Accuracy of cone-beam computed tomography, digital mammography and digital breast tomosynthesis for microcalcifications and margins to microcalcifications in breast specimens

Abstract Accurate determination of resection margins in breast specimens is important as complete removal of malignancy is a prerequisite for patients’ outcome. Mammography (DM) as 2D-technique provides only limited value in margin assessment. Therefore, we investigated whether cone-beam computed tomography (CBCT) or digital breast tomosynthesis (DBT) has incremental value in assessing margins to microcalcifications. Three independent readers investigated breast specimens for presence of microcalcifications and the smallest distance to margins. Histopathology served as gold standard. Microcalcifications were detected in 15 out of 21 included specimens (71%). Pooled sensitivity for DM, DBT and CBCT for microcalcifications compared to preoperative DM was 0.98 (CI 0.94–0.99), 0.83 (CI 0.73–0.94) and 0.94 (CI 0.87–0.99), pooled specificity was 0.99 (CI 0.99–0.99), 0.73 (CI 0.51–0.96) and 0.60 (CI 0.35–0.85). Mean measurement error for margin determination for DM, DBT and CBCT was 10 mm, 14 mm and 6 mm (p = 0.002) with significant difference between CBCT and the other devices (p