Utilization of preoperative breast MRI among women with newly diagnosed breast cancer.

Abstract

1546 Background: The use of preoperative (pre-op) breast MRI remains controversial. Current practice may rely on patient characteristics and providers’ clinical judgment. This national study examined factors associated with pre-op breast MRI among women with newly diagnosed breast cancer (BC) and explored sources of variations. Methods: We applied the Nattinger algorithm to identify women with incident BC diagnosed between Mar 2008 and Dec 2018 from OPTUM Clinformatics database. Patients who had 26 months of full enrollment, 14 months before and 12 months after the first (index) BC surgery, and no pre-op radiotherapy were included. We defined pre-op MRI as patients who had an MRI between the date of BC diagnosis and date of index surgery. We conducted multivariable logistic regression models to examine factors associated with pre-op MRI and performed separate analyses for elderly (age > = 65) and non-elderly (age < 65) women. Results: 61,865 women (non-elderly: 27,309, elderly: 34,556) were included in the analysis. The crude rate of pre-op MRI increased from 7.4% in 2008 to 14.6% in 2018 (p-value <.001). For the non-elderly, women who were older (adjusted rates: 60-64, 10% vs 20-49, 12.1%), had no distant metastasis (10.6% vs. 12.3% with metastasis), no neoadjuvant chemotherapy (9.9% vs 15.0% with neoadjuvant), and 2 or more comorbidities (9.0% vs. 11.1% with zero comorbidity) were less likely to undergo pre-op MRI (all p-value <.001). Compared to white women (adjusted rate 10.6%), African Americans were more likely to have pre-op MRI (12.7%, p <.001) and Hispanics were less likely (8.14%, P <.001). There was no association between Health Management Organization (HMO) status and receipt of pre-op MRI among non-elderly. For elderly women, older age, more comorbidities, no distant metastasis, and no neoadjuvant chemotherapy were similarly associated with less pre-op MRI use. There was no significant association between race and receipt of pre-op MRI. Moreover, elderly women with HMO insurance were less likely to receive pre-op MRI. In both age groups, we observed wide geographic variations, with significant interaction between census division and HMO enrollment among elderly group only (Table). Conclusions: The use of pre-op MRI nearly doubled from 2008 and 2018. In addition to demographic and clinical characteristics, regional practice pattern variations and insurance type played a key role in the receipt of pre-op MRI among newly diagnosed breast cancer patients. Future study to understand this phenomenon is warranted.[Table: see text]