Search

Introduction:Breast arterial calcification (BAC), which can be seen on mammograms, is emerging as a risk stratification tool for cardiovascular disease (CVD) outcomes including heart failure, though how prediction varies by race and ethnicity is unknown.Methods:This single-center, retrospective study included 18,092 women age ≥40 years who had a digital mammogram between 2008-2016 and clinical follow-up through 2019. BAC was quantified using an automated algorithm (CureMetrix), resulting in a continuous Bradley score. Heart failure (HF) events and self-reported race and ethnicity were obtained via medical records. Adjusted hazard ratios (aHR) were calculated using Cox proportional hazard regression models with interaction terms to evaluate the effect of race and ethnicity on the association between BAC (presence/absence; or Bradley score quartile among those with BAC) with incident HF, adjusted for age, diabetes, smoking, systolic and diastolic blood pressure, high-density lipoprotein cholesterol, low-density lipoprotein cholesterol, and history of CVD and chronic kidney disease.Results:Mean age was 57±11 years with 11,319 (63%) of women White, 907 (5%) Black, 1,694 (9%) Hispanic, 2,321 (13%) Asian, and 1,851 (10%) Other. Over a median follow-up of 3.0 years, 149 of the 4,076 women with BAC (3.7%) developed HF compared to 149 of the 14,016 women without BAC (1.1%). Presence of BAC was associated with incident HF (aHR 1.38, 95% CI 1.21-1.58, p<0.001) overall and among all racial and ethnic groups, with increased risk seen for increasing quartiles of Bradley score (aHR 1.17 in the first, 1.30 in the second, 1.79 in the third, 1.69 in the fourth; p-trend<0.001). The presence of BAC was an even stronger predictor of HF in Black women compared to White women (Figure; aHR 1.51 [1.00-2.30], p interaction=0.05).Conclusions:The presence and degree of BAC on a mammogram is an independent predictor of HF and may serve as an especially strong risk factor among Black women.

BackgroundCardiovascular disease (CVD) remains the number one cause of death in women. Current screening tools in this population do not identify all women who are at risk. Screening for breast artery calcification (BAC) via routine mammography is emerging as a risk enhancing tool; however, how risk prediction varies by age is unknown.MethodsThis single-center, retrospective study included 14,544 women who had a digital mammogram from 2008-2016. BAC was quantified with the Bradley score derived from an automated algorithm developed by CureMetrix; presence of BAC was defined as a Bradley score ≥10. Mortality data were collected after index mammogram by review of medical records. Adjusted hazard rations (aHR) were calculated using Cox regression models with interaction terms to evaluate the influence of age on the association between presence of BAC and mortality, after adjusting for age, diabetes, medications, race, current smoking, systolic blood pressure, total and HDL cholesterol, plus prior CVD.ResultsIn the overall study population 8% had diabetes, 30% hypertension, 30% hyperlipidemia and 5% were current smokers. BAC was detectable in 522 of 9,027 (6%) of women age 40-59 years, 657 of 2,865 (23%) 60-69 years, and 702 of 1,358 (52%) 70+. Over mean follow-up of 4.2 years, there were 328 deaths (15.1%) among women with BAC. When subjects were separated by age, presence of BAC was associated with mortality among younger women (40-59 years: aHR 1.50 [95% CI 1.06-2.12]; 60-69 years: aHR 1.59 [95% CI 1.19-2.12], Figure). Among women ages 70+, BAC was not a significant predictor of mortality (aHR 1.11 [95% CI 0.86-1.43]). There was a significant negative interaction effect of age with BAC for prediction of mortality (p < 0.001).Conclusions:BAC is predictive of mortality, especially in younger women. BAC may be a useful screening adjunct to traditional CVD risk factors, to help identify women at an early age who may benefit from more targeted preventive measures.Figure

