ASPO Abstracts
Breast cancer screening recall rate on digital breast tomosynthesis with synthesized two-dimensional mammography
Category: Early Detection & Risk Prediction
Conference Year: 2019
Abstract Body:
Purpose. Evidence to date suggests that digital breast tomosynthesis (DBT) reduces breastcancer screening recall rates when performed concurrently with conventional 2D digitalmammography (DM). However, many breast imaging centers have now replaced DM withsynthesized 2D mammography (SM); the radiation dose is substantially reduced by exchangingthe DM images with SM images reconstructed from the DBT acquisition. The goal of thisstudy is to evaluate screening recall rates for DBT screening with SM compared to DM/DBTimaging.Methods. We conducted an analysis of breast cancer screening recall rates on DBT examsbefore and after the adoption of SM at the University of Vermont Medical Center and theUniversity of Pennsylvania. We identified 90,403 DM/DBT exams and 100,231 SM/DBT examsconducted between 2011 and 2018. Recall rate was defined as the proportion of examsrecommended for additional imaging due to an abnormal finding. Logistic regression wasused to estimate the association between modality and likelihood of a positive assessmentafter adjusting for known predictors of recall including age, race, body mass index,breast density, family history of breast cancer, baseline vs. subsequent screen, andradiology practice.Results. The raw recall rate on DM/DBT exams was 8.9% at the University of Pennsylvaniaand 7.2% at the University of Vermont Medical Center, with a combined recall rate of 7.9%. On SM/DBT exams, the raw recall rate was 7.0% at the University of Pennsylvania and 7.1%at the University of Vermont Medical Center, with a combined recall rate of 7.0%. In themultivariable model, SM/DBT screening was associated with a 12% reduction in recall rate(odds ratio [OR] = 0.88; 95% confidence interval [CI]: 0.85-0.91). There was astatistically significant interaction according to radiology practice (p<0.001), such thatSM/DBT screening was associated with reduced recall at the University of Pennsylvania(OR=0.77; 95% CI: 0.73-0.81) but not at the University of Vermont Medical Center (OR=0.98;95% CI: 0.94-1.03). Conclusions. Our results suggest that SM adoption may further reduce recall rate forbreast cancer screening with DBT, but that the impact of SM adoption on screening recallvaries across screening centers.
Keywords: breast cancer,screening, digitalbreast tomosynthesis