Effect of arrival of prior mammograms on recall negation for screening mammograms performed with digital breast tomosynthesis in a clinical setting

Abstract Purpose This retrospective study evaluates the effect of comparison with prior mammograms on recall negation for screening mammography performed with digital breast tomosynthesis (DBT) in a clinical setting and compares it with that performed without DBT. Methods This is an Institutional Review Board–approved, HIPAA-compliant retrospective review of the electronic medical record for all nonbaseline screening mammograms performed in clinical practice over 13 months. For each mammogram, we recorded if DBT were used, the BI-RADS assigned at initial interpretation, and whether prior mammograms were available at initial interpretation. If prior mammograms arrived later for comparison, the final BI-RADS assigned after comparison was recorded. A mammogram assigned a BI-RADS 0 at initial interpretation and assigned a BI-RADS 1 or BI-RADS 2 after prior mammograms arrived for comparison was labeled as a recall that was negated by the arrival of prior mammograms. The number of recalls negated for mammograms that used DBT was compared with that for mammograms that did not use DBT. Results Arrival of prior mammograms for comparison negated the need for recall for mammograms performed with DBT by 67.67% and negated the need for recall for mammograms performed without DBT by 55.80%. After adjusting for age, density, and time between mammograms, the percentage of recalls negated by comparison with prior mammograms was not significantly different for mammograms performed with DBT than it was for those performed without DBT. Conclusion Comparison with prior mammograms remains important for the minimization of recall rates during the use of DBT for screening mammography in the clinical setting.