E4679. Comparing the Performance of a Wide-Angle and Narrow-Angle Tomosynthesis System on Screening Mammography in Community Outpatient Setting
Authors
Abid Irshad;
Medical University of South Carolina
Ian Smith;
Medical University of South Carolina
Brielle Paolini;
Medical University of South Carolina
Madelene Lewis;
Medical University of South Carolina
Rebecca Leddy;
Medical University of South Carolina
Leonel Vasquez;
Medical University of South Carolina
Abbie Cluver;
Medical University of South Carolina
Objective:
This study aims to evaluate if tomosynthesis systems using different arc-angles and different numbers of image slices through the breast have an impact on overall cancer detection, early-stage cancer detection, positive axillary node rates, and recall rates.
Materials and Methods:
Six years of 3D screening mammography data (January 1,2016–December 31, 2022) were collected after IRB exemption from two different outpatient centers that utilized different equipment systems, one using wide-angle technique (Siemens Inc.) and the other using narrow-angle technique (Hologic Inc.) for obtaining tomosynthesis images. The performance of these units was compared by retrieving the overall anonymized pathology data through our Penrad Imaging mammography reporting system. The data for both systems was compared for overall cancer detection rate (CDR), detection of minimal cancers (DCIS or < 1.0 cm), rate of positive axillary nodes, and the radiologists’ callback rates. The same group of 6–8 radiologists read mammograms from both centers. Chi-square statistics were used for the comparative analysis, and p-value was considered significant at p<0.05.
Results:
A total of 24,589 3D mammograms were performed on the wide-angle tomosynthesis system, which detected total of 131 cancers (CDR = 5.3/1000), of which 92 (70.2%) were minimal cancers. Out of 59 patients with invasive cancers who had axillary node sampling, 9 (15.3%) of these showed positive axillary nodes. The radiologists reading mammograms from wide-angle system had 2666 callbacks (callback rate of 10.8%). On the other hand, 18,081 mammograms were performed on the narrow-angle system, which detected 76 total cancers (CDR = 4.2/1000), of which 56 (73.6%) were minimal cancers. Out of 32 patients with invasive cancers who had axillary node sampling, 7 (21.9%) of these showed positive axillary nodes. Radiologists reading mammograms on the narrow-angle system showed 1947 callbacks (callback rate of 10.8%). Comparative analysis showed no significant difference between the performance the two systems for overall cancer detection (x²=2.70, p = .100), detection of minimal cancers (x² = 1.24, p = .265), cancers with positive axillary nodes (x² = 0.31, p = .575), or callback rates (x²=0.05, p= .827).
Conclusion:
The comparative performance of 3D mammography (tomosynthesis) with systems using wide-angle versus narrow-angle techniques showed no significant difference in overall cancer detection, detection of minimal cancers, axillary node positivity rates, or radiologists’ callback rates.
