Radiographic markers of cardiovascular risk based on digital mammograms: a cross-sectional study

Full Text

BACKGROUND

According to published studies and guidelines, mammography is the primary screening modality for breast cancer and is recommended for women aged ≥40 years, either annually or biennially [1, 2].

This imaging method has demonstrated high effectiveness in the early detection of cancer, establishing itself as a rapid, accessible, and cost-effective diagnostic tool capable of covering large populations of women [1, 2].

As a result, there is growing interest in the potential use of mammography not only for breast cancer detection but also for obtaining additional information regarding women’s overall health status. In particular, the identification of mammographic features associated with an increased risk of socially relevant diseases, such as ischemic heart disease and atherosclerosis, opens new opportunities for expanding preventive strategies and may contribute to improved prognosis [3, 4].

Calcification of the mammary arterial walls demonstrates high diagnostic value regarding the risk of coronary artery calcification, which is considered a reliable marker of coronary atherosclerosis and ischemic heart disease, as well as a substantial predictor of adverse cardiovascular events [3–5]. Morphologically, medial calcification of the breast arteries differs from the intimal atherosclerotic lesions of the coronary arteries associated with ischemic heart disease. Nevertheless, multiple studies have confirmed a significant correlation between breast arterial calcification and cardiovascular risk factors, as well as the prevalence of ischemic heart disease [6].

Recent studies have further demonstrated that mammographic markers of cardiovascular risk are not limited to vascular calcifications. In a review by Bochkareva et al. [7], based on large cohort studies, mammographic density was also shown to be a predictor of adverse cardiovascular events. Sardu et al. [8] reported that women with the lowest mammographic density before menopause (American College of Radiology [ACR] category A) exhibited a 10-year incidence of major adverse cardiovascular events of 19.6%, which was several-fold higher than that in women with higher mammographic density: 7.6%, 3.3%, and 2.0% for ACR categories B, C, and D, respectively. Moreover, low mammographic density (ACR A) during the premenopausal period was associated with a threefold increase in the 10-year cardiovascular risk. Despite the increasing interest in the relationship between mammographic features and cardiovascular risk, existing studies in this field have several limitations that hinder the broad clinical application of their findings.

• First, the scientific evidence on the diagnostic and prognostic value of breast arterial calcification as a marker of cardiovascular risk is still inconsistent. This parameter is not currently incorporated into international guidelines for the diagnosis and prevention of cardiovascular diseases (CVDs), which limits its adoption in clinical algorithms [4].
• Second, methodological limitations of previous studies substantially restrict the generalizability of their results. Some studies were conducted using a cross-sectional design without long-term follow-up [5, 6], which precludes evaluation of the trends of CVD development and complications. Moreover, breast arterial calcification has often been evaluated using simplified binary classification (present vs. absent) without quantitative characterization of calcification severity [5, 6], thereby reducing the sensitivity of assessment.

Of particular interest are studies in which breast arterial calcification is analyzed not in isolation but in combination with other mammographic markers. Aldous et al. [9] reported that integrated assessment of mammographic density and breast arterial calcification improves the prognostic accuracy for ischemic heart disease risk compared with evaluation of either parameter alone.

These findings underscore the promise of a comprehensive approach and highlight the need for the development and implementation of standardized assessment protocols. However, most previous investigations have not accounted for the combined effect of mammographic density and the presence of breast arterial calcification on cardiovascular risk [6–8]. Importantly, these markers reflect different aspects of a woman’s vascular and hormonal status and may therefore possess synergistic prognostic value.

AIM

To evaluate the association of mammographic density and the extent of breast arterial calcification with cardiovascular risk category in asymptomatic women aged 40 years and older.

METHODS

Study Design

A cross-sectional, sample-based, single-center study was conducted.

Study Setting

The study included data from women who underwent preventive screening mammography at Lomonosov Moscow State University Clinic between 2019 and 2023. No deviations from the planned time frames occurred.

Study Duration

The study was conducted between September and December 2024.

Eligibility Criteria

Inclusion criteria:

• Women with available preventive screening mammography findings;
• Women aged 40 years and older;
• Preventive screening mammography performed at Lomonosov Moscow State University Clinic between 2019 and 2023.

Non-inclusion criteria:

• Presence of artifacts on mammograms;
• Technically inadequate screening mammography (nonstandard positioning, incomplete breast coverage).

Exclusion criteria:

• Absence of medical records required for cardiovascular risk calculation;
• Confirmed ischemic heart disease.

