AS729 - Women's Health Initiative breast arterial calcification ancillary study

Investigator Names and Contact Information

Matthew Nudy (mnudy@pennstatehealth.psu.edu)

Introduction/Intent

Cardiovascular disease (CVD) is the number one cause of death globally¹ and in the United States more women die from CVD than men.² Sudden cardiac arrest may be the presenting manifestation of CVD. Therefore, identifying women at higher risk for developing CVD could save many lives. Unfortunately, current CVD risk prediction tools are limited and have been shown to underestimate women’s risk of CVD compared to men.^3,4 The currently available tool used to estimate CVD risk-the ACC/AHA pooled cohort risk assessment equation (PCE)-does not identify the majority of women who experience a first heart attack or stroke, who could have been targeted with primary prevention efforts.⁵ Thus, a critical unmet need exists for better screening and risk stratification strategies in women to identify those at increased risk for CVD.

Mammography is a well-accepted imaging modality for breast cancer screening that is widely available. Most women over the age of 50 years have a mammogram performed at least every two years.⁶ Medium sized arteries in the breast can become calcified and subsequently visible on mammography. This is called a breast arterial calcification (BAC, Figure 1) and is observed in up to 25% (1 in 4) of mammograms.^7,8 BAC is not commonly mentioned on radiology reports since there is no known association between BAC and breast cancer. However, this “incidental” finding is emerging as a potential marker of subclinical CVD.⁹ My preliminary data evaluated 1,309 women with no known CVD at baseline and found that BAC was associated with a 3-fold increase in coronary artery disease and 5-fold increase in stroke after 10 years of follow-up.¹⁰ Mammography may be an ideal way to improve CVD risk prediction in women without additional cost, radiation exposure, time and resources. However, clinically important questions remain. BAC incidence has not been analyzed longitudinally in an ethnically diverse population and quantification of BAC severity has not been validated. An objective measure of BAC severity exists using densitometry but this method is expensive, requires frequent calibration, is not widely available and cannot be applied to most image formats.¹¹ A subjective BAC severity grade has been described but is not validated.¹² Furthermore, it is unknown if BAC severity associates with a higher risk of CVD.

The main purpose of this NHLBI K23 proposal is for me to gain the necessary research/statistics skills, career development training and mentoring to conduct independent, funded, CVD prevention research. For the research component of this award, 3 annual mammograms from 848 ethnically diverse postmenopausal women enrolled in the Women’s Health Initiative-Mammogram Density Ancillary Study^13,14 will be analyzed for the annual incidence of BAC and will have BAC severity graded by two blinded radiologists. The main hypothesis of the research component of this K23 award is that BAC incidence will increase over a 3-year study period and subjective BAC severity grades can be reliably determined by radiologists without the use of additional costly equipment. Also, we aim to determine if an association between BAC severity and CVD risk exists. Finally, we aim to determine whether follow-up BAC severity (progression or regression) correlates with CVD risk factors and/or risk prediction scores. In other words, if CVD risk factors improve from baseline to follow-up assessment, will BAC likewise improve over that same time period.

Specific Aim 1: Determine the annual incidence of BAC on screening mammograms (baseline, year 1, and year 2) in ethnically diverse postmenopausal women and establish the inter-rater reliability of BAC severity graded by two blinded radiologists. We hypothesize that BAC incidence will increase from baseline to year 1 and to year 2 on screening mammograms. In addition, we hypothesize that the subjective BAC severity grade will have acceptable inter-rater reliability using Cohen’s weighted kappa statistic.

Specific Aim 2: In a cross-sectional manner, ascertain the association between BAC severity, CVD risk factors, and ASCVD risk scores. We hypothesize that severe BAC will be associated CVD risk factors including hypertension, hyperlipidemia, obesity, diabetes, age and tobacco use. Furthermore, BAC severity will associate with blood pressure, body mass index, lipid concentrations, glucose concentrations, and ASCVD risk scores in a linear manner.

Specific Aim 3: Determine the relationship between baseline 10-year-risk of CVD (via the PCE) and observed 10-year incidence of major adverse cardiac events (MACE) in postmenopausal women and assess the role BAC status has in CVD risk prediction when applied to the PCE. We hypothesize that the addition of BAC presence to the PCE will improve CVD prediction in postmenopausal women.