Mammographic density and its association with molecular subtype of breast cancer.

Background: Breast cancer is the leading cause of death in women globally. The present study aimed to determine the prevalence of different mammographic density of breast parenchyma along with the association with the molecular subtype of breast cancer. Methodology: The present cross-sectional study was conducted among breast cancer females age 40 and above. The study participants' demographic and clinical data, such as age, sex, menstrual status, gravida, parity, and lactation, were collected using a standard questionnaire. The anthropometric parameters were taken by using the standard techniques. Breast physical examination, mammography, and ultrasound were done. The ultrasound did the breast tumor biopsy -guided techniques and histopathology, and immunohistochemistry was used to identify the molecular subtype of breast cancer. The univariate and multivariate statistic was performed on SPSS version 20.0. Results: Most of the patients<50 years were human epidermal growth factor receptor (HER2), estrogen receptor (ER), progesterone receptor (PR) positive, and are overweight too. Similar findings were noted among lactating and multigravida patients. The Grade I tumor and dense breast patients were 100% with HER2 negative and ER, PR positive. Most heterogeneous dense and fibroglandular containing breast patients were ER, PR positive. There is a significant correlation between BMI, microcalcification, and HER2 positive. Most of the triple Negative and Luminal-B were heterogeneous dense. Conclusion: In conclusion, a significant association between the mammographic density parameters and molecular breast cancer subtype, particularly Luminal-A and luminal B, was seen. The ER, PR, and HER-2 have a positive correlation with the physical radiographical and mammodensity parameters.


Introduction
Breast cancer is one of the leading causes of death among women, with most of the share belonging to Europe, North America, Australia, and New Zealand 1 . Almost 2.4 million new cases are predicted worldwide, accounting for one in every four cases among western women 2 . Although Asian countries fall behind the western world in the incidence of breast cancer, still Pakistan has a rising trend among Asian countries where one in every nine women experience breast cancer once in life time 3 , and total breast cancer incidence is estimated to be increased by approximately 23.1% in 2020 to 60.7% in 2025 4 . The worldwide overall mortality rate is reduced due to advancement and timely treatment, but unfortunately in the developing world like Pakistan, we still are striving hard for delayed diagnosis and overburdened health care system.
Dense breast tissue is a significant risk factor for breast carcinoma, and approximately 50% of females younger than 50 years have a high mammographic density (MD) of the breast. By definition, MD refers to fibro glandular mammary tissue comprising of epithelial cells, fibroblasts, and connective tissue 5 . MD is usually assessed by breast imaging reporting and data system (BI-RADS), that divides it into four categories 6 . These categories include fatty parenchyma, scattered fibro glandular parenchyma, heterogeneously dense parenchyma, and too dense parenchyma and are found in approximately 10%, 40%, 40%, and 10% respectively 7 . Molecular subtypes were assigned by hormone receptor status, tumor grade and include luminal A (ER/PR+ and grade I tumor, ER/PR+, and grade II tumor and ER+/PRand grade I), luminal B (ER+ and tumor having grade III, ER+/PR-and grade tumor and ER/PR+ and tumor having grade II), HER2-positive (ER+ or ER-and HER-2 neu positivity) a triple-negative (ER/PR-and HER-2 neu negative). Few studies have been conducted on the association of mammographic density with molecular breast cancer subtype type 8 . However, most of them were carried out in West or developed in Asian countries. No published data could be retrieved on the current topic in our population. There is a difference in MD and breast cancer in different ethnic groups 9 . Therefore, there might be a difference in the association between MD and molecular subtype type of breast cancer. Therefore, this study was done to determine the prevalence of different mammographic density of breast parenchyma along with the association of breast mammographic density with the molecular subtype of breast cancer.

Methodology
This cross-sectional study was conducted at the Dow University of Health Sciences. A total of 250 females age 40 or above diagnosed with breast cancer were considered for this study. Whereas, Women who do not undergo mammography at our institute, who have started treatment of breast cancer without having a mammogram, or women having breast implants or having lumpectomy/breast conservation surgery were not included in this study. The standard questionnaire was designed to take the demographic and clinical data of the study participants such as age, sex, menstrual status, gravida, parity, and lactation. The anthropometric parameters height, weight was recorded by using standard techniques.
The procedure was performed by an experienced radiologist having ten years of women imaging expertise. Physical examination of the breast was done before the ultrasound to evaluate the softness and hardness of the breast and location of the lesion before the procedure. Ultrasound was then performed to look for the imaging characteristics of the lesion followed by an ultrasound-guided trucut biopsy using a 16 gauge needle. Immunohistochemistry for grading and molecular subtypes were followed later. A mammogram was performed on the Senographe digital mammographic machine by General Electronics and two standard Craniocaudal (CC) and Mediolateral oblique (MLO) views of both breasts are taken. Mammographic features that were recorded include breast mammographic density, lesion margin, architectural distortion, skin thickness, nipple retraction, and microcalcification.  Statistical analysis was performed by using SPSS version 20.0. Qualitative variables such as breast density, menstrual status, marital status, the molecular subtype of breast cancer were expressed as frequency and percentage, and quantitative variables such as age as mean and standard deviation. A Chi-square test was applied to evaluate the association of clinical and breast density parameters with the molecular subtype of breast cancer. P-value ≤ 0.05 was considered as significant. Multivariate logistic regression was used to evaluate the correlation.