Screening Mammography

Digital mammography was performed using a General Electric Senographe Essential® system (GE Healthcare, USA) in two standard projections.

To assess mammographic density and types of calcifications in each breast, a radiologist analyzed craniocaudal and mediolateral oblique images at a SenoIris® workstation (GE Healthcare, USA).

Study Outcomes

Primary study outcome

The primary outcome was the association between mammographic density and the presence of breast arterial calcification with cardiovascular risk category.

Measurement of Study Outcomes

Assessment of breast density

Mammographic findings were interpreted using the Breast Imaging Reporting and Data System (BI-RADS) classification [10].

Breast density was evaluated according to the standardized classification proposed by the American College of Radiology (ACR) and incorporated into the BI-RADS [10]. According to this classification, there are four categories of breast density (see Fig. 1). To present data in tables, we converted the conventional ACR letter designations of density categories to numerical values as follows:

• A (1), predominantly fatty breast tissue (hereinafter, density refers to the degree of X-ray attenuation by breast tissue). Mammography has high sensitivity for lesion detection;
• B (2), scattered areas of fibroglandular tissue with low radiographic density;
• C (3), heterogeneously dense breast tissue or sufficiently dense focal areas that may obscure small lesions;
• D (4), extremely dense breasts. Mammography has low sensitivity for lesion detection.

 

Fig. 1. American College of Radiology breast density scale: a, ACR A; b, ACR B; c, ACR C; d, ACR D. ACR, American College of Radiology.

 

Assessment of breast arterial calcification

The type of breast arterial wall calcification was evaluated using the following scale developed by the authors:

• 1, no calcifications;
• 2, non-vascular calcifications only;
• 3, vascular calcifications only;
• 4, both vascular and non-vascular calcifications.

Moreover, the extent of breast quadrant involvement was recorded using the following scale:

• 0, no involved quadrants;
• 1, one quadrant;
• 2, two quadrants;
• 3, three or four quadrants.

The number of calcifications within each quadrant was also graded as follows:

• 0, no calcifications;
• 1, single;
• 2, multiple.

Assessment of cardiovascular risk category

The cardiovascular risk category was determined in accordance with the guidelines of the Russian Society of Cardiology using the Systematic Coronary Risk Estimation 2 (SCORE2) model [11].

Data required to estimate the 10-year risk of fatal and non-fatal CVDs were obtained from patients’ medical records. The following modifiable and non-modifiable risk factors were considered:

• Age;
• Body mass index;
• Harmful habits (smoking);
• Blood glucose level;
• Total cholesterol level;
• Systolic blood pressure;
• Medical history: type 2 diabetes mellitus, hypertension, and CVDs.

Group Analysis

Ten distinct patient subgroups (G0–G9) were formed based on the presence of calcifications, breast density, type of calcifications, and the number of involved breast quadrants (1 or ≥2) (see Table 1).

 

Table 1. Criteria for the initially formed groups
Group	Code	Description
G9	{1,2} (3,23,*)	Low breast density (ACR 1 or 2); Vascular and nonvascular calcifications; Two or more quadrants involved.
G8	{1,2} (2,23,*)	Low breast density (ACR 1 or 2); Vascular calcifications only; Two or more quadrants involved.
G7	{1,2} (23,1,*)	Low breast density (ACR 1 or 2); Vascular only or combined vascular and nonvascular calcifications; One quadrant involved.
G6	{1,2} (1,123,*)	Low breast density (ACR 1 or 2); Nonvascular calcifications only.
G5	{1,2} (0,0,0)	Low breast density (ACR 1 or 2); No calcifications.
G4	{3,4} (3,23,*)	High breast density (ACR 3 or 4); Vascular and nonvascular calcifications; Two or more quadrants involved.
G3	{3,4} (2,23,*)	High breast density (ACR 3 or 4); Vascular calcifications only; Two or more quadrants involved.
G2	{3,4} (23,1,*)	High breast density (ACR 3 or 4); Vascular only or combined vascular and nonvascular calcifications; One quadrant involved.
G1	{3,4} (1,123,*)	High breast density (ACR 3 or 4); Nonvascular calcifications only.
G0	{3,4} (0,0,0)	High breast density (ACR 3 or 4); No calcifications.

