Comparison of Computerised Assessment of Breast Density with Subjective BI-RADS Classification and Tabar's Pattern from Two-View CR Mammography

Computerised assessment, the objective classification of breast density provides a second opinion to the radiologist in classifying breast density. Subjective classification of breast density by the radiologist involves using: i) Wolfe's classification, ii) Breast Imaging Reporting and Data System (BI-RADS) classification, and iii) Tabar's pattern. The objective of this study was to compare the results of breast density obtained by using a computerised assessment technique with the subjective assessment by BI-RADS classification and Tabar's pattern from two-view computed radiography (CR) mammography. A computerised assessment technique to quantify breast density from two-view CR-mammography was developed using the MATLAB GUI applications that utilise basic MATLAB functionality. Hundred sets of CR-mammograms (from fifty cases) were initially classified by the radiologist (SR) into parenchymal patterns, according to BI-RADS schemes and Tabar's pattern. Median age of the patients was 53.3 years (range of 40-69 years). The radiologist then reanalysed the CR-mammograms using the computerised technique. The correlation between computerised results with BI-RADS classification and Tabar's pattern were analysed respectively. Classification performance of each class and pattern was also analysed. The breast density calculated using the computerised assessment technique correlated well with the subjective estimation of BI-RADS classification (0.82) and Tabar's pattern (0.95). The computerised assessment technique correctly classified 78.3% and 73.2% of the total cases based on BI-RADS classification and Tabar's pattern respectively. This computerised technique can he useful in providing a clinically accurate measurement of breast density and an assessment of future risk of developing breast cancer. This technique can also effectively assist in patient management.