Breast size and mammographic pattern in relation with age, height, weight, and parity

Abstract

To analyze the relationship between breast size and mammographic pattern, considering the contribution of age, anthropometric data, and parity. After a preliminary concordance study of 300 mammograms, we analyzed 500 successive cases of asymptomatic women sent by their physicians to a Breast Diagnostic Imaging Unit. Patients with prior surgery or radiation of any breast were excluded. Height, weight, age, and parity were recorded for each woman. Every mammogram was assigned a mammographic pattern, and measurements were taken of the width, height, and anteroposterior diameter of mammograms without knowledge of the age or other characteristics of the case. A multiple linear regression of the estimated volume of the left breast was performed, including mammographic pattern, age, parity, weight, height, and body mass index (BMI) as dependent variables. Concordance in the assignment of mammographic patterns between two radiologists was excellent (weighed kappa being 0,83 and 0,86 for right and left breasts respectively). The 500 women studied had a mean age of 52,9 years (boundaries 32-81 years), an average of 2,6 children, and 28% of them had a BMI over 30 and thus could be considered obese. The estimated volumes for both breasts were slightly higher for the left breast, but the difference was not significant (P=0.12) in the paired t test. The variables that remained significant predictors for the volume of the left breast were weight, mammographic pattern, and BMI, with age falling outside significance (P=0.063). Autocorrelation and multicollinearity were rejected in this regression model. Breast volume increases significantly with increasing weight, BMI, and mammographic involution.