 

Furthermore, Uniform Manifold Approximation and Projection (UMAP), a modern machine learning method, was applied. This algorithm was developed by McInnes et al. in 2020 [12]. It is currently among the most advanced dimensionality reduction techniques. A distinctive feature of this method is its relatively high computational efficiency compared with other algorithms. The objective of UMAP is to model a manifold with a fuzzy topological structure from a given dataset (set of points) and subsequently embed it into a low-dimensional representation that preserves a topologically equivalent fuzzy structure. Based on UMAP analysis of the mammographic data, the patients were divided into three clusters:

• Cluster I: high breast density (ACR C or D) without calcifications;
• Cluster II: low breast density (ACR A or B) without calcifications;
• Cluster III: presence of any calcifications regardless of breast density.

A more detailed analysis was then performed within cluster III (patients with calcifications), resulting in the final classification into 8 groups (G7–G0) (see Table 2).

 

Table 2. Criteria for the final eight groups
Group	Code	Description
G7	{1,2} (3,23,*)	Low breast density (ACR 1 or 2); Vascular and nonvascular calcifications; Two or more quadrants involved.
G6	{1,2} (2,23,*)	Low breast density (ACR 1 or 2); Vascular calcifications only; Two or more quadrants involved.
G5	{1,2} (23,1,*)	Low breast density (ACR 1 or 2); Vascular only or combined vascular and nonvascular calcifications; One quadrant involved.
G4	{1,2} (1,123,*)	Low breast density (ACR 1 or 2); Nonvascular calcifications only.
G3	{1,2} (0,0,0)	Low breast density (ACR 1 or 2); No calcifications.
G2	{3,4} (123,23,*)	High breast density (ACR 3 or 4); Calcifications of any type; Two or more quadrants involved.
G1	{3,4} (123,1,*)	High breast density (ACR 3 or 4); Calcifications of any type; One quadrant involved.
G0	{3,4} (0,0,0)	High breast density (ACR 3 or 4); No calcifications.

 

Stratification by type of calcifications and number of involved breast quadrants was based on binary logistic regression, in which the dependent variable was the cardiovascular risk category, taking two values: 0 (low or moderate risk) and 1 (high or very high risk).

Group comparisons were performed according to the cardiovascular risk category using odds ratios (ORs). The reference group was defined as the most favorable mammographic profile: women with high breast density (ACR C or D) and no calcifications. The OR for each group is presented relative to the reference.

Ethics Approval

The study protocol was approved by the local Ethics Committee of the Medical Research and Education Center of Lomonosov Moscow State University (Minutes No. 5 of October 16, 2023). All participants provided informed consent prior to radiographic mammography. The data analysis was performed retrospectively; therefore, no additional patient consent was obtained.

Statistical Analysis

Because the study employed a predefined period of consecutive patient enrollment, the sample size was not calculated in advance.

Statistical analysis was conducted using the open-source software environment R version 4.4.2. Age was expressed as M ± SD, where M is the mean and SD is the standard deviation. Categorical variables were described using absolute values and percentages (%). Clustering was performed using the UMAP machine learning algorithm. ORs and their 95% confidence intervals (CIs) were calculated using univariate logistic regression. Contingency tables were analyzed using the Pearson’s chi-square (χ²) test. To construct groups according to the severity of mammographic findings, multivariable logistic regression was applied. The significance level for all tested hypotheses was set at p = 0.05.

RESULTS

Sample Characteristics

The study sample consisted of 1030 women aged 40–89 years who underwent preventive mammography at Lomonosov Moscow State University Clinic. The distribution of participants by cardiovascular risk category, as determined by the SCORE2 scale, is presented in Table 3. Notably, the largest proportion of women (33.7%) had a moderate cardiovascular risk.

 

Table 3. Distribution of patients according to cardiovascular risk category based on SCORE2
Risk category	n (%)
Low risk	36 (3.5)
Moderate risk	347 (33.7)
High risk	306 (29.7)
Very high risk	341 (33.1)
Note. Systematic Coronary Risk Estimation 2 (SCORE2) is a scale for estimating the 10-year cardiovascular risk.

 

Table 4 presents the characteristics of the 10 initially formed patient groups (G9–G0) according to mammographic density, presence and type of calcifications, and number of involved breast quadrants (1 or ≥2). Mammographic findings in groups G9–G0 are presented in Figs. 2–11, respectively.

