Purpose: To use digital breast tomosynthesis (DBT)-guided vacuum-assisted biopsy (VAB) to sample target lesions identified at full-field digital screening mammography and compare clinical performance with that of prone stereotactic (PS) VAB.
Materials and methods: In this institutional review board-approved study, 205 patients with 216 mammographic findings suspicious for cancer were scheduled to undergo mammography-guided VAB. Written informed consent was obtained. PS VAB was performed in 159 patients with 165 target lesions. DBT VAB was performed in 46 consecutive patients with 51 target lesions. Tissue-sampling methods and materials (9-gauge needles) were the same with both systems. For calcifications, specimen radiographs were obtained, and for masses or architectural distortions, control mammography or DBT was performed to confirm adequate target lesion sampling. χ(2) and Student t tests were used to compare biopsy time, and the Fisher exact test was used to compare lesion type distribution for DBT versus PS VAB.
Results: Technical success was achieved in 51 of 51 lesions (100%) with DBT VAB versus 154 of 165 lesions (93%) with PS VAB. In one of 11 lesions in which PS VAB failed, DBT VAB was performed successfully. Mean time to complete VAB was 13 minutes ± 3.7 for DBT VAB versus 29 minutes ± 10.1 for PS VAB (P < .0001). Reidentifying and targeting lesions during PS VAB took longer than it did during DBT VAB (P < .0001). Tissue sampling took about the same time for PS VAB and DBT VAB (P = .067). Significantly more "low-contrast" (ie, uncalcified) target lesions were biopsied with DBT VAB (13 of 51 lesions) versus PS VAB (nine of 165 lesions) (P < .0002). No major complications were observed with either system. One patient who underwent DBT VAB in the sitting position and one patient who underwent PS VAB developed self-limiting vasovagal reactions.
Conclusion: Clinical performance of DBT VAB was significantly superior to PS VAB. Because DBT VAB allows use of the full detector size for imaging and provides immediate lesion depth information without requiring triangulation, it facilitates target lesion reidentification and sampling of even low-contrast targets, such as uncalcified masses.