Results
The general characteristics of the patient and immunomolecular markers are illustrated in Table1  and Table 2. Patient characteristics stratified by HER2 gene, ER, and PR status are also presented in Table 1. The age of the patient < 50 years showed no difference.
The relationship between immunomolecular markers and clinical characteristics of breast tumor are presented in Table 3 where HER2 score shows a significant probability value with Gravida (p = 0.056) and Microcalcification (p = 0.047). Significant values were also observed between PR score and Architectural distortion (p = 0.027). PR also showed significant results with Skin Thickness (p = 0.038), whereas Mammographic density is different with Lactation (p = 0.053) and Menstrual status (p = 0.007). BMI showed significant differences with HER2 (p = 0.002), ER (p = 0.022), PR score (p = 0.033) and Parity category (p = 0.030). Microcalcification illustrated significant difference with Gravida-category (p = 0.009) and HER2 Score (p = 0.047). Skin thickness showed significant probability difference with Gravidacategory (p = 0.023), HER2, (p = 0.008), PR (p = 0.0.038) and PR score (p = 0.032). However significant difference with nipple retraction with HER2 score (p = 0.039) and architectural distortion with ERP score (p = 0.027) were also observed. No significant relationship was observed between HER2, ER, and PR score and Margins and Mammogram density in Table 3.