 

Table 4. Characteristics of the ten initially formed groups for the left and right breast
Group	Breast	n (%)	Mean age, years	High and very high cardiovascular risk, n (%)	Odds ratio (95% confidence interval)
G9	Left	100 (9.7)	70.3 ± 7.6	92 (92)	23.7 (9.9–63.8)
	Right	95 (9.2)	70.1 ± 7.6	86 (90.5)	19 (8.3–48.4)
G8	Left	179 (17.4)	63.7 ± 8.9	135 (75.4)	6.4 (3.6–11.6)
	Right	175 (17)	63.5 ± 9.2	132 (75.4)	6.2 (3.5–11)
G7	Left	190 (18.4)	63.1 ± 9.4	131 (68.9)	4.6 (2.7–8.2)
	Right	205 (19.9)	63 ± 9.5	142 (69.3)	4.5 (2.7–7.9)
G6	Left	49 (4.8)	65.5 ± 8.7	33 (67.3)	4.3 (2–9.7)
	Right	52 (5)	66.3 ± 8.4	38 (73.1)	5.4 (2.5–12.4)
G5	Left	222 (21.6)	59.2 ± 9.4	125 (56.3)	2.7 (1.6–4.7)
	Right	213 (20.7)	59.2 ± 9.3	117 (54.9)	2.5 (1.5–4.2)
G4	Left	21 (2)	69.8 ± 9.9	16 (76.2)	6.6 (2.1–25.3)
	Right	26 (2.5)	69.4 ± 10.7	21 (80.8)	8.4 (2.8–30.9)
G3	Left	78 (7.6)	58.6 ± 11.4	39 (50)	2.1 (1.1–4.1)
	Right	68 (6.6)	58.4 ± 11.9	31 (45.6)	1.7 (0.9–3.3)
G2	Left	91 (8.8)	57 ± 10.7	40 (44)	1.6 (0.9–3.1)
	Right	83 (8.1)	56.5 ± 10.4	40 (48.2)	1.9 (1–3.6)
G1	Left	6 (0.6)	72.2 ± 14.9	5 (83.3)	10.3 (1.1–502.9)
	Right	9 (0.9)	66.8 ± 13.5	6 (66.7)	4 (0.8–26.2)
G0	Left	93 (9)	53.1 ± 10.4	30 (32.3)	Reference
	Right	103 (10)	54.1 ± 10.6	34 (33)	Reference

 

Fig. 2. Group G9: low breast density, vascular and nonvascular calcifications with two or more quadrants involved.

 

Fig. 3. Group G8: low breast density, vascular calcifications only, with two or more quadrants involved.

 

Fig. 4. Group G7: low breast density, vascular or combined vascular and nonvascular calcifications, with one quadrant involved.

 

Fig. 5. Group G6: low breast density, nonvascular calcifications only.

 

Fig. 6. Group G5: low breast density, no calcifications.

 

Fig. 7. Group G4: high breast density, vascular and nonvascular calcifications with two or more quadrants involved.

 

Fig. 8. Group G3: high breast density, vascular calcifications only, with two or more quadrants involved.

 

Fig. 9. Group G2: high breast density, vascular or combined vascular and nonvascular calcifications, with one quadrant involved.

 

Fig. 10. Group G1: high breast density, nonvascular calcifications only.

 

Fig. 11. Group G0: high breast density, no calcifications.

 

After applying the UMAP machine learning method to the mammographic data, the participants were stratified into three clusters: cluster I (n = 93), cluster II (n = 222), and cluster III (n = 715). After a detailed analysis of cluster III (patients with calcifications), the final eight groups (G7–G0) were formed, the characteristics of which are presented in Table 5.

 

