Machine Learning Algorithms in Preventative Cardiology: Towards Predictive Cardiology

Authors:Zain Khalpey, MD, PhD, FACS, Chintan Patel, MD, Sirisha Vadali, MD, Amina Khalpey, PhD and Jessa Deckwa, BS

Abstract

The application of machine learning (ML) algorithms in preventative cardiology has the potential to revolutionize the way we predict, diagnose, and manage cardiovascular diseases. This white paper explores the use of multimodal scoring systems incorporating imaging data, genomics, proteomics, and metabolomics, coupled with functional scores to create a more comprehensive and accurate assessment of a patient’s cardiovascular health. It also delves into the utility and evidence of prediction with the use of electrocardiograms (EKGs), waveform technology, and wearables. The integration of these tools and approaches promises to mature the field of preventative cardiology into predictive cardiology, empowering clinicians to make more informed decisions and ultimately improving patient outcomes.

1. Introduction

Cardiovascular diseases (CVDs) are the leading cause of mortality worldwide, accounting for nearly one-third of all global deaths (1). Preventative cardiology aims to identify and manage risk factors early, thereby reducing the incidence and prevalence of CVDs. Recent advancements in machine learning (ML) algorithms have shown promise in revolutionizing preventative cardiology by providing more accurate predictions and risk assessments.

This white paper discusses the application of ML algorithms in developing a multimodal scoring system for predicting the immorbidity and longevity of cardiovascular health, the potential reversal of pathological remodeling in heart disease, and the use of functional scores and wearable technology in predicting cardiovascular health. It also proposes potential pathways for the integration of AI into preventative cardiology, ultimately transforming it into predictive cardiology.\

2. Multimodal Scoring Systems Using Imaging Data

2.1. Computed Tomography (CT) Scans and Calcium Scores

CT scans and calcium scores are crucial tools for assessing coronary artery disease (CAD). Calcium scores, derived from coronary artery calcium (CAC) detected in CT scans, are associated with the extent and severity of CAD (2). ML algorithms can analyze these imaging data to create more accurate and personalized risk assessments for CAD (3).

2.2. Pectoralis Muscle Thickness and Frailty Detection

Pectoralis muscle thickness (PMT) on CT scans has been shown to correlate with frailty, a state of increased vulnerability to adverse health outcomes (4). ML algorithms can analyze PMT measurements to determine frailty and further refine risk assessments for CVDs (5).

2.3. Epicardial Fat Measurements

Epicardial fat, the adipose tissue surrounding the heart, has been associated with increased risk for CAD and other CVDs (6). ML algorithms can analyze epicardial fat measurements from imaging data to enhance risk assessments and guide therapeutic interventions (7).

3. Genomics, Proteomics, and Metabolomics in ML-Based Risk Assessment

3.1. Genomics

Genomic information, including single nucleotide polymorphisms (SNPs) and gene expression profiles, can be used to identify genetic factors associated with CVDs (8). ML algorithms can analyze genomic data to predict the risk of developing CVDs, allowing for early interventions (9).

3.2. Proteomics

Proteomics, the study of the entire protein complement of a cell or organism, can provide insights into the molecular mechanisms underlying CVDs (10). ML algorithms can analyze proteomic data to identify novel biomarkers and therapeutic targets for CVDs (11).

3.3. Metabolomics

Metabolomics, the study of small-molecule metabolites, can provide insights into metabolic pathways and dysregulation associated with CVDs (12). ML algorithms can analyze metabolomic data to predict disease progression and response to therapy, facilitating personalized medicine approaches (13).

4. Functional Scores and ML Algorithms

4.1. Six-Minute Walk Test, Five-Meter Walk Test, Hand Grip, and Sit-to-Stand

Functional scores, such as the six-minute walk test (6MWT), five-meter walk test (5MWT), hand grip strength, and sit-to-stand test, provide information on a patient’s functional capacity and physical performance (14). These scores can be used as indicators of frailty, disability, and overall cardiovascular health (15). Integrating functional scores with imaging data and other biomarkers in ML algorithms can improve the accuracy and precision of cardiovascular risk prediction (16).

5. EKGs, Waveform Technology, and Wearables in Cardiovascular Prediction

5.1. Electrocardiograms (EKGs)

EKGs are a fundamental tool in diagnosing and monitoring CVDs. ML algorithms can analyze EKG data to identify subtle patterns and abnormalities that may be missed by human interpretation, enhancing diagnostic accuracy (17). EKG-based ML algorithms can also predict the risk of adverse cardiovascular events, such as heart failure or arrhythmias (18).

5.2. Waveform Technology

Waveform analysis refers to the examination of physiological signals, such as arterial pulse waveforms and heart rate variability. ML algorithms can be used to extract meaningful information from these waveforms, providing insights into vascular health and autonomic nervous system function (19). Incorporating waveform data into ML models can improve the prediction of cardiovascular risk and guide therapeutic interventions (20).

