Comparison of Cancers Detected by Screening Breast Ultrasound and Digital Breast Tomosynthesis.

PURPOSE: Review of outcomes of screening patients imaged with both digital breast tomosynthesis (DBT) and screening ultrasound (US) to compare the cancer detection rates and characteristics of cancers detected by the imaging modalities.

MATERIAL AND

METHODS

This retrospective study reviewed a total of 24,787 screening US exams performed in the time period of January 2013 through December 2017. These exams were in patients with heterogeneously dense or extremely dense breast tissue. In this population, 21,220 (86%) had DBT screening mammography. These cases were further reviewed to identify any pathology-proven malignancy detected with US and/or DBT.

RESULTS

The study cohort consisted of 115 breast cancers in patients having screening US and DBT. Of the 115 cancers, 100 were invasive cancers and 15 were ductal carcinoma in situ: 64/115 were seen on DBT, 9 of which were seen only on DBT, and 106 were seen on US, with 51 seen only on US. The cancer detection rate of DBT only was 0.4/1000 (9/21,220) and 3.0/1000 (64/21,220) for those detected on DBT whether with or without additional US, with detection on US only having an incremental cancer detection rate of 2.4/1000 (51/21,220) above DBT detected malignancies. Differences in DBT-detected lesions and US only lesions when comparing median lesion size, lesion type, tumor type (invasive vs noninvasive) and tumor stage were statistically significant (p = 0.0045, p = 0.0113, p = 0.0003, and p = 0.0153, respectively).

CONCLUSION

In review of the outcomes of a screening US program, we found a similar number of breast cancers were detected by DBT and US, and US alone (47.8% vs 44.3%, respectively). Ninety-six percent of the cancers detected by US alone were invasive; 89% of those seen on both modalities were invasive, while most of the breast cancers seen on DBT only were in situ carcinoma. Statistically significant differences between DBT and US, and US alone were found for many lesion characteristics including lesion size, type, tumor size, and tumor stage.