PB.24: Is digital breast tomosynthesis helpful within a symptomatic one-stop breast clinic for characterisation of subtle findings?

Introduction

The study investigated the accuracy of two-view full-field digital (2D) mammograms (FFDM) by comparing its performance with two-view digital breast tomosynthesis (3D) plus FFDM combination in a symptomatic setting.

Methods

A multi-case multi-reader study was conducted involving four imagers of varying experience. A total of 109 lesions from 103 patients who attended symptomatic breast clinics during a 7-month period between March 2012 and September 2012 were retrospectively read. All patients who had subtle signs on 2D images were retrospectively double read in a free response study and the findings recorded. The performance quality of the methods was evaluated using receiver operating characteristic (ROC) curves. Reader agreement (k values) was calculated and the area under curve (AUC) was compared between 2D imaging and 2D+3D imaging, regarding both readers and cases as random events.

Results

There was more inter-reader agreement between the two readers with 2D+3D combination (k = 0.391) compared with 2D images alone (k = 0.153). In total, 85.19% of M3 mammograms on 2D imaging were changed to 11.01% M3 and 77.06% M2 on combination of FFDM and DBT. There was more correlation between the mammographic scores and the final result for combination mode compared with 2D images alone (P = 0.0001). ROC analysis revealed the AUC for the combination (0.898) was significantly greater than 2D mammograms alone (0.734; P = 0.009).

Conclusion

DBT increases diagnostic sensitivity in a symptomatic setting and reduces the number of M3 mammograms, when used as an adjuvant to 2D images. Thus DBT has the potential of increasing work-flow efficiency in a symptomatic setting by reducing benign biopsies.