Purpose: To evaluate interobserver agreement in full-field digital mammography (FFDM) and digital breast tomosynthesis (DBT) in terms of both lesion detection and characterization, and to evaluate the cancer detection rate of standard two-view FFDM compared to various combinations of DBT.
Materials and methods: Thirty-five women (mean age 59.7; range 50-80 years) with 37 breast cancers who underwent both two-view DBT and two-view FFDM were included. DBT images were obtained using an investigational prototype. We performed interobserver agreement analyses using kappa (k) statistics. The cancer detection rate of various combinations of DBT compared to standard two-view FFDM was estimated using a generalized estimation equation.
Results: There was fair to moderate agreement on detectability (k = 0.59-0.62) in both views of FFDM and DBT, while fair to substantial agreement was found for lesion location (k = 0.52-0.84) and fair to moderate agreement for lesion type (k = 0.46-0.70) and BI-RADS final assessment (k = 0.48-0.69). In generalized estimation equations, standard two-view FFDM was inferior to any combination of DBT. The detection rate ratio was significantly higher in the combined four views of DBT and FFDM compared to standard FFDM (p < 0.046).
Conclusion: Our study showed good agreement in lesion detection and characterization between FFDM and DBT images. Our findings also demonstrated that combining DBT and FFDM is superior in detecting cancer compared to standard FFDM.
Keywords: Breast; Comparison; Digital breast tomosynthesis; Full-field digital mammography.