A comparative analysis of BI-RADS and the Japanese Recall Criteria and Categorization in breast ultrasound screening

Abstract Background Japan widely utilizes breast ultrasound screening and the Japanese Recall Criteria and Categorization (JRCC) to identify breast cancer cases, while other countries commonly employ the Breast Imaging Reporting and Data System (BI-RADS). The differences between the two systems have not been thoroughly investigated. Methods This study analyzed 3,113 lesions in 2,000 breast ultrasound images with confirmed diagnoses and compared the accuracy and disparities between JRCC and BI-RADS. Results Among patients who were not recalled under BIRADs, 94 (16.2%) patients were assigned to recall under JRCC whereas among patients who were not recalled under JRCC, 332 (40.6%) patients were assigned to recall. Of the 423 BI-RADS 4a and 528 BI-RADS 4b cases, 264 (62.4%) and 68 (12.8%) would not have been recalled under JRCC, of which 9 (2.1%) and 22 (4.2%) were diagnosed as malignant, respectively. Of the 579 cases classified as BI-RADS 3 or lower, only two (0.35%) were diagnosed with breast cancer, while 33 of the 817 cases (4.04%) classified as JRCC 2 or lower were malignant. The sensitivity and the specificity of BI-RADS were 99.9% and 40.0% while those of JRCC were 98.0% and 54.3%, respectively. While JRCC exhibited higher specificity, it also showed lower sensitivity and false-negative rates. If the threshold for BI-RADS was changed from 4a or higher to 4b or higher, the sensitivity and specificity for diagnosing malignancy would be 99.0% and 68.3%, respectively, surpassing those for JRCC. Conclusions It is important to understand the differences between JRCC and BI-RADS and translate screening results from other countries.