Author + information
- Published online January 16, 2019.
- Jessica M. Peña, MD, MPH∗ ()
- Dalio Institute of Cardiovascular Imaging, Weill Cornell Medical College and NewYork-Presbyterian Hospital, New York, New York
- ↵∗Address for correspondence:
Dr. Jessica M. Peña, Dalio Institute of Cardiovascular Imaging NewYork-Presbyterian Hospital and Weill Cornell Medical College, New York-Presbyterian Hospital and Weill Cornell Medicine, 1305 York Avenue, WGC-8, New York, New York 10021.
- breast arterial calcification
- coronary artery calcification
- coronary artery disease
Trends in cardiovascular disease (CVD) among women have sharpened calls for improved risk stratification and adoption of sex-specific approaches to CVD prevention. For U.S. women age 40 to 79 years, risk assessment in its current state is largely predicated on estimation of an individual’s 10-year risk of atherosclerotic cardiovascular disease (ASCVD) using the pooled cohorts equation (PCE) (1). However, for many women, doubts remain about this approach. For example, the PCE has yet to be validated in Hispanic and Asian ethnicities—who together represent nearly a quarter of the U.S. population. As age and sex are major determinants of risk, there also remain questions about both overestimation (2) and underestimation of risk (3), as many women will be classified as low-risk even in the face of a significant burden of risk factors. Application of risk prediction tools in practice may also be lacking, with only 40% of women in 1 survey reporting an assessment of their heart health in the past year (4). These limitations underscore the need for improved risk stratification in women and race/ethnic minorities.
Nearly two-thirds of women older than the age of 40 years in the United States have undergone a recent mammogram for breast cancer screening (5) and screening for osteoporosis is recommended for women older than 65 years. Although age-related diseases such as cancer, osteoporosis, and CVD may share risk factors, their prevention has traditionally been viewed through separate lenses. Breast arterial calcifications (BACs) noted incidentally on screening mammography have been consistently associated with risk factors such as age and diabetes and an increased risk of coronary artery calcification (CAC) in several observational studies (6). In these prior studies, the sensitivity and specificity of BAC for CAC have ranged widely from 17% to 91% and 54% to 86%, respectively (6). While in aggregate these studies have suggested an association with impressive effect estimates, some limitations are noteworthy and leave outstanding questions. These limitations include selection bias, variable inclusion of women with known coronary artery disease or at high risk of CVD, less contemporary cohorts, and less sensitive mammography techniques for the detection of BAC. Given these limitations, a number of questions remain. Is the pathophysiology of BAC, a predominately medial process in the breast artery, related to traditional cardiovascular risk factors? Moreover, what is the clinical utility of BAC as an imaging biomarker beyond contemporary risk estimation? There have been calls for routine reporting of BAC on screening mammography reports (7); however, the implications for or optimal management of a woman found to have BAC are not known.
It is with this background that Yoon et al. (8) shed light on the interplay between BAC, low bone mass, and subclinical atherosclerosis in this issue of iJACC. Using a cross-sectional registry of 2,100 asymptomatic Korean women undergoing contemporaneous screening digital mammography, dual energy x-ray absorptiometry, and coronary computed tomography angiography at a single center, the authors sought to determine the independent association between BAC, low bone mass, and CAC or coronary artery plaque (CAP). Moreover, they also determined whether the addition of these imaging biomarkers improves prediction of subclinical coronary artery disease beyond 10-year risk estimates. Overall, women in the study were at low risk of CVD with a mean age of 52 years and estimated 10-year risk of atherosclerotic vascular disease of 3.3% using the Korean Risk Prediction Model (2.1% using PCE). The prevalence of BAC and low bone mass was 9.5% and 34%, respectively. CAC and CAP were present in 11.2% and 15.6% of women, respectively. As expected, women with evidence of subclinical coronary atherosclerosis had a higher burden of traditional risk factors.
