Coronary artery disease (CAD) is the most common type of heart disease, affecting an estimated 18.2 million Americans, or about one in 20 adults. It is the leading cause of death in the U.S., claiming the lives of 375,476 people in 2021 alone, 20% of whom were younger than 65.[1] By 2050, 61% of adults are expected to have some type of cardiovascular disease.[2]
Although CAD is highly prevalent, it is preventable and treatable. Effective treatment plans depend on accurately assessing the severity of the disease. The current standard of care for determining CAD risk based on risk factors like high cholesterol often gives an incomplete picture of a patient’s overall heart health. Quantitative analysis of plaque can be an effective way to obtain a more complete picture by including measurement of the patient’s total heart atherosclerosis burden. However, manual assessment by physicians is time-consuming and prone to subjectivity.[3]
Artificial intelligence (AI) provides a simpler and faster alternative and could be a potential game-changer for their heart patients. For example, HeartFlow’s AI-enabled Plaque Analysis tool leverages an automated, deep learning-based, non-invasive method for identifying, characterizing, and segmenting plaque throughout the coronary arteries. By offering a more comprehensive view, plaque analysis tools can improve diagnostic accuracy, enhance treatment planning, and help stratify risk.
Improving diagnosis and stratifying risk
Diagnostic protocols vary, but electrocardiography (EKG), stress testing, single photon emission computed tomography (SPECT) and coronary CT angiography (CCTA) are commonly used options. CCTA scans visualize the patient’s plaque burden, but manual quantification of plaque is highly labor-intensive, which has proven to be challenging due to time requirements and limited inter- and intra-observer reproducibility.[4]
Information gained from analyzing plaque using AI and proprietary algorithms can help physicians understand critical components of a patient’s plaque burden. More specifically, plaque analysis can help effectively identify, characterize, and segment plaque in the coronary arteries. By being able to assess the type of plaque, its characteristics, location and severity, physicians can stratify a patient’s risk for CAD.
When Plaque Analysis data are paired with insights from Fractional Flow Reserve – Computed Tomography (FFRCT), physicians can have lesion-specific FFRCT values to determine the need for immediate intervention as well as accurate quantification of calcified, non-calcified and low attenuation plaque, the latter of which is a marker of plaque instability and a major predictor of risk.[5]
Clinical implications of proper analysis of plaque
HeartFlow’s Plaque Analysis assessment of a patient’s CAD and risk of other cardiovascular events is highly correlated to invasive intravascular ultrasound (IVUS) for quantified plaque volumes for each vessel and lesion. The technology provides physicians with actionable information, such as specific plaque characteristics that correlate with higher risk. [6],[7],[8] According to the REVEALPLAQUE study, Plaque Analysis has 95% agreement for total plaque volume measures compared with the IVUS gold standard.[9] The ability to reference the cross-sectional views with plaque by type while reviewing a patient’s workup instills confidence in physicians that the insights are accurate and actionable. The quantitative measurement of a patient’s plaque volume enables physicians to more accurately assess risk and then develop a personalized, targeted treatment plan.
In fact, using the Plaque Analysis tool prompted physicians to modify medical management for two thirds of patients, according to the 2024 DECODE study.[10] Furthermore, nearly 50% of patients with a calcium score of zero were reclassified following clinician review of the Plaque Analysis, highlighting the importance of quantifying total plaque beyond calcium score.
To evaluate how physicians are changing their treatment decisions when afforded more detailed information about plaque, earlier this year HeartFlow established the DECIDE Registry, which will collect multi-site, real-world information from approximately 10,000 patients at about 25 sites across the U.S.
More detailed information about plaque could enable providers to perform a more detailed and accurate CAD risk assessment and identify more appropriate thresholds for preventive care, which are currently being underestimated by risk assessment scores.[11] This information may also help providers to assess response to therapeutic interventions and potentially identify patients with residual risk factors at an earlier stage in their disease process. By more accurately assessing patient CAD risk, physicians can produce more effective personalized treatment plans and
[1] Web MD. What is Coronary Artery Disease. https://www.webmd.com/heart-disease/coronary-artery-disease. Retrieved July 25, 2024.
[2] Heart disease and stroke could affect at least 60% of adults in U.S. by 2050. www.heart.org. (2024, June 4). https://www.heart.org/en/news/2024/06/04/heart-disease-and-stroke-could-affect-at-least-60-percent-of-adults-in-us-by-2050
[3] Rinehart S, et al. Utility of Artificial Intelligence Plaque Quantification: Results of the DECODE Study. JSCAI (2024)3;3:: https://www.jscai.org/article/S2772-9303(24)00003-6/fulltext?dgcid=raven_jbs_etoc_email
[4] Narula J, et al. Prospective Deep Learning-based Quantitative Assessment of Coronary Plaque by CT Angiography Compared with Intravascular Ultrasound. Eur Heart J Cardiovasc Imaging. 2024 May 3: jeae115. doi: 10.1093/ehjci/jeae115. https://academic.oup.com/ehjcimaging/advance-article/doi/10.1093/ehjci/jeae115/7663685
[5] Meah MN, Wereski R, Bularga A, et al. Coronary low-attenuation plaque and high-sensitivity cardiac troponin. Heart 2023;109:702-709. https://heart.bmj.com/content/109/9/702
[6] Lee, et al. JACC Cardiovasc Imaging 2019.
[7] Narula J, et al. Prospective Deep Learning-based Quantitative Assessment of Coronary Plaque by CT Angiography Compared with Intravascular Ultrasound. Eur Heart J Cardiovasc Imaging. 2024 May 3:jeae115. doi: 10.1093/ehjci/jeae115. https://academic.oup.com/ehjcimaging/advance-article/doi/10.1093/ehjci/jeae115/7663685
[8] Newby D. et al. Lancet 2015.
[9] Lee, et al. JACC Cardiovasc Imaging 2019.
[10] Rinehart S, et al. Utility of Artificial Intelligence Plaque Quantification: Results of the DECODE Study. JSCAI (2024)3;3:: https://www.jscai.org/article/S2772-9303(24)00003-6/fulltext?dgcid=raven_jbs_etoc_email
[11] Rinehart S, et al. Utility of Artificial Intelligence Plaque Quantification: Results of the DECODE Study. JSCAI (2024)3;3:: https://www.jscai.org/article/S2772-9303(24)00003-6/fulltext?dgcid=raven_jbs_etoc_email
Editor’s Note: Campbell Rogers, M.D., F.A.C.C. serves as the Chief Medical Officer. Prior to joining HeartFlow, he was the Chief Scientific Officer and Global Head of Research and Development at Cordis Corporation, Johnson & Johnson, where he was responsible for leading investments and research in cardiovascular devices. Prior to Cordis, he was Associate Professor of Medicine at Harvard Medical School and the Harvard-M.I.T. Division of Health Sciences and Technology, and Director of the Cardiac Catheterization and Experimental Cardiovascular Interventional Laboratories at Brigham and Women’s Hospital. He served as Principal Investigator for numerous interventional cardiology device, diagnostic, and pharmacology trials, is the author of numerous journal articles, chapters, and books in the area of coronary artery and other cardiovascular diseases, and was the recipient of research grant awards from the NIH and AHA.