Breast arterial calcification (BAC) is commonly observed on screening mammography and may provide a low-cost opportunistic marker to enhance cardiovascular risk assessment in women, in whom cardiometabolic risk is often underrecognized by traditional scores. We performed a PRISMA-guided systematic review and meta-analysis of PubMed, Embase, Scopus, and ClinicalTrials.gov through April 30, 2025, to evaluate the association between BAC and both future cardiovascular events and underlying coronary pathology. Cohort studies reporting adjusted hazard ratios (HRs) for incident cardiovascular events were pooled using random-effects meta-analysis, with prespecified subgroup analyses by BAC ascertainment method and study design, while additional studies reporting odds ratios (ORs) for coronary artery disease (CAD) or coronary artery calcium (CAC) were synthesized separately. Four cohort studies, including approximately 25,000 women with 6 to 12 years of follow-up demonstrated that BAC was associated with significantly higher incident cardiovascular events (pooled HR 1.82, 95% CI 1.37 to 2.43; p <0.001), with the strongest association observed for radiologist-reported BAC and concordant findings across artificial intelligence-derived and densitometric measures, and no evidence of small-study effects. Across OR-based analyses including approximately 5,000 women, BAC was strongly associated with underlying coronary pathology, with a pooled adjusted OR of 4.00 (95% CI 2.44 to 6.56) for CAD and similarly elevated odds for CAC. In conclusion, BAC detected on routine mammography identifies women at substantially higher future cardiovascular risk and is strongly associated with subclinical coronary disease, supporting its potential role as a scalable, no-added-cost marker to enhance cardiovascular risk assessment in women.
Cardiovascular disease (CVD) remains the leading cause of death in women globally, yet the risk in women is often underreported and underrecognized. Around 35% of all female deaths worldwide are due to CVD. Despite improvements in treatment and awareness, many women experience delays in diagnosis, poorer risk stratification, and suboptimal preventive care. This calls for an urgent need for better, sex-specific tools to identify at-risk women before overt disease manifests.
One increasingly recognized marker is breast arterial calcification (BAC), often noted incidentally during routine screening mammography. BAC, a form of medial arterial calcification (MAC), reflects a pathophysiology distinct from atherosclerosis, which is caused due to intimal calcification. In a study by Iribarren et al, breast vascular calcification was independently associated with elevated risks of coronary heart disease, stroke, and heart failure over decades to follow-up. These findings suggest that BAC may signal systemic vascular pathology with prognostic implications, along with being a benign radiographic finding. A recent meta-analysis examined the association between breast arterial calcification and cardiovascular outcomes. While informative, that study primarily synthesized a single effect measure and did not distinguish between prevalent coronary disease and incident cardiovascular risk. Moreover, emerging evidence using artificial intelligence–based BAC detection was not systematically evaluated. These gaps warrant an updated and methodologically stratified synthesis.
Recent reviews have demonstrated an association between BAC and cardiovascular pathology; however, most have relied primarily on cross-sectional designs and a single effect measure, limiting their ability to distinguish between underlying coronary disease and future cardiovascular risk. In this meta-analysis, BAC is evaluated across two complementary clinical domains. First, BAC is examined as an anatomic marker of prevalent coronary pathology using cross-sectional evidence reporting odds ratios (ORs), assessing its association with existing coronary artery disease (CAD) or coronary artery calcification (CAC) at the time of mammography. Second, BAC is evaluated as a prognostic marker of future cardiovascular risk using longitudinal cohort studies reporting hazard ratios (HRs) for incident cardiovascular events over time. These analyses address distinct but complementary clinical questions, prevalent disease versus future risk, and are therefore conducted separately rather than pooled together. ORs and HRs are not interchangeable, but together provide a comprehensive characterization of BAC as both an anatomic and prognostic cardiovascular marker in women.
The association between breast arterial calcification detected on screening mammography and cardiovascular outcomes in women, including both prevalent coronary disease and incident cardiovascular events, is summarized in the Central Illustration.
Methods
Study design and reporting framework
This systematic review and meta-analysis were conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) 2020 guidelines. The protocol was not prospectively registered. All analyses followed a predefined PICO framework evaluating the association between BAC detected on mammography and subsequent cardiovascular (CV) outcomes in women.
Data sources and search strategy
A comprehensive electronic search of PubMed/MEDLINE, Embase and Scopus was performed from database inception through August 10, 2025, without language or regional restrictions. Search strings combined controlled vocabulary and free-text terms related to BAC and cardiovascular disease using Boolean operators (“AND,” “OR”). Reference lists of eligible publications and relevant reviews were manually screened to identify additional studies. The full database-specific search strategies are provided in the Supplementary Table 1 .
Eligibility criteria
Studies were included if they met all of the following criteria:
-
•
Population: adult women (≥18 years) undergoing screening or diagnostic mammography.
-
•
Exposure: presence of BAC identified by (1) radiologist visual assessment, (2) artificial intelligence (AI)-based quantification, or (3) densitometric measurement.
-
•
Comparator: absence of BAC on mammogram.
-
•
Outcomes: incident composite cardiovascular outcomes (major adverse cardiovascular events [MACE] or ASCVD) or specific endpoints (myocardial infarction, stroke, heart failure, or cardiovascular mortality).
Design: observational studies, prospective or retrospective cohorts, case-control, or cross-sectional designs, reporting relative measures of association (hazard ratio [HR], odds ratio [OR], or risk ratio [RR]) and corresponding 95% confidence intervals (CIs). Random-effects models were used as the primary analytic approach, with alternative models applied for sensitivity analyses.
