RESUMO
To explore the effect of digital breast tomosynthesis (DBT) on the classification of breast imaging-reporting and data system (BI-RADS) and its significance. Methods: A total of 832 patients with breast diseases, who came from Second Clinical College of Fujian Medical University from May 2013 to November 2013, were collected. The patients were examined by double position radiography (including craniocaudol and mediolateral oblique) and COMBO mode [including DBT and full-field digital mammography (FFDM)]. Meanwhile, the results of FFDM and DBT were classified. The number of glands, the characteristics of mass and other indirect signs were compared by COMBO and FFDM modes. Paired Wilcoxon rank sum text was adopted to investigate the differences between COMBO mode and FFDM mode in the 832 patients, and receiver operator characteristic curve (ROC) was applied to analyze the 79 patients with the pathological results. Results: The patients with large amount of glands (including Class c and Class d) accounted for 87.6% in the 832 patients, while the patients with small amounts of glands (including Class a and Class b) accounted for 11.7%. In estimating the content of glands, more details about the distribution of glands were found in the COMBO mode compared with those in the FFDM mode. According to the results of paired Wilcoxon rank sum test, there was significant statistical difference in BI-RADS classcification in breast masses between the COMBO mode and the FFDM mode (P<0.05), though the overall classification of the COMBO mode is higher than that of the FFDM mode. The pathology was served as a standard to estimate the diagnostic efficiency. The area under ROC curve was 0.805 in the FFDM mode, while that in the COMBO mode was 0.941. The optimal sensibility in the COMBO mode was 82.9%, which was higher than that in the FFDM mode. However, the specificity was 93.2% in both COMBO mode and the FFDM mode. Conclusion: DBT has a high clinical significance in BI-RADS classification for breast X-ray examination.