Non-calcified ductal carcinoma in situ of the breast: comparison of diagnostic accuracy of digital breast tomosynthesis, digital mammography, and ultrasonography

To retrospectively compare the diagnostic accuracy of digital breast tomosynthesis (DBT), digital mammography (DM), and ultrasonography (US) in non-calcified ductal carcinoma in situ (DCIS, include DCIS with micro-invasion).

Ninety-eight patients with non-calcified DCIS (include DCIS with micro-invasion) were enrolled in our study. Breast carcinoma in situ was confirmed by surgical pathologic evaluation. Our Institutional Review Board granted approval and the participating women provided written informed consent. The imaging findings were evaluated according to the Breast Imaging Reporting and Data System (BI-RADS) of the American College of Radiology (ACR) by comparing the differences in the detection rate and diagnostic accuracy among the three techniques in all cases, in dense breasts, and in non-dense breasts.

The detection rates of DBT, DM, and US for non-calcified DCIS in all cases were 83.7, 68.4, and 94.9%, respectively, and in patients with dense breasts were 81.2, 63.8, and 95.0%. The detection rate of US was higher than DBT, which, in turn, was higher than DM both in all cases and in dense breasts. Pairwise comparisons among the three techniques showed that the differences were statistically significant (P = 0.000 and P = 0.000, respectively). The experts identified a case as abnormal for all criteria (BI-RADS score of 4B-5) in 68.4% of ratings using DBT, 43.9% of ratings using DM, and 66.3% of ratings using US; for dense breasts, the positive identification rates were 62.5% of ratings using DBT, 41.2% of ratings using DM, and 61.2% of ratings using US. The diagnostic accuracy of DBT and US was significantly higher than that of DM in all cases (P = 0.001 and P = 0.006, respectively) and in dense breasts (P = 0.007 and P = 0.011, respectively). The diagnostic accuracy of DBT was slightly higher than US in all cases and in dense breasts, but the difference was not statistically significant (P = 0.761 and P = 0.871, respectively). By DBT, most non-calcified cases of DCIS presented as a mass lesion (54.9%) with an irregular shape (46.7%), indistinct margin (53.3%), and isodense composition (71.1%). Using US, 72 of 93 patients (77.4%) were shown to have a mass. Most mass lesions had an irregular shape (83.3%), indistinct margin (55.5%), and parallel the skin (82.8%).

DBT and US gave better detection rates and diagnostic accuracy for non-calcified DCIS compared with DM in all cases and in dense breasts. The detection rate of DBT was lower than that of US in all cases and in dense breasts. The diagnostic accuracy of DBT was slightly higher than that of US in all cases and in dense breasts, but the difference was not statistically significant. Imaging findings for non-calcified DCIS were relatively non-specific.