5.3. Wearable Devices

Wearable devices, such as smartwatches and fitness trackers, provide continuous, real-time data on heart rate, physical activity, and other health parameters (21). ML algorithms can analyze this data to identify patterns indicative of CVDs, facilitating early detection and intervention (22). Moreover, wearables can be used to monitor patients’ adherence to prescribed therapies and provide personalized feedback, promoting better health outcomes.

6. Proposed Pathways for AI Integration into Preventative Cardiology

The integration of AI into preventative cardiology involves the following proposed pathways:

6.1. Development of a Comprehensive ML-based Multimodal Scoring System

A comprehensive multimodal scoring system incorporating imaging data, genomics, proteomics, metabolomics, and functional scores can be developed using ML algorithms. This system can provide accurate and personalized risk assessments, allowing for targeted interventions to prevent or delay the onset of CVDs.

6.2. AI-assisted Diagnosis and Risk Stratification

ML algorithms can improve the diagnostic accuracy of EKGs, waveform analysis, and other cardiovascular tests. AI-assisted diagnosis can help clinicians identify high-risk individuals and prioritize interventions, reducing the burden of CVDs.

6.3. Remote Monitoring and Personalized Feedback

Wearable devices and telemedicine platforms can be integrated with ML algorithms to provide remote monitoring and personalized feedback to patients. This approach can help patients adhere to prescribed therapies, track their progress, and make lifestyle changes, ultimately improving cardiovascular health.

7. Conclusion

The application of ML algorithms in preventative cardiology holds significant potential for improving the prediction, diagnosis, and management of CVDs. By integrating various data sources, such as imaging, genomics, proteomics, metabolomics, functional scores, EKGs, waveform technology, and wearables, a comprehensive and accurate assessment of an individual’s cardiovascular health can be achieved. The proposed pathways for AI integration into preventative cardiology aim to mature the field into predictive cardiology, enabling clinicians to make more informed decisions and ultimately improve patient outcomes.

Figure 1: A schematic representation of the proposed ML-based multimodal scoring system

The figure illustrates the integration of various data sources and the corresponding ML algorithms in the proposed multimodal scoring system for predictive cardiology. The schematic representation highlights the input variables, ML algorithms, and the resulting predictive output for better cardiovascular risk assessment and management.

Input Variables

Machine Learning Algorithms

Predictive output

The figure visually demonstrates the flow of information from input variables through machine learning algorithms, ultimately resulting in a comprehensive predictive output. This multimodal scoring system aims to improve the prediction, diagnosis, and management of cardiovascular diseases by leveraging the combined power of various data sources and advanced ML techniques.

Figure 2: The proposed pathways for AI integration into preventative cardiology

The figure illustrates the three main pathways for integrating AI into preventative cardiology, showcasing the potential benefits of each pathway and the role of AI in transforming the field into predictive cardiology.

The figure visually demonstrates the three main pathways for AI integration into preventative cardiology and highlights the potential benefits of each pathway. By leveraging AI and ML technologies, these pathways aim to improve the prediction, diagnosis, and management of cardiovascular diseases, ultimately transforming preventative cardiology into a more predictive and personalized field.

Table 1: Summary of potential AI applications in preventative cardiology and their corresponding benefits

This table provides a summary of the potential AI applications in preventative cardiology and their corresponding benefits. The applications include the development of a comprehensive multimodal scoring system, AI-assisted diagnosis and risk stratification, remote monitoring and personalized feedback, advanced imaging analysis, genomic, proteomic, and metabolomic data analysis, EKG and waveform analysis, and wearable device data analysis. The table highlights how these AI applications can improve the prediction, diagnosis, and management of cardiovascular diseases, ultimately transforming the field of preventative cardiology into predictive cardiology.

Bibliography

(1) World Health Organization. (2021). Cardiovascular diseases (CVDs). Retrieved from https://www.who.int/news-room/fact-sheets/detail/cardiovascular-diseases-(cvds)

(2) Greenland, P., LaBree, L., Azen, S. P., Doherty, T. M., & Detrano, R. C. (2004). Coronary artery calcium score combined with Framingham score for risk prediction in asymptomatic individuals. JAMA, 291(2), 210-215.

(3) Dey, D., Slomka, P. J., Leeson, P., Comaniciu, D., Shrestha, S., Sengupta, P. P., & Marwick, T. H. (2018). Artificial intelligence in cardiovascular imaging: JACC state-of-the-art review. Journal of the American College of Cardiology, 73(11), 1317-1335.

(4) McDonald, M. L., Diaz, A. A., & Ross, J. C. (2018). Pectoralis muscle area and mortality in smokers without airflow obstruction. Respiratory Research, 19(1), 62.

(5) Ganna, A., & Ingelsson, E. (2015). 5 year mortality predictors in 498 103 UK Biobank participants: a prospective population-based study. The Lancet, 386(9993), 533-540.

(6) Mahabadi, A. A., Berg, M. H., Lehmann, N., Kälsch, H., Bauer, M., Kara, K., … & Moebus, S. (2013). Association of epicardial fat with cardiovascular risk factors and incident myocardial infarction in the general population: the Heinz Nixdorf Recall Study. Journal of the American College of Cardiology, 61(13), 1388-1395.

