Return To Abstract Listing

Scientific Session 03 — Breast - Digital Breast Tomosythesis

Monday, May 1, 2017

Abstracts 1108-2863



2362. Comparing Call Back Characteristics Between Digital Breast Tomosynthesis and 2D Mammography

Chan T1*,  Schetter S1,  Zhu J2,  Chetlen A1 1. Penn State Milton S. Hershey Medical Center, Hershey, USA ; 2. Penn State College of Medicine, Hershey, PA

Address correspondence to T. Chan (tchan@hmc.psu.edu)

Objective: The purpose of our study was to determine if digital breast tomosynthesis (DBT) in conjunction with full field digital mammography (FFDM) detects a different distribution of BI-RADS category 0 lesions when compared with FFDM alone. Our study evaluates these findings on a technology not previously published to our knowledge.

Materials and Methods: This retrospective review was approved by the institutional review board. Methodology was in compliance with federal HIPAA regulations. We searched the institutional database from December 1, 2013, to December 3, 2015, for DBT-FFDM and FFDM radiology reports with a BI-RADS category 0 assessment. Pregnant patients and those with prior diagnoses of breast cancer were excluded. Our search yielded 435 radiology reports corresponding to 205 patients (age range, 37–86 years) who underwent DBT-FFDM and 184 patients (age range, 53–88 years) who underwent FFDM. All patients were imaged on the SenoClaire DBT platform (GE Healthcare). We reviewed 435 electronic medical records corresponding to 389 patients. Age, breast density, and the BI-RADS category 0 findings of mass, focal asymmetries, asymmetries, calcifications, architectural distortion, and axillary adenopathy leading to recall were recorded in a spreadsheet. To evaluate age difference between the two patient groups, the nonparametric two-sample Mann-Whitney U test was performed. To compare the distribution of BI-RADS category 0 findings and breast density within each patient group, a Fisher exact test was performed. A multivariate logistic regression model evaluated the relationship between BI-RADS category 0 findings and group while controlling for breast density and age. Statistical significance level was set as p = 0.05.

Results: There was no significant difference in age (p = 0.1068) between patients in the DBT-FFDM and FFDM groups. The mean and median ages were 56.4 and 55.0 years in the DBT-FFDM group and 54.6 and 53.0 years in the FFDM group, respectively. There was no significant difference in breast tissue density distribution (p = 0.5417) between the two groups. There was a significant difference (p < 0.0001) in the distribution of BI-RADS category 0 findings detected on DBT-FFDM compared with FFDM, independent of age or breast density. Masses (36.0%) were the second most commonly recalled finding on DBT-FFDM but accounted for only 8.3% of findings on FFDM. Asymmetry was the most commonly recalled finding in both groups, though a smaller proportion of DBT-FFDM recalls were due to this finding when compared with FFDM (43.0% vs 68.9%). The remaining findings were each responsible for a similar proportion of patient recalls in both groups.

Conclusion: DBT allows breast tissue lesions that would otherwise be categorized as asymmetries to be characterized as masses. These results may contribute to the documented 29% increase in breast cancer detection following implementation of tomosynthesis. Data published to date on this topic have been derived from a different vendor of tomosynthesis technology. Because the Food and Drug Administration considers each vendor’s tomosynthesis product as a different technology, it is important to verify these results across all platforms.