Table 5. Characteristics of the eight final groups for the left and right breasts
Group	Breast	n (%)	Mean age, years	High and very high cardiovascular risk, n (%)	Odds ratio (95% confidence interval)
G7	Left	100 (9.7)	70.3 ± 7.6	92 (92)	23.7 (9.9–63.8)
	Right	95 (9.2)	70.1 ± 7.6	86 (90.5)	19 (8.3–48.4)
G6	Left	179 (17.4)	63.7 ± 8.9	135 (75.4)	6.4 (3.6–11.6)
	Right	175 (17)	63.5 ± 9.2	132 (75.4)	6.2 (3.5–11)
G5	Left	190 (18.4)	63.1 ± 9.4	131 (68.9)	4.6 (2.7–8.2)
	Right	205 (19.9)	63 ± 9.5	142 (69.3)	4.5 (2.7–7.9)
G4	Left	49 (4.8)	65.5 ± 8.7	33 (67.3)	4.3 (2–9.7)
	Right	52 (5)	66.3 ± 8.4	38 (73.1)	5.4 (2.5–12.4)
G3	Left	222 (21.6)	59.2 ± 9.4	125 (56.3)	2.7 (1.6–4.7)
	Right	213 (20.7)	59.2 ± 9.3	117 (54.9)	2.5 (1.5–4.2)
G2	Left	103 (10)	61.8 ± 12.4	59 (57.3)	2.8 (1.5–5.3)
	Right	98 (9.5)	61.6 ± 12.6	54 (55.1)	2.5 (1.4–4.6)
G1	Left	93 (9)	57 ± 10.6	41 (44.1)	1.7 (0.9–3.1)
	Right	88 (8.5)	57.2 ± 10.9	44 (50)	2 (1.1–3.8)
G0	Left	93 (9)	53.1 ± 10.4	30 (32.3)	Reference
	Right	103 (10)	54.1 ± 10.6	34 (33)	Reference

 

Primary Results

High mammographic density in the presence of calcifications was associated with a lower probability of high and very high cardiovascular risk compared with low mammographic density combined with calcifications. The strongest associations with high and very high cardiovascular risk were observed in groups G9 and G8. Conversely, the absence of calcifications corresponded to the lowest probability of high and very high cardiovascular risk (see Table 4).

A significant association was found between the number of involved breast quadrants and the severity of involvement in both the left and right breast (see Table 6) (p <  0.001).

 

Table 6. Association between the number of involved quadrants and the severity of calcification
Number of involved quadrants	Left breast, n (%)		Right breast, n (%)
	Severity
	1	2	1	2
1	278 (87 7)	39 (12.3)	294 (89.4)	35 (10.6)
2	99 (73.3)	36 (26.7)	83 (68)	39 (32)
3	97 (37)	165 (63)	99 (37.8)	163 (62.2)

 

The main challenge in classifying women with calcifications was that the proportion of patients with nonvascular calcifications alone was relatively small compared with the overall group of patients with any type of calcifications (see Table 7).

 

Table 7. Distribution of patients by type of calcification in the left and right breasts
Breast	No calcifications, n (%)	Nonvascular calcifications (type 1), n (%)	Vascular calcifications (type 2), n (%)	Vascular and nonvascular calcifications (type 3), n (%)
Left	315 (30.6)	55 (5.3)	518 (50.3)	142 (13.8)
Right	316 (30.7)	61 (5.9)	515 (50)	138 (13.4)

 

Notably, when stratifying patients into 10 groups, women with high mammographic density were not subclassified according to calcification type but were subdivided according to the number of involved quadrants.

Moreover, the small sample sizes of groups G1 and G4 resulted in high ORs and wide CIs. For this reason, group G4 was merged with group G3, and group G1 was redistributed: patients with one involved quadrant were reassigned to group G2, whereas those with two or more involved quadrants were reassigned to group G3.

Thus, eight new mutually exclusive groups were formed (see Tables 1 and 2). Under this classification, women with high breast density remained unstratified by calcification type but continued to be stratified by the number of involved quadrants. Figure 12 illustrates the resulting ORs plotted on a logarithmic scale.

 

Fig. 12. Odds ratios with 95% confidence intervals for each of the seven groups (G1–G7) compared with the reference group G0 on a base-2 logarithmic scale: a, left breast; b, right breast.

 

DISCUSSION

Summary of Primary Results

The results of the mathematical analysis of cardiovascular risk factors in relation to mammographic density, the presence of calcifications, and their extent of distribution are presented. In women with low mammographic density (ACR A or B) and vascular calcifications involving more than one breast quadrant, the probability of being classified as having a high or very high cardiovascular risk exceeded 75%. In turn, when vascular and nonvascular calcifications were present in two or more quadrants, the probability of assignment to these risk categories exceeded 90%.

Interpretation of Study Results

The findings demonstrate that, in addition to conventional cardiovascular risk factors, mammographic density and the severity of breast calcifications play an important role in the female population. The findings emphasize the necessity of a comprehensive assessment that integrates standard clinical risk factors with radiographic characteristics, improving the accuracy of risk stratification and promoting timely identification of women with high and very high cardiovascular risk according to SCORE2, in agreement with the findings reported by Aldous et al. [9]. The authors demonstrated that the presence of breast arterial calcifications and low mammographic breast density, both independently and in combination, is associated with ischemic heart disease and improves cardiovascular risk prediction compared with standard risk assessment. This underscores the need for an integrated approach and supports the development of unified evaluation algorithms. Nevertheless, the majority of previously published studies did not assess the combined impact of mammographic density and breast calcifications, including their type and extent, on cardiovascular risk [6–8].