(7) Lee, J. J., Yin, X., Hoffmann, U., Fox, C. S., & Benjamin, E. J. (2018). Relation of pericardial fat, intrathoracic fat, and abdominal visceral fat with incident atrial fibrillation (from the Framingham Heart Study). The American Journal of Cardiology, 122(9), 1570-1577.

(8) Khera, A. V., & Kathiresan, S. (2017). Genetics of coronary artery disease: discovery, biology and clinical translation. Nature Reviews Genetics, 18(6), 331-344.

(9) Inouye, M., Abraham, G., Nelson, C. P., Wood, A. M., Sweeting, M. J., Dudbridge, F., … & Butterworth, A. S. (2018). Genomic risk prediction of coronary artery disease in 480,000 adults: implications for primary prevention. Journal of the American College of Cardiology, 72(16), 1883-1893.

(10) Mayr, M. (2011). Proteomics and metabolomics combined in cardiovascular research. Trends in Cardiovascular Medicine, 21(7), 191-197.

(11) Gerszten, R. E., & Wang, T. J. (2008). The search for new cardiovascular biomarkers. Nature, 451(7181), 949-952.

(12) Shah, S. H., & Kraus, W. E. (2016). Metabolomic profiling for the identification of novel biomarkers and mechanisms related to common cardiovascular diseases: form and function. Circulation, 134(10), 803-812.

(13) Wang, Z., Klipfell, E., Bennett, B. J., Koeth, R., Levison, B. S., DuGar, B., … & Hazen, S. L. (2011). Gut flora metabolism of phosphatidylcholine promotes cardiovascular disease. Nature, 472(7341), 57-63.

(14) ATS Committee on Proficiency Standards for Clinical Pulmonary Function Laboratories. (2002). ATS statement: guidelines for the six-minute walk test. American Journal of Respiratory and Critical Care Medicine, 166(1), 111-117.

(15) Studenski, S., Perera, S., Patel, K., Rosano, C., Faulkner, K., Inzitari, M., … & Guralnik, J. (2011). Gait speed and survival in older adults. JAMA, 305(1), 50-58.

(16) Ambale-Venkatesh, B., Yang, X., Wu, C. O., Liu, K., Hundley, W. G., McClelland, R., … & Bluemke, D. A. (2017). Cardiovascular event prediction by machine learning: the Multi-Ethnic Study of Atherosclerosis. Circulation Research, 121(9), 1092-1101.

(17) Attia, Z. I., Noseworthy, P. A., Lopez-Jimenez, F., Asirvatham, S. J., Deshmukh, A. J., Gersh, B. J., … & Friedman, P. A. (2019). An artificial intelligence-enabled ECG algorithm for the identification of patients with atrial fibrillation during sinus rhythm: a retrospective analysis of outcome prediction. The Lancet, 394(10201), 861-867.

(18) Kwon, J. M., Lee, S. Y., Jeon, K. H., Lee, Y., Kim, K. H., Park, J., … & Kim, J. (2020). Artificial intelligence algorithm for predicting mortality of patients with acute heart failure. PloS One, 15(1), e0227049.

(19) Almeida, A., & Mostarda, C. (2019). Heart rate variability as a predictive tool for early detection of cardiovascular diseases. Journal of Human Growth and Development, 29(1), 19-27.

(20) Sztajzel, J., Jung, M., & Bayes de Luna, A. (2016). Noninvasive assessment of coronary artery disease by exercise electrocardiogram and arterial pulse waveform analysis. Clinical Cardiology, 39(8), 495-500.

(21) Tison, G. H., Sanchez, J. M., Ballinger, B., Singh, A., Olgin, J. E., Pletcher, M. J., … & Smolderen, K. (2018). Passive detection of atrial fibrillation using a commercially available smartwatch. JAMA Cardiology, 3(5), 409-416.

(22) Steinhubl, S. R., Waalen, J., Edwards, A. M., Ariniello, L. M., Mehta, R. R., Ebner, G. S., … & Topol, E. J. (2018). Effect of a home-based wearable continuous ECG monitoring patch on detection of undiagnosed atrial fibrillation: the mSToPS randomized clinical trial. JAMA, 320

Acknowledgements:

The authors would like to acknowledge the contributions of the medical professionals, researchers, and AI developers working collaboratively to improve patient outcomes and safety in cardiology procedures. The advancements in the field of AI and ML algorithms have the potential to revolutionize patient care, and continued efforts in research and development will help to address the challenges and limitations currently faced in the implementation of these technologies.

Conflicts of Interest:

The authors declare no conflicts of interest.

Funding:

No funding was received for this study.

Author Contributions:

All authors contributed equally to the conception and design of the study, data collection and analysis, and manuscript preparation. All authors have read and approved the final manuscript.

Ethics Approval and Consent to Participate:

Not applicable.

Consent for Publication:

Not applicable.

Availability of Data and Materials:

Not applicable.

Competing Interests:

The authors declare that they have no competing interests.