J. (2023, January 6). A Functional Medicine Approach To Obesity And Weight Management. Rupa Health. https://www.rupahealth.com/post/an-integrative-approach-to-obesity
Cleveland Clinic. (2022). Understanding Heart Failure | Cleveland Clinic. Cleveland Clinic. https://my.clevelandclinic.org/health/diseases/17069-heart-failure-understanding-heart-failure
Cloyd, J. (2023a, April 7). Functional Medicine High Cholesterol Protocol. Rupa Health. https://www.rupahealth.com/post/functional-medicine-high-cholesterol-protocol
Cloyd, J. (2023b, April 10). A Functional Medicine Hypertension Protocol. Rupa Health. https://www.rupahealth.com/post/functional-medicine-hypertension-protocol
Cloyd, J. (2023c, May 17). The Role Of Nutrition And Dietary Supplements In Preventing And Managing Cardiovascular Disease. Rupa Health. https://www.rupahealth.com/post/the-role-of-nutrition-and-dietary-supplements-in-preventing-and-managing-cardiovascular-disease
Cloyd, J. (2023d, June 19). A Functional Medicine Post Stroke Protocol: Testing, Therapeutic Diet, and Integrative Therapy Options. Rupa Health. https://www.rupahealth.com/post/a-functional-medicine-post-stroke-protocol-testing-supplements-and-integrative-therapy-options
Cloyd, J. (2023e, October 9). Rhythms of the Heart: Demystifying Common Types of Heart Arrhythmia. Rupa Health. https://www.rupahealth.com/post/rhythms-of-the-heart-demystifying-common-types-of-heart-arrhythmia
DePorto, T. (2022, December 9). Worried About Heart Disease? Ask Your Provider for These 6 Specialty Labs at Your Next Appointment. Rupa Health. https://www.rupahealth.com/post/worried-about-heart-disease-ask-your-provider-for-these-6-specialty-labs-at-your-next-appointment
DePorto, T. (2023, January 10). Timeline: What Happens Inside Your Body When You Quit Smoking? Rupa Health. https://www.rupahealth.com/post/what-happens-to-our-bodies-when-we-quit-smoking-a-timeline
Hayes, S. M., Dupuis, M., & Murray, S. (2008). Issues and challenges in the assessment, diagnosis and treatment of cardiovascular risk factors: Assessing the needs of cardiologists. BMC Medical Education, 8(1). https://doi.org/10.1186/1472-6920-8-30
Johnson, K. B., Wei, W., Weeraratne, D., Frisse, M. E., Misulis, K., Rhee, K., Zhao, J., & Snowdon, J. L. (2020). Precision Medicine, AI, and the Future of Personalized Health Care. Clinical and Translational Science, 14(1).
Karatzia, L., Aung, N., & Aksentijevic, D. (2022). Artificial intelligence in cardiology: Hope for the future and power for the present. Frontiers in Cardiovascular Medicine, 9. https://doi.org/10.3389/fcvm.2022.945726
Khakham, C. (2023, April 6). Understanding Your Risk of Cardiovascular Disease With Functional Medicine Labs. Rupa Health. https://www.rupahealth.com/post/understanding-your-risk-of-cardiovascular-disease-with-functional-medicine-labs
Kulkarni, P., Mahadevappa, M., & Chilakamarri, S. (2021). The emergence of artificial intelligence in cardiology: current and future applications. Current Cardiology Reviews, 17. https://doi.org/10.2174/1573403x17666211119102220
Ledziński, Ł., & Grześk, G. (2023). Artificial Intelligence Technologies in Cardiology. Journal of Cardiovascular Development and Disease, 10(5), 202. https://doi.org/10.3390/jcdd10050202
Mathur, P., Srivastava, S., Xu, X., & Mehta, J. L. (2020). Artificial Intelligence, Machine Learning, and Cardiovascular Disease. Clinical Medicine Insights: Cardiology, 14, 117954682092740. https://doi.org/10.1177/1179546820927404
Stanford Medicine. (2020). The Rise of the Data-Driven Physician. https://med.stanford.edu/content/dam/sm/school/documents/Health-Trends-Report/Stanford%20Medicine%20Health%20Trends%20Report%202020.pdf
Sweetnich, J. (2023, April 25). Complementary and Integrative Medicine Approaches to Type 2 Diabetes Management. Rupa Health. https://www.rupahealth.com/post/complementary-and-integrative-medicine-approaches-to-type-2-diabetes-management
Ueda, D., Matsumoto, T., Ehara, S., Yamamoto, A., Walston, S. L., Ito, A., Shimono, T., Shiba, M., Takeshita, T., Fukuda, D., & Miki, Y. (2023). Artificial intelligence-based model to classify cardiac functions from chest radiographs: a multi-institutional, retrospective model development and validation study. The Lancet. Digital Health, 5(8), e525–e533. https://doi.org/10.1016/S2589-7500(23)00107-3
Visco, V., Izzo, C., Mancusi, C., Rispoli, A., Tedeschi, M., Virtuoso, N., Giano, A., Gioia, R., Melfi, A., Serio, B., Rusciano, M. R., Di Pietro, P., Bramanti, A., Galasso, G., D’Angelo, G., Carrizzo, A., Vecchione, C., & Ciccarelli, M. (2023). Artificial Intelligence in Hypertension Management: An Ace up Your Sleeve. Journal of Cardiovascular Development and Disease, 10(2), 74. https://doi.org/10.3390/jcdd10020074
WHO. (n.d.). Ethics and governance of artificial intelligence for health. Www.who.int. https://www.who.int/publications-detail-redirect/9789240029200
Yoshimura, H. (2023, November 7). The Remarkable Power of Exercise on Our Health: A Comprehensive Overview. Rupa Health. https://www.rupahealth.com/post/the-remarkable-power-of-exercise-on-our-health-a-comprehensive-overview
Zhao, J., Feng, Q., Wu, P., Lupu, R. A., Wilke, R. A., Wells, Q. S., Denny, J. C., & Wei, W.-Q. (2019). Learning from Longitudinal Data in Electronic Health Record and Genetic Data to Improve Cardiovascular Event Prediction. Scientific Reports, 9(1), 717. https://doi.org/10.1038/s41598-018-36745-x