Discussion
In the present study, we examine the association between Breast mammographic density and Immunological markers such as HER2, ER, and PR. It was determined in many recent studies that the percent mammographic density (PMD) (expressed as a percentage of the breast area) associated with many risk factors, notably including BMI, Parity, Gravida, and Menopausal status, that are mainly associated with variations in mammographic density, and also were also associated with differences in one or more of these tissue features 10 . In our study findings, most of the study participants with age < 50 were ER and PR similar positive trend was seen among overweight patients. According to the finding of the study, similar findings were reported by de Kruijf (2014); the ER-positive were common in the advanced age group and had some significant associations 11 . Eppenberger-Castori's (2002) study also supported our findings 12 .
Furthermore, the Nattenmüller (2018) study was aligned with our study results that BMI was associated with ER/PR positive patients 13 . Although at the same time frame Ma (2018) noted that BMI is inversely associated with ER/PR positive cases 14 .
Our findings can be due to the genetic makeup of the study population. In our findings, the ER was positive significantly correlate with BMI; similarly, Phipps (2012) found that the BMI was positively associated with risks of ER+ in advanced age women 15 . The rare of our findings were PR and HER-2 inversely significantly correlated with BMI. The Babu (2018) and Kawai (2014) found similar findings; in both research pieces, there were variations in the correlation of BMI with the subtype of breast cancer 16,17 .
Another unique finding of our study was that the Grade-I tumor patients were 100% with the presentation of ER/PR positive 18 . Paul (2019) supported our findings that most of the grade-I tumor patients were ER/PR positive 19 . On the other hand, Musa (2018) also accessed the association of HR expression with grading and staging of a tumor; the findings appear that higher-grade tumor (II) was observed (76.93%) in type IV (ER/PR-, HER2+) and higher stage (III) was observed (80.95 %) in triple-negative subtype 20 . Although, Kaur (2016) found that the Grade-I tumor patients were 0 % ER/PR positive 21 . These findings suggest that the ER+/PR+ is a prognostic marker for the noninvasive tumor of the breast. The patients with tumors in upper outer areas also have the highest numbers of patients with ER/PR positive. No study was found to show the relationship between the tumor location and immunological markers to our best knowledge. However, our study does not demonstrate a statistically non-significant difference between all parameters regarding the immunological markers on bivariate analysis.
The breast density, which is considered an essential element by the previous studies, it was also noted that the Percent Mammographic Density (PMD) (expressed as a percentage of the breast area) associated with many risk factors, notably including BMI, parity, and gravida, and menopausal status, that are mainly associated with variations in mammographic density, and in addition, we are also associated with differences in one or more of these tissue features 22 . Thereby, it was shown previously in researches that parity and gravida status were inversely and significantly associated with breast tissue density and collagen percent 23 . In our findings, most of the multigravida study patients were ER-positive, but in the case of the PR+/PR-the study subjects were equal in percentage. Moreover, the HER-2 score had a positive significance correlation gravida. Palmer's (2011) findings were aligned with our findings. These findings supported that at the time of pregnancy and lactation, the breast can be affected by variously specific and unique disorders, including benign disorders closely related to physiologic changes, inflammatory and infectious diseases, etc. 24 . He also found that lactating women are at higher risk of a triple-negative breast cancer type.
Furthermore, in the present study, different breast density parameters were studied to evaluate the best correlation with the immunomarkers. Most of the high-density breast patients' ER+, but the PR+ mostly fell into the soft, dense breast category. The HER2 + were very low in the percentage. Furthermore, it was determined that although the mammographic density is the most crucial risk factor for breast cancer 23 but according to our results, there are variations associated with the subtypes of breast cancer. One more study have assessed breast cancer's molecular subtypes with different mammographic densities, and contrasting results have been shown 15 . In alignment with our study findings, Tang (2017) also observed similar findings 25 . Thereby, it was determined that though the different subtypes of breast cancer are separate biological entities, lifestyle, however, they are not resulted due to the differences in the mammographic densities and their association.
In our study, the different mammographic density parameters (Table 2) were accessed using the Physical examination, Mammographic and ultrasonic modalities similar to the Tang (2017) study 25 . The findings showed that most of the patients with all density parameters were PR+, although HER2 negative patients were in higher prevalence. A similar study was recently conducted by . According to the findings, the results may partially explain the higher proportion of HER2+ tumors 26 , and these findings were not aligned with our study findings.
Among the features, microcalcification and architectural distortion have a significant positive correlation with HER-2 score and PR. Bae found that HER2 status correlated positively with the mammographic calcifications noted radiologically 18 . Furthermore, WU (2017) also supported our findings that microcalcification can be conveniently used to facilitate the preoperative prediction of HER2 and Luminal A molecular subtype in patients with infiltrating ductal carcinoma 27 .
Previous studies have also assessed the association of breast cancer's molecular subtypes with different mammographic densities, and contrasting results have been shown. Yang Researches have previously been done to examine the association between mammographic densities and different breast cancer subtypes where receptor status (ER-, PR-, and HER2 status) was used as proxies. The data suggest that increasing dense volume was associated with luminal B and HER2-enriched subtypes among Chinese 28 . Norman (2017) 29 results suggest mammographic density measures are associated with all 'intrinsic' molecular subtypes. Our study findings found that most of the Luminal-A, Luminal-B, and triple-negative subjects were with heterogeneously dense and fibroglandular breast findings. Phipps (2012) conducted a case-control study to assess the association between breast densities, using a BI-RADS classification with different subtypes of breast cancers 15 . They achieved the same results as the current study; that is, the density was similarly associated with all subtypes. However, Edwards (2017) and Razzaghi (2012) found no correlation with the ER/PR positive subtype of breast cancer 8,9 . Thereby, it was determined that though the different subtypes of breast cancer are separate biological entities, they are not due to the differences in the mammographic densities and their association. Another study by Eriksson (2012) was done to analyze the association between density and the luminal A and basal-like subtypes of breast cancer 30 and, specifically, does not found any association. We thus believe that our results are not in agreement with both studies. To align with our results Arora (2010), where the association between density and the luminal A, luminal B, basal-like, and ERBB-2 subtypes of breast cancer was studied, using BI-RADS classification. They observed that women with extremely dense breasts had a higher frequency of luminal-A tumors 31 . Therefore, the results are similar to our study.
Moreover, one more study have also demonstrated that mammographic densities might be partially correlated with breast cancer's molecular pathology and with its subtypes 32 . Shaikh et al. showed similar results that there was a discordant relationship between molecular breast densities and subtypes of breast cancer 33 . According to the latest review, it was demonstrated that most studies had found no association between mammographic densities, hormone receptor status, and molecular subtypes of breast cancer 34 . Moreover, the two recent studies examining the relationship between HER2 status and mammographic density also do not found any association 34 . These studies showed similar results to the current study.
High mammographic density was considered a significant risk factor for breast cancer in females. However, the significantly positive association between mammographic densities and breast cancer subtypes did not vary materially by hormone receptor status, tumor tissue status, or by molecular subtypes defined by hormone marker status. These findings suggested that the association between breast cancer subtypes with mammographic breast densities may be because of other causal pathways. Therefore, future studies are recommended to assess the association with comprehensive information to different parameters like demographic and other risk factors.

Conclusion
In the current study, it was found that there is a significant association between the mammographic density parameters and molecular breast cancer subtype, particularly Luminal-A and luminal B. The ER, PR, and HER2 have a positive correlation with the physical, mammodensity parameters. However; future studies with more sample sizes are recommended to assess the strong prognostic predictor and their association with comprehensive information to different parameters like demographic and risk factors.

Conflicts of Interest
None.