Exclusion criteria were: animal or autopsy studies, conference abstracts without extractable data, case reports or series with <10 participants, and duplicate cohorts (the largest or most recent study retained).
Study selection and data extraction
All retrieved records were screened independently by two reviewers using a structured screening platform, followed by full-text review. Duplicates were removed before two independent reviewers (PA and KM) screened titles and abstracts, followed by full-text review of potentially relevant studies. Disagreements were resolved by discussion or consultation with a third reviewer (AB).
Data were extracted independently by two reviewers using a standardized Microsoft Excel form capturing first author, publication year, country, cohort design (prospective/retrospective), sample size, mean age, detection method (radiologist/AI/densitometry), duration of follow-up, adjusted covariates, effect sizes (HRs and 95% CIs) for each CV endpoint, and reported quality metrics. Any discrepancies were adjudicated by consensus.
Risk of bias assessment
Study quality was assessed using the Newcastle–Ottawa Scale (NOS), an eight-item tool evaluating selection, comparability, and outcome/exposure domains for nonrandomized studies. Scores range from 0 to 9, with ≥7 indicating high quality. For cross-sectional studies, an adapted NOS version was applied. Assessments were performed independently by two reviewers, and disagreements were resolved by consensus ( Table 1 ).
Table 1
Newcastle–Ottawa scale (NOS) risk of bias
| Study | Selection | Comparability | Outcome | Total (0-9) | Risk of Bias |
|---|---|---|---|---|---|
| Newallo et al. | ★★★ | ★★ | ★★★ | 8 | Low |
| Pudil et al. | ★★★ | ★★ | ★★ | 7 | Moderate-Low |
| Nudy et al. | ★★★ | ★★ | ★★★ | 8 | Low |
| Yoon et al. | ★★★★ | ★★ | ★★★ | 9 | Very Low |
| Sykes | ★★★ | ★ | ★★ | 6 | Moderate |
| Hu et al. | ★★★★ | ★★ | ★★ | 8 | Low |
| Allen et al. | ★★★★ | ★★ | ★★★ | 9 | Very Low |
| Iribarren et al. | ★★★★ | ★★ | ★★★ | 9 | Very Low |
| Hendriks et al. | ★★★ | ★ | ★★ | 6 | Moderate |
Outcomes
The primary outcome was the association between the presence of BAC and the risk of composite cardiovascular events (MACE/ASCVD/global CVD) during follow-up. Secondary outcomes included individual components: myocardial infarction, stroke, heart failure, and cardiovascular mortality. For studies reporting stratified HRs (e.g., by BAC severity), the overall estimate was used for the primary analysis.
Statistical analysis
Cross-sectional analyses (prevalent coronary disease outcomes)
Statistical analyses were performed in RevMan using restricted maximum likelihood (RML) estimation. Both random-effects and mixed-effects models were applied to synthesize ORs for coronary artery disease (CAD) and coronary artery calcification (CAC) outcomes across cross-sectional and cohort studies. Odds ratios (ORs) were used to assess the association between BAC and prevalent coronary pathology, including coronary artery disease (CAD) and coronary artery calcification (CAC). These analyses quantify the relationship between BAC and existing coronary disease at the time of mammography.
Between-study heterogeneity was assessed using Chi-square (Q) and I² statistics, and publication bias was evaluated with funnel plots.
Cohort analyses (incident cardiovascular event outcomes)
Cohort studies reporting HRs were analyzed using MedCalc Statistical Software, version 23.3.7 (MedCalc Software Ltd, Ostend, Belgium). Hazard ratios (HRs) were used to evaluate the association between BAC and incident cardiovascular events over time. HRs reflect time-to-event risk and were derived exclusively from longitudinal cohort studies. OR-based and HR-based analyses were conducted separately, as they address distinct clinical questions and are not directly comparable. Heterogeneity was evaluated using Cochran’s Q (p <0.10) and quantified with I² (<25% low, 25 to 75% moderate, >75% high). Small-study effects and publication bias were assessed using standard visual and regression-based methods. Prespecified subgroup analyses examined heterogeneity by: BAC detection method (radiologist vs AI vs densitometry). Forest plots were generated in MedCalc, and heterogeneity was cross-validated in R (version 4.3.2).
Data sharing statement
This study used aggregate data extracted from previously published studies. No individual participant data were generated or analyzed. The extracted datasets and analytic code used for the meta-analysis are available from the corresponding author upon reasonable request.
Results
Following PRISMA 2020 guidelines, 170 records were identified through database searches. After removing 52 duplicates, 118 unique records were screened, and 98 were excluded. Twenty full-text reports were sought for retrieval; three could not be obtained. Seventeen articles were assessed for eligibility, of which eight were excluded (three wrong population and five abstract-only reports). Ultimately, nine studies were included in the review ( Figure 1 ). All nine studies included both cross-sectional and cohort designs, representing more than 25,000 women. ,,,,,,,, Results are presented separately for OR-based analyses assessing prevalent coronary disease and HR-based analyses assessing incident cardiovascular risk over time. The studies varied in methodology, including radiologist visual grading, densitometric quantification, and artificial intelligence-based detection, and follow-up in cohort studies ranged from 6 to 12 years. Cardiovascular endpoints included composite outcomes such as MACE, ASCVD, and global CVD, as well as specific events including myocardial infarction, ischemic stroke, heart failure, and cardiovascular mortality.