After risk adjustment using a prediction model validated in the Korean population, as well as traditional risk factors, the presence of any BAC was associated with a 2.0- to 2.9-fold increase in the odds of CAC or any coronary plaque. Although this is significant in univariate analysis, low bone mass was not significantly associated with subclinical coronary artery disease in a fully adjusted model. Moving toward measures of accuracy, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value of BAC for the detection of CAC was calculated across strata of estimated 10-year risk. Among women who were at low risk for ASCVD (on the basis of an estimated 10-year risk <5%) the sensitivity, specificity, PPV, and negative predictive value of BAC for CAC was 13%, 94%, 13%, and 94%, respectively. The majority of women in this study (81%) fell into this low-risk category. In women whose estimated 10-year risk was between 5% and 7.5%, sensitivity, specificity, and PPV were 23%, 84%, 33%, respectively. The PPV was highest, 58%, among the minority of women with an estimated 10-year risk of ≥7.5%, where guidelines would already advocate consideration of statin therapy.
The ability of BAC to discriminate between individuals who have and do not have subclinical coronary disease was determined by adding BAC to a base model using the Korean Risk Prediction Model. The addition of BAC resulted in a very small, but statistically significant improvement in the area under the curve of 0.02 to 0.03. Similarly, the addition of BAC resulted in improvement in category-free net reclassification index and categorical net reclassification index for the surrogate of coronary computed tomography angiography–derived coronary plaque. In addition to being limited by the absence of hard cardiovascular events, application of the categorical NRI was also limited by use of risk thresholds that were driven by the prevalence of CAC and CAP, rather than clinically meaningful strata of risk used to make decisions regarding preventive therapies. Using a 10-year risk threshold of 11.2%, BAC resulted in 6.8% of women without CAC inappropriately reclassified to a higher risk, whereas 11.9% of women with CAC were appropriately reclassified to a higher risk category. Overall, these findings confirm the association between BAC and subclinical atherosclerosis, but highlight the limited utility of using BAC as a screening test for subclinical atherosclerosis in a low-risk population. They also remind us of the importance of disease prevalence in measures such as PPV. In this population, ∼8 women with BAC would need to undergo CAC testing to detect the presence of CAC in 1 woman.
The results of the BBC study by Yoon et al. (8) are timely and strengthened by a large sample size and use of digital mammography and contemporary risk assessment methods in a low-risk screening population. The authors should also be congratulated in their attempt to extend the knowledge of BAC beyond measures of association and determine the clinical utility of this imaging biomarker in a screening population. Although the current study was performed in an East Asian population, it is estimated that nearly 80% of U.S. women have an estimated 10-year ASCVD risk <7.5%, and thus these findings are of broad interest (1).
Beyond surrogate measures, future longitudinal studies will need to address whether the addition of BAC to contemporary risk estimation improves prediction of hard CVD events in a screening population. The ongoing MINERVA (Multiethnic Study of Breast Arterial Calcium Gradation and Cardiovascular Disease) study will examine the role of BAC for prediction of CVD in older women (9). As with any biomarker, routine adoption of BAC reporting will also require evidence that knowledge of BAC results in superior outcomes in a clinical setting.
↵∗ Editorials published in JACC: Cardiovascular Imaging reflect the views of the authors and do not necessarily represent the views of iJACC or the American College of Cardiology.
Dr. Peña has no relationships relevant to the contents of this paper to disclose.
- 2019 American College of Cardiology Foundation
- Cook N.R.,
- Ridker P.M.
- Karmali K.N.,
- Goff D.C.,
- Ning H.,
- Lloyd-Jones D.M.
- Bairey Merz C.N.,
- Andersen H.,
- Sprague E.,
- et al.
- ↵FastStats. Available at:. https://www.cdc.gov/nchs/fastats/mammography.htm. 3172018.
- Ryan A.J.,
- Choi A.D.,
- Choi B.G.,
- Lewis J.F.
- Hecht H.,
- Blaha M.J.,
- Berman D.S.,
- et al.
- Yoon Y.,
- Kim K.M.,
- Han J.S.,
- Kang S.-H.
- Iribarren C.,
- Sanchez G.,
- Husson G.,
- et al.