RETURN TO ABSTRACT LISTING


Objective:
The aim of this study is to compare patients recalled from screening with 2D digital mammography versus tomosynthesis.

Materials and Methods:
An IRB-approved HIPPA-compliant retrospective review was conducted on women recalled from screening mammograms during the period of January 2012 to March 2020. Institutional database and EMRs were reviewed for mammographic findings, patient follow-up, cancer detection, breast cancer risk and demographic variables. Differences in age between 2D and 3D mammography groups were evaluated with an independent-samples t-test, and data were summarized with means (M) and standard deviations (SD). Cancer Risk (%) differences between 2D and 3D groups was evaluated with an independent samples Mann-Whitney U test, and data were summarized with medians (Mdn), median absolute deviations (MAD), and range. Associations between mammography type (2D, 3D) and race, finding, BI-RADS, follow-up results, and cancer outcomes were evaluated with chi-square tests. Statistical significance was set at the <em>a</em> = .05 threshold, 2-sided p-values were reported, and statistical analyses were conducted with SPSS v.28 (Armonk, NY: IBM).

Results:
During the study period, a total of 2190 women were recalled from screening mammography. There were 1622 patients (74.1%; mean age = 55.51 years, SD = 10.51) called back from 2D mammography and 568 patients (25.9%; mean age = 58.42 years, SD = 10.39) called back from 3D mammography. The distribution of race was significantly different for 2D and 3D mammography groups, with 2D mammography composed of 45.3% Black patients versus 31.2% for the 3D group, and the 3D group composed of 65.1% White patients versus 50.6% in the 2D group, <em>p</em> < .001. The type of finding varied significantly by mammography type with 28.7% of 2D patients having a finding of an asymmetry versus 18.5% of 3D patients, 3.7% of 2D patients having a finding of architectural distortion versus 12.9% of patients, and 15.9% of 2D patients having a finding of mass versus 24.1% of 3D patients, <em>p</em> < .001. There was no significant difference in final BI-RADS assessment, <em>p</em> = .36, adherence to follow-up, <em>p</em> = .29, or cancer grade, <em>p</em> = .06. Patients who underwent 3D mammography (Mdn = 8.4%, MAD = 3.1%, range = .2% - 44.5%) had significantly higher cancer risk scores than patients who underwent 2D mammography (Mdn = 7.7%, MAD = 3.0%, range = 0.1% - 52.1%), <em>p</em> = .10.

Conclusion:
Significantly more patients were recalled for an asymmetry in the 2D group, and significantly more patients were recalled for architectural distortion and mass in the 3D group. The call-back rate was higher in the 2D group compared to the 3D group. This study also highlights a disparity between patients receiving 2D versus 3D screening with a larger number of Black women receiving 2D exams. Patients called back from 3D screening had a higher lifetime risk of developing breast cancer compared to the 2D group. There was no significant difference between final BI-RADS assessment, adherence to follow up, cancer detection rates or cancer grade.