Mammographic breast density classification among Ghanaian population

Abstract

Introduction

Breast density is an established risk factor for breast cancer. However, there is limited literature on breast density evaluation among the African population. This study aimed to evaluate the breast density classification of patients presenting for mammography examination in a selected hospital in Ghana to inform national policy guidelines on breast cancer screening.

Methods

The study analyzed retrospective data of 960 digital screening and diagnostic mammograms of 245 patients from January to December 2020. Two consultant radiologists with over 10 years of clinical experience evaluated mammogram breast density using the fifth edition of the Breast Imaging-Reporting and Data System (BI-RADS) descriptors. Cohen kappa was used in calculating inter and intraobserver reliability for the breast density. Parameters including compressed breast thickness, and age were also recorded.

Results

The majority (62.0 % by Radiologist 1 and 63.3 % by Radiologist 2) of the patients were classified as having ‘almost entirely fatty’ breasts while a few (3.7 % by Radiologist 1 and 4.5 % by Radiologist 2) had ‘extremely dense’ breasts. Most of the patients aged 50–59 years were found to have ‘almost entirely fatty’ breasts. In general, the interobserver and intraobserver reliability showed an almost perfect agreement for the breast density (κ = 0.844, κ = 0.892). However, the interobserver reliability agreement varied from substantial to almost perfect score, with significant p-values, within each BI-RADS category.

Conclusion

The radiologists demonstrated consistency in both interobserver and intraobserver reliability with an almost perfect agreement for the breast density categories.

Implications for practice

The consistency observed in the BI-RADS assessment will guide future practice to develop models for breast cancer risk stratification and decision making for asymptomatic patients.