Mammographic Breast Density in Infertile and Parous Women

Meggiorini Maria Letizia; Vestri Anna Rita; De Stefano Maria Grazia; Cipolla Valentina; Bellati Filippo; Maffucci Diana; Nusiner Maria Paola; Aragona Cesare; De Felice Carlo


BMC Womens Health. 2016;16(8) 

In This Article


This study was carried out from January 2007 to November 2009. Data collection methods and study design were approved by the institutional review board of the University of Rome "Sapienza" and all patients signed an informed consent form. The study included women with primary infertility referred to our Breast Unit at the Department of Gynaecology and Obstetrics, University of Rome "Sapienza" as well as a control group selected from the database of the Department of Radiology.

A total of 440 patients were referred to our Breast Unit from the IVF Centre for a clinical breast examination before undergoing specific infertility treatment. Of these patients, 51 % were >35 years old, 37 % had previously undergone fertility drug treatment and 9 % had secondary infertility. Of the 440 patients, 151 (34 %) were women aged >35 with primary infertility who had never undergone fertility drug treatment (case group). The case group was compared to a control group consisting of 154 parous women selected from the database of the Department of Radiology who had at least one previous pregnancy (average 1.3 children per woman). The control group patients had undergone mammography examination for other reasons, such as family history or clinical suspicion of breast tumours, and they were negative for cancer. The age range of the control group was chosen in accordance with the case group: premenopausal women age ≥35. The information collected in both groups included: age, family history of breast cancer with at least 2 first degree relatives, previous administration of hormonal contraceptive therapy (HCT), age at menarche, previous pregnancy or etiology of infertility (if known) as well as height and weight for the calculation of body mass index (BMI). According to the FONCAM recommendations, all recruited case group patients underwent clinical examination and X-ray mammography (XRM) after signing the informed consent form. In all cases, conventional XRM was performed at our Department of Radiological Science using digital image formation and computed radiography. At least two views per breast were obtained.

All mammograms were classified using BIRADS and Boyd scales. Classification was carried out by three physicians (two radiologists and a breast specialist) blinded to the clinical data in accordance with the guidelines of the American College of Radiology (ACR) BIRADS. Diagnostic quality of the mammograms was assessed according to the PGMI evaluation system that includes four image quality categories: P (perfect), G (good), M (moderate) and I (inadequate).

BIRADS is a qualitative system, but the recently modified, more quantitative version was employed.[18]

According to the BIRADS lexicon, patients were assigned to one of the four categories of breast parenchymal density distribution: BIRADS-1: the breast is almost entirely fat (glandular parenchyma <25 % of the total area of both breasts); BIRADS-2: scattered fibroglandular densities (25%-50 %); BIRADS-3: heterogeneously dense breast tissue (51%-75 %); BIRADS-4: extremely dense (>75 % glandular tissue). The Boyd scale is based on a visual estimation of the proportion of mammographic density in the projected area, divided into six categories of unequal intervals: type A: 0 %; type B: 0–10 %; type C:10–25 %; type D: 25–50 %; type E: 50–75 % and type F: >75 %. In case of contradictory judgments, the classification assigned by at least two readers out of three was considered correct.

Patients were excluded from the study if mammography revealed focal disease, and/or if further diagnostic tests such as breast ultrasound, breast magnetic resonance or needle biopsy were required. To assess whether classification as Dense Breast (DB) and Non Dense Breast (NDB) was consistent, agreement between the three referents was tested using Cohen's kappa before further statistical analysis. It was decided to divide BIRADS into two categories and BOYD into three categories because absolute frequencies for certain categories were very small and this would create problems in the multivariate analysis.

The sample of 151 women in the case group and 154 in the control group were evaluated as having a 80 % power to detect an OR (Odds Ratio) of 2 at a 0.05 significance level for an exposure (dense breast according to BIRADS score) expected the control group equal to 40 %. The proportional odds were tested using the Brant test. For both groups of cases, the proportional odds assumption was verified (chi square = 2.316 p = 0.678 in the case group; chi square = 4.106 p = 0.392 in the control group).

Continuous variables were presented as mean ± Standard Deviation (SD), and differences were evaluated by the Student's t test or the Mann–Whitney test. Categorical variables were expressed as count and percentages. To assess the association the chi-square test or Fisher's exact test were used as appropriate. Multivariate analyses (ordered logistic regression and binary logistic regression) were used to calculate ORs, including interacting or confounding effects of age, BMI, family history and age at menarche. All tests were two-tailed, and p-value <0.05 was considered significant. All computations were carried out using STATA v.12.