E1886. Inter-Reader Agreement Regarding Background Parenchymal Enhancement Assessment on Breast MRI
  1. Shabnam Mortazavi; UCLA
  2. Wakana Murakami; UCLA
  3. Tiffany Yu; UCLA
  4. stephanie Lee-Felker; UCLA
  5. Kyung Sung; UCLA
To investigate the inter-reader agreement in evaluation of background parenchymal enhancement (BPE) on breast MRI.

Materials and Methods:
In this IRB-approved, HIPAA-compliant, retrospective study, 431 breast MRIs performed from 2017-2019 were reviewed independently by two breast fellowship-trained radiologists. In accordance with BI-RADS, BPE was assigned as minimal, mild, moderate, or marked to each breast (n = 862 breasts). Weighted kappa statistic was used to assess inter-reader agreement.

We divided the breast MRIs into three subgroups: 197 MRIs in average-risk women without prior breast cancer (Group 1), 196 in BRCA mutation carriers without prior breast cancer (Group 2), and 34 in BRCA carriers with newly diagnosed breast cancer (Group 3). There was substantial agreement between the two radiologists (? = 0.74, 95% confidence interval (CI): 0.71,0.76) with overall agreement of 70.4% for 431 MRIs. Agreement was substantial for all groups: ? 0.75 (95% CI 0.62-0.89) with 71.0% agreement for Group 1; ? 0.74 (95% CI 0.61-0.88) with 73.4% agreement for Group 2; and? 0.67 (95% CI 0.54- 0.80) with 50% agreement for Group 3.

We showed overall high inter-reader agreement regarding BPE on MRI on a qualitative basis. Although we had high inter-reader agreement regarding BPE on breast MRI, a reliable quantitative tool to assess BPE would be beneficial to limit variability.