In turn, when determining the cardiovascular risk category, it is important to consider not only the presence of vascular calcifications but also their combination with nonvascular calcifications. Vascular calcifications have characteristic imaging features and appear on mammograms as parallel hyperdense lines following the course of the vessel walls. They are also referred to as Mönckeberg-type calcifications, located in the media of the arterial wall (see Fig. 13) [13, 14]. Nonvascular calcifications include microcalcifications, defined as calcium deposits within breast tissue measuring < 0.5 mm. In most cases, these are benign and are detected in approximately half of examined women. Microcalcifications are believed to be associated with the deposition of calcium-rich compounds arising from various physiological and pathophysiological processes, including secretory, inflammatory, and involutional changes, as well as trauma or necrosis.¹ Moreover, intradermal microcalcifications, predominantly found in the dermis, are caused by calcification of sebaceous gland ducts. They usually appear in clusters and have ring-shaped or punctate morphology (see Fig. 14). Weddell-type calcifications, characterized by square, triangular, or trapezoidal shapes, are typical of fibrocystic changes (see Fig. 15). Round calcifications with central lucency and eggshell-type calcifications are observed in macro- and microcyst calcification (see Fig. 16). For small breast cysts, milk of calcium calcifications are typical; they result from calcium sedimentation within the cyst cavity and appear cup-shaped or crescent-shaped on lateral views (see Fig. 17). Multiple scattered punctate microcalcifications are characteristic of sclerosing adenosis, a benign condition associated with enlargement of glandular tissue lobules and their compression by a fibrous component [13] (see Fig. 18). Coarse microcalcifications with multiple internal lucencies are observed in calcified breast papillomas (see Fig. 19). Linear and rod-like calcifications oriented along ducts develop during plasma cell infiltration of the periductal stroma and basal layer proliferation; they are a pathognomonic sign of previous plasma cell mastitis (see Fig. 20).

 

Fig. 13. Calcification of the breast arterial walls.

 

Fig. 14. Intradermal microcalcifications.

 

Fig. 15. Weddell-type calcifications.

 

Fig. 16. Calcification of a small breast cyst.

 

Fig. 17. Cup-shaped calcification.

 

Fig. 18. Multiple punctate microcalcifications in sclerosing adenosis.

 

Fig. 19. Calcification of a papilloma within the breast tissue.

 

Fig. 20. Radiographic appearance of previous plasma cell mastitis: linear calcifications along the milk ducts.

 

An important parameter associated with cardiovascular risk is the number of breast quadrants containing calcifications. When analyzing their presence and distribution within a single breast, the accuracy of the association with the cardiovascular risk category is not significantly reduced; however, the workload required for radiologic assessment is reduced by approximately twofold.

We believe that one of the most important aspects of applying these findings as radiographic markers of cardiovascular risk is the possibility of their integration into screening programs using artificial intelligence technologies. This is particularly relevant given the subjective nature of assessing mammographic density and calcification severity, which largely depends on radiologist experience. In our view, software systems capable of automated assessment of breast density and calcification presence and extent may substantially improve objectivity, reproducibility, and standardization of image interpretation.

Further studies are required to more comprehensively investigate the relationship between breast arterial calcifications, mammographic density, and cardiovascular risk category.

Study Limitations

The study did not assess the incidence of cardiovascular events.

Other limitations of this study are a relatively small sample size and uneven distribution of participants among the study groups.

Given that the sample size required to achieve adequate statistical power was not calculated before or during the study, the resulting sample cannot be considered sufficiently representative. This precludes extrapolation of the findings and their interpretation in the general population of comparable patients beyond the scope of this study.

CONCLUSION

The findings confirm the association between mammographic density, the severity of calcification of breast arteries, and cardiovascular risk.

Machine learning methods in combination with statistical analysis made it possible to stratify patients into ranked groups. However, the current number of identified groups remains relatively large, which complicates the implementation of these findings in routine clinical practice. Nevertheless, the findings may serve as a basis for further systematization and the development of a convenient decision-making algorithm for patient routing.

In our view, the combined use of mammographic density and the degree of breast arterial calcification is a promising approach for predicting cardiovascular risk categories. This strategy has the potential to improve both the diagnosis and prevention of CVDs.

