Advertisement

Mortality impact of low CAC density predominantly occurs in early atherosclerosis: explainable ML in the CAC consortium

Published:November 11, 2022DOI:https://doi.org/10.1016/j.jcct.2022.10.001

      Abstract

      Background

      Machine learning (ML) models of risk prediction with coronary artery calcium (CAC) and CAC characteristics exhibit high performance, but are not inherently interpretable.

      Objectives

      To determine the direction and magnitude of impact of CAC characteristics on 10-year all-cause mortality (ACM) with explainable ML.

      Methods

      We analyzed asymptomatic subjects in the CAC consortium. We trained ML models on 80% and tested on 20% of the data with XGBoost, using clinical characteristics ​+ ​CAC (ML 1) and additional CAC characteristics of CAC density and number of calcified vessels (ML 2). We applied SHAP, an explainable ML tool, to explore the relationship of CAC and CAC characteristics with 10-year all-cause and CV mortality.

      Results

      2376 deaths occurred among 63,215 patients [68% male, median age 54 (IQR 47–61), CAC 3 (IQR 0–94.3)]. ML2 was similar to ML1 to predict all-cause mortality (Area Under the Curve (AUC) 0.819 vs 0.821, p ​= ​0.23), but superior for CV mortality (0.847 vs 0.845, p ​= ​0.03).
      Low CAC density increased mortality impact, particularly ≤0.75. Very low CAC density ≤0.75 was present in only 4.3% of the patients with measurable density, and 75% occurred in CAC1-100. The number of diseased vessels did not increase mortality overall when simultaneously accounting for CAC and CAC density.

      Conclusion

      CAC density contributes to mortality risk primarily when it is very low ≤0.75, which is primarily observed in CAC 1–100. CAC and CAC density are more important for mortality prediction than the number of diseased vessels, and improve prediction of CV but not all-cause mortality. Explainable ML techniques are useful to describe granular relationships in otherwise opaque prediction models.

      Keywords

      Abbreviations:

      CAC (Coronary Artery Calcium), ML (Machine Learning), SHAP (SHapley Additive exPlanations), AUC (Area under the receiver operating curve), NRI (Net reclassification index), IDI (Integrated discrimination index)
      To read this article in full you will need to make a payment

      Purchase one-time access:

      Academic & Personal: 24 hour online accessCorporate R&D Professionals: 24 hour online access
      Full access to the journal is a member benefit for SCCT Members, Login via the SCCT website to access all journal content.

      Subscribe:

      Subscribe to Journal of Cardiovascular Computed Tomography
      Already a print subscriber? Claim online access
      Already an online subscriber? Sign in
      Institutional Access: Sign in to ScienceDirect

