Evaluation of Concordance of Breast Imaging Reporting and Data System with Interpretation of Digital Mammography

Abstract

Mammography is a valuable tool in the early detection of breast cancer, the American College of Radiology’s (ACR) Breast Imaging Reporting and Data System (BI-RADS) provide a lexicon that give a standard way of reporting mammogram that help constant reporting and ensure better follow up of benign and suspicious findings. The research conduct to assess the concordance of final findings of mammograms reported using BI-RADS lexicons and the final impression of radiologist those using routine interpretation of mammogram. The research was retrospective study design, and included 300 mammographic image for women aged between 15 to 90 years old, the mammograms obtained by digital mammography machine and used the basic mammographic projections (CC: Craniocaudal and MLO: Mediolateral Oblique).The data descriptively analyzed and measured the different variables in the research and determined the relationship between them, Kappa value was determined to measure the concordance. Regarding to age, the most abundant group was (46-60 years) as (40%).Regarding to Breast composition, the most abundant group was scattered fibroglandular breast (b) group for both routine interpretation and BIRAD.The result appeared as the concordance of the breast composition was (k = 0.5) that point Moderate agreement. All pathological findings were founded and the concordance among the descriptive terms measured, According to Mass, the presence of the mass was (k = 0.83) that was a Very Good agreement, the shape of the mass was (k = 0.53) that a Moderate agreement, the margin of the mass was (k = 0.475) that point Moderate agreement, the density of the mass was (k = 0.74) that point a good agreement, in addition the associate calcifications was (k = 1) that point a total agreement. According to Calcification the IV Kappa value was for presence (k = 0.87) that point to Very Good agreement, the morphology of the calcifications was (k = 0.85) that point to Very Good agreement and distribution was (k = 1) that point Very Good agreement. In addition, the Architectural distortion was (k = 0.85) that point to Very Good agreement, and A symmetry was (k = 1.3) it was poor agreement. The intramammary lymph node was (k = 0.65) that point a good agreement and the overall agreement and concordance to the Final finding (BIRAD Categories) was (k = 0.58) that point to a moderate agreement. The research found a significant relationship between the different variables such as breast composition with age as (Pvalue = 0.000), final findings with age(P-value = 0.000), mass presence (Pvalue = 0.000), mass shape (P-value = 0.000), mass margin (P-value = 0.000), mass density (P-value = 0.000), mass associated calcification (P-value = 0.000), mass location (P-value = 0.000), presence of calcification (P-value = 0.000), architectural distortion (P-value = 0.001), presence of asymmetry (Pvalue = 0.000), presence of inflammatory lymph node (P-value = 0.000), presence of associated features (P-value = 0.000) and presence of solitary dilated duct as (P-value = 0.000). As the conclusion to the research, there is a moderateagreement with radiologist when reporting mammogram, which mean some variation in wording of the mammography report between the radiologists. Therefore, it is very importance to the similarity wording of the mammography report to avoid any misunderstanding or confusion, and following of BIRAD lexicon leads to this goal.