Given that mammography is already a standard method for breast cancer screening, incorporating breast density assessment and evaluation of breast arterial calcification into routine mammographic analysis may provide a broader range of relevant health information without requiring additional costs.

ADDITIONAL INFORMATION

Author contributions: D.D. Tsurskaya: conceptualization, methodology, investigation, formal analysis, writing—original draft, writing—review & editing, visualization; E.A. Mershina: conceptualization, methodology, investigation, formal analysis, writing—review & editing; V.E. Sinitsyn: conceptualization, methodology, investigation, writing—review & editing; O.E. Ivlev: investigation, software, formal analysis, writing—review & editing; E.M. Filichkina: investigation, formal analysis, writing—original draft, writing—review & editing, visualization; E.B. Yarovaya: conceptualization, methodology, formal analysis, writing—review & editing; G.O. Dolgushin, Ia.A. Orlova: resources, investigation, writing—review & editing. All the authors approved the version of the manuscript to be published and agreed to be accountable for all aspects of the work, ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Ethics approval: The study protocol was approved by the local Ethics Committee of the Medical Research and Educational Center Moscow at the Lomonosov Moscow State University (Minutes No. 5, dated October 16, 2023).

Funding sources: The work was performed by the university’s research teams as part of Moscow State University’s interdisciplinary research project under a state assignment in the interest of the University’s Interdisciplinary Research and Educational Schools. Project No. 23-Sh05-08: An Integrated Method for Assessing Cardiovascular Risk Using Mammography-Based Probabilistic Models.

Disclosure of interests: The authors have no relationships, activities, or interests for the last three years related to for-profit or not-for-profit third parties whose interests may be affected by the content of the article.

Statement of originality: No previously obtained or published material (text, images, or data) was used in this study or article.

Data availability statement: The editorial policy regarding data sharing does not apply to this work.

Generative AI: No generative artificial intelligence technologies were used to prepare this article.

Provenance and peer-review: This article was submitted unsolicited and reviewed following the standard procedure. The peer review process involved three external reviewers and a member of the Editorial Board.

 

¹ Udoh AI, He J. Microcalcifications [Internet]. In: PathologyOutlines.com. Available at: https://www.pathologyoutlines.com/topic/breastcalcification.html Accessed on: December 22, 2024.

About the authors

Daria D. Tsurskaya

Lomonosov Moscow State University

Author for correspondence.
Email: dashnom03@gmail.com
ORCID iD: 0009-0008-7732-4093
SPIN-code: 5298-8707

MD

Russian Federation, Moscow

Elena A. Mershina

Lomonosov Moscow State University

Email: elena_mershina@mail.ru
ORCID iD: 0000-0002-1266-4926
SPIN-code: 6897-9641

MD, Cand. Sci. (Medicine), Assistant Professor

Russian Federation, Moscow

Valentin E. Sinitsyn

Lomonosov Moscow State University

Email: vsini@mail.ru
ORCID iD: 0000-0002-5649-2193
SPIN-code: 8449-6590

MD, Dr. Sci. (Medicine), Professor

Russian Federation, Moscow

Oleg E. Ivlev

Lomonosov Moscow State University

Email: oleg.ivlev@math.msu.ru
ORCID iD: 0000-0002-3663-6305
SPIN-code: 8257-0252
Russian Federation, Moscow

Elena M. Filichkina

Lomonosov Moscow State University

Email: elena.filichkina1999@yandex.ru
ORCID iD: 0000-0003-3715-6896
SPIN-code: 3153-4281
Russian Federation, Moscow

Elena B. Yarovaya

Lomonosov Moscow State University

Email: yarovaya@mech.math.msu.su
ORCID iD: 0000-0002-6615-4315
SPIN-code: 5591-8439

Dr. Sci. (Physics and Mathematics)

Russian Federation, Moscow

Grigory O. Dolgushin

Lomonosov Moscow State University

Email: grdolgushin@yandex.ru
ORCID iD: 0000-0002-5981-3933
SPIN-code: 3452-9799

MD

Russian Federation, Moscow

Iana А. Orlova

Lomonosov Moscow State University

Email: 5163002@bk.ru
ORCID iD: 0000-0002-8160-5612
SPIN-code: 3153-8373

MD, Dr. Sci. (Medicine), Professor

Russian Federation, Moscow

References

Supplementary files