      References

        • Grundy S.M.
        • Stone N.J.
        • Bailey A.L.
        • et al.
        AHA/ACC/AACVPR/AAPA/ABC/ACPM/ADA/AGS/APhA/ASPC/NLA/PCNA guideline on the management of blood cholesterol: a report of the American college of cardiology/American heart association Task Force on clinical practice guidelines.
        Circulation. 2018; 139 (2019): e1082-e1143
        • McClelland R.L.
        • Chung H.
        • Detrano R.
        • Post W.
        • Kronmal R.A.
        Distribution of coronary artery calcium by race, gender, and age: results from the Multi-Ethnic Study of Atherosclerosis (MESA).
        Circulation. 2006; 113: 30-37
        • McClelland R.L.
        • Jorgensen N.W.
        • Budoff M.
        • et al.
        10-Year coronary heart disease risk prediction using coronary artery calcium and traditional risk factors: derivation in the MESA (Multi-Ethnic study of atherosclerosis) with validation in the HNR (Heinz Nixdorf recall) study and the DHS (Dallas heart study).
        J Am Coll Cardiol. 2015; 66: 1643-1653
        • Razavi A.C.
        • Agatston A.S.
        • Shaw L.J.
        • et al.
        Evolving role of calcium density in coronary artery calcium scoring and atherosclerotic cardiovascular disease risk.
        JACC Cardiovasc Imaging. 2022; 15: 1648-1662
        • Razavi A.C.
        • van Assen M.
        • De Cecco C.N.
        • et al.
        Discordance between coronary artery calcium area and density predicts long-term atherosclerotic cardiovascular disease risk.
        JACC Cardiovasc Imaging. 2022;
        • Shea S.
        • Navas-Acien A.
        • Shimbo D.
        • et al.
        Spatially weighted coronary artery calcium score and coronary heart disease events in the multi-ethnic study of atherosclerosis.
        Circ Cardiovasc Imaging. 2021; 14e011981
        • Blaha M.J.
        • Budoff M.J.
        • Tota-Maharaj R.
        • et al.
        Improving the CAC score by addition of regional measures of calcium distribution: multi-ethnic study of atherosclerosis.
        JACC Cardiovasc Imaging. 2016; 9: 1407-1416
        • Criqui M.H.
        • Denenberg J.O.
        • Ix J.H.
        • et al.
        Calcium density of coronary artery plaque and risk of incident cardiovascular events.
        JAMA. 2014; 311: 271-278
        • Foldyna B.
        • Eslami P.
        • Scholtz J.E.
        • et al.
        Density and morphology of coronary artery calcium for the prediction of cardiovascular events: insights from the Framingham Heart Study.
        Eur Radiol. 2019; 29: 6140-6148
        • Blaha M.J.
        • Mortensen M.B.
        • Kianoush S.
        • Tota-Maharaj R.
        • Cainzos-Achirica M.
        Coronary artery calcium scoring: is it time for a change in methodology?.
        JACC (J Am Coll Cardiol). 2017; 10: 923-937
        • Nakanishi R.
        • Slomka P.J.
        • Rios R.
        • et al.
        Machine learning adds to clinical and CAC assessments in predicting 10-year CHD and CVD deaths.
        JACC (J Am Coll Cardiol). 2020; 14: 615-625
        • Sánchez-Cabo F.
        • Rossello X.
        • Fuster V.
        • et al.
        Machine learning improves cardiovascular risk definition for young, asymptomatic individuals.
        J Am Coll Cardiol. 2020; 76: 1674-1685
        • Engelhard M.M.
        • Navar A.M.
        • Pencina M.J.
        Incremental benefits of machine learning—when do we need a better mousetrap?.
        JAMA Cardiol. 2021; 6: 621-623
        • Quer G.
        • Arnaout R.
        • Henne M.
        • Arnaout R.
        Machine learning and the future of cardiovascular care: JACC state-of-the-art review.
        J Am Coll Cardiol. 2021; 77: 300-313
        • Wenzl F.A.
        • Kraler S.
        • Ambler G.
        • et al.
        Sex-specific evaluation and redevelopment of the GRACE score in non-ST-segment elevation acute coronary syndromes in populations from the UK and Switzerland: a multinational analysis with external cohort validation.
        Lancet. 2022; 400: 744-756
        • Kakadiaris I.A.
        • Vrigkas M.
        • Yen A.A.
        • Kuznetsova T.
        • Budoff M.
        • Naghavi M.
        Machine learning outperforms ACC/AHA CVD risk calculator in MESA.
        J Am Heart Assoc. 2018; 7e009476
        • Sengupta P.P.
        • Shrestha S.
        • Berthon B.
        • et al.
        Proposed requirements for cardiovascular imaging-related machine learning evaluation (PRIME): a checklist: reviewed by the American college of cardiology Healthcare Innovation council.
        JACC (J Am Coll Cardiol). 2020; 13: 2017-2035
        • Shaw L.J.
        • Min J.K.
        • Nasir K.
        • et al.
        Sex differences in calcified plaque and long-term cardiovascular mortality: observations from the CAC Consortium.
        Eur Heart J. 2018; 39: 3727-3735
        • Osei A.D.
        • Mirbolouk M.
        • Berman D.
        • et al.
        Prognostic value of coronary artery calcium score, area, and density among individuals on statin therapy vs. non-users: the coronary artery calcium consortium.
        Atherosclerosis. 2021; 316: 79-83
        • Dzaye O.
        • Razavi A.C.
        • Dardari Z.A.
        • et al.
        Mean versus peak coronary calcium density on non-contrast CT: calcium scoring and ASCVD risk prediction.
        JACC Cardiovasc Imaging. 2022; 15: 489-500
        • Molnar C.