Introduction:Breast arterial calcification (BAC) is gaining acceptance as a CVD risk factor (RF), though overlap with traditional RFs remains unclear. We tested the hypothesis that BAC, assessed via the Bradley score, is associated with CVD RFs.Methods:We analyzed baseline cross-sectional data from 16,217 women who underwent screening digital mammography between 2008-2016 at a single center. BAC was quantified using an algorithm from a trained, deep neural network (CureMetrix), resulting in an automated and continuous Bradley score; >10 was considered BAC positive. CVD RFs were obtained via medical records. We calculated odds ratios (OR) for RFs per Bradley score quartile and created distribution plots.Results:At baseline (mean age 55 ± 11 yrs; 70% White, 45% Black, 25% other), 8% had diabetes, 30% hypertension, 30% hyperlipidemia, and 5% smoked; 22% and 6% took lipid-lowering and anticoagulant medication, respectively. There were 1,904 women (13%) with BAC (median Bradley score 42). BAC prevalence increased with age, and those in the highest quartile of the Bradley score, compared to lowest, were significantly more likely to have hypertension (OR 4.1, 95% CI 3.3-5.0; p<0.001), diabetes (OR 2.8, 95% CI 2.2-3.6; p<0.001), and hyperlipidemia (OR 2.4, 95% CI 2.0-3.0; p<0.001) (Figure).Additionally, 492 women (7%) with zero traditional RFs had BAC (Table).Conclusions:BAC prevalence on mammography is robustly associated with traditional CVD RFs, however BAC was present in 7% of women without RFs. Future studies are needed to determine the significance of BAC among women without RFs, as it may provide insight and opportunity for prevention before traditional CVD RFs develop.

Introduction:Breast arterial calcification (BAC), seen on mammography, has emerged as a risk stratification tool and surrogate marker of cardiovascular disease (CVD). Studies evaluating the association of BAC with clinical outcomes are limited.Methods:This single-center, retrospective study included 14,544 women who had a digital mammogram from 2008-2016. BAC was quantified with the Bradley score using an automated algorithm based on a trained, deep neural network (CureMetrix). The presence of BAC was defined as Bradley score >10. Clinical outcomes were collected via medical records using ICD-10 diagnoses. Cox regression models were used to evaluate the association between BAC (continuous and binary) and clinical outcomes including mortality and composite outcome (mortality, stroke, and myocardial infarction). Models were adjusted for age, diabetes, medications, race, current smoking, systolic blood pressure, total & HDL cholesterol, and prior CVD diagnosis.Results:The mean age was 54.7 ±11.0 years with 8% diabetes, 30% hypertension, 30% hyperlipidemia and 5% smokers. BAC was present in 13%. Over a mean follow-up of 4.2 years, there were 1,687 deaths (6.8%) and 2,113 composite outcomes (8.5%). On multivariable analysis, each 10-point increase in BAC score was significantly associated with clinical outcomes: mortality (aHR 1.05 [95% CI 1.02-1.08], p < 0.001) and composite outcome (aHR 1.07 [95% CI 1.04-1.09], p < 0.001). With BAC as a binary variable, women with BAC had a 1.52 (95% CI 1.22-1.89) higher risk of mortality and a 1.72 (95% CI 1.40-2.11) higher risk of composite outcome compared to those without BAC.Conclusion:BAC is significantly associated with clinical outcomes, including mortality and CVD. Further study is needed to confirm these findings.

Purpose:Mammographically detected breast arterial calcification (BAC) has been reported as a surrogate marker for coronary vascular disease (CVD). BAC is often not reported, reducing the power of this risk marker. Here we assess a deep-learning algorithm for BAC quantification, and in cases where the software identified BAC but the readers did not, assess correlation with clinical CVD.Methods:Screening digital mammograms (n=285, 1232 images) were selected from a tertiary Australian hospital from a sample of patients who underwent CVD screening (137 confirmed CVD). Two readers identified binary presence/absence of BAC. A deep learning software (cmAngio™ , CureMetrix, USA) quantified BAC score (scale 0-100). Reader-software discordance was adjudicated by a third observer. Area under the curve (AUC) was used to assess performance with 95% confidence intervals (CI) presented. Interobserver agreement was assessed by Cohen’s kappa (k).Results:Interobserver agreement for visual BAC was good (k=0.78, CI:0.71-0.85, p<0.01). BAC scores ranged from 0-100, with 12% scoring 0. The algorithm had high performance, AUC 0.92 (CI:0.88-0.96, p<0.001) vs initial reader assessment, increasing to AUC 0.98 (CI:0.97-0.99, p<0.001) after adjudicated assessment. At an empirical threshold of BAC score ≥5 to denote BAC presence, BAC prevalence was 38% (109/285) by software, and 31% (87/285) visually. In cases where only the software identified BAC (n=35), BAC scores were all <25, with 37% (n=13) having confirmed CVD and 9% (n=3) myocardial infarction.Conclusion:cmAngio demonstrates excellent performance for detecting BAC on screening mammograms. Faint BAC may be missed by human readers, is better identified by the software, and in these cases there is a high rate of CVD suggesting use in a screening population may improve risk prediction for CVD.