        Interpretable machine learning: a guide for making black box models explainable.
        https://christophm.github.io/interpretable-ml-book/
        Date: 2019
        Date accessed: February 16, 2021
        (Accessed)
        • Lundberg S.M.
        • Erion G.
        • Chen H.
        • et al.
        From local explanations to global understanding with explainable AI for trees.
        Nat Mach Intell. 2020; 2: 56-67
        • Lundberg S.M.
        • Nair B.
        • Vavilala M.S.
        • et al.
        Explainable machine-learning predictions for the prevention of hypoxaemia during surgery.
        Nat Biomed Eng. 2018; 2: 749-760
        • Deshmukh F.
        • Merchant S.S.
        Explainable machine learning model for predicting GI bleed mortality in the intensive care unit.
        Am J Gastroenterol. 2020; 115: 1657-1668
        • Vaid A.
        • Johnson Kipp W.
        • Badgeley Marcus A.
        • et al.
        Using deep-learning algorithms to simultaneously identify right and left ventricular dysfunction from the electrocardiogram.
        JACC (J Am Coll Cardiol). 2022; 15: 395-410
        • Fahmy Ahmed S.
        • Csecs I.
        • Arafati A.
        • et al.
        An explainable machine learning approach reveals prognostic significance of right ventricular dysfunction in nonischemic cardiomyopathy.
        JACC (J Am Coll Cardiol). 2022; 15: 766-779
        • Bhatt S.
        • Cohon A.
        • Rose J.
        • et al.
        Interpretable machine learning models for clinical decision-making in a high-need, value-based primary care setting.
        NEJM Catal. 2021; 2
        • Blaha M.J.
        • Whelton S.P.
        • Al Rifai M.
        • et al.
        Rationale and design of the coronary artery calcium consortium: a multicenter cohort study.
        J Cardiovasc Comput Tomogr. 2017; 11: 54-61
        • Sun X.
        • Xu W.
        Fast implementation of DeLong's algorithm for comparing the areas under correlated receiver operating characteristic curves.
        IEEE Signal Process Lett. 2014; 21: 1389-1393
        • Pencina M.J.
        • D'Agostino R.B.
        • Sr .,
        • D'Agostino Jr., R.B.
        • Vasan R.S.
        Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond.
        Stat Med. 2008; 27 (discussion 207-112): 157-172
        • Mortensen M.B.
        • Fuster V.
        • Muntendam P.
        • et al.
        Negative risk markers for cardiovascular events in the elderly.
        J Am Coll Cardiol. 2019; 74: 1-11
        • Tota-Maharaj R.
        • Blaha M.J.
        • Blankstein R.
        • et al.
        Association of coronary artery calcium and coronary heart disease events in young and elderly participants in the multi-ethnic study of atherosclerosis: a secondary analysis of a prospective, population-based cohort.
        Mayo Clin Proc. 2014; 89: 1350-1359
        • Yahagi K.
        • Kolodgie F.D.
        • Otsuka F.
        • et al.
        Pathophysiology of native coronary, vein graft, and in-stent atherosclerosis.
        Nat Rev Cardiol. 2016; 13: 79-98
        • Motoyama S.
        • Ito H.
        • Sarai M.
        • et al.
        Plaque characterization by coronary computed tomography angiography and the likelihood of acute coronary events in mid-term follow-up.
        J Am Coll Cardiol. 2015; 66: 337-346
        • van Rosendael A.R.
        • Narula J.
        • Lin F.Y.
        • et al.
        Association of high-density calcified 1K plaque with risk of acute coronary syndrome.
        JAMA Cardiol. 2020; 5: 282-290
        • Raffield L.M.
        • Cox A.J.
        • Criqui M.H.
        • et al.
        Associations of coronary artery calcified plaque density with mortality in type 2 diabetes: the Diabetes Heart Study.
        Cardiovasc Diabetol. 2018; 17: 67
        • Parikh R.B.
        • Teeple S.
        • Navathe A.S.
        Addressing bias in artificial intelligence in Health care.
        JAMA. 2019; 322: 2377-2378
        • Orimoloye O.A.
        • Budoff M.J.
        • Dardari Z.A.
        • et al.
        Race/ethnicity and the prognostic implications of coronary artery calcium for all-cause and cardiovascular disease mortality: the coronary artery calcium consortium.
        J Am Heart Assoc. 2018; 7e010471
        • Jones N.
        • Marks R.
        • Ramirez R.
        • Rios-Vargas M.
        Census illuminates racial and ethnic composition of the country.
        (Published 2021. Updated August 12, 2021)
        • European Commission
        Proposal for a Regulation laying down harmonised rules on artificial intelligence. Shaping Europe’s digital future web site.
        (Published 2021. Updated 4/21/2021)
        • Ikemura K.
        • Bellin E.
        • Yagi Y.
        • et al.
        Using automated-machine learning to predict COVID-19 patient mortality.
        J Med Internet Res. 2021;
        • Vaid A.
        • Somani S.
        • Russak A.J.
        • et al.
        Machine learning to predict mortality and critical events in a cohort of patients with COVID-19 in New York city: model development and validation.
        J Med Internet Res. 2020; 22e24018
        • Whelton S.P.
        • Rifai M.A.
        • Marshall C.H.
        • et al.
        Coronary artery calcium and the age-specific competing risk of cardiovascular versus cancer mortality: the coronary artery calcium consortium.
        Am J Med. 2020; 133: e575-e583