Association of MRI and locoregional recurrence (LRR) rates in ductal carcinoma in situ (DCIS) patients treated with or without radiation therapy (RT).

Abstract

57 Background: Perioperative MRI is frequently obtained in women with breast cancer; however, studies have not shown decreased rates of re-excision, and some report unnecessary increases in mastectomy rates. We examined LRR rates among women with DCIS who underwent perioperative MRI as compared to those who did not. Methods: All women who underwent breast-conserving surgery for DCIS in 1997-2010 were included from a prospectively maintained database. Patient characteristics and rates of LRR were compared in women with and without an MRI. Univariate and multivariate analyses were performed. Analysis was repeated in the subset of women who did not receive RT. Results: 2,321 cases were identified; 596 had MRI and 1,725 did not. Women who had MRI were younger, more likely to be premenopausal, have a family history of breast cancer, have a clinical presentation, receive RT and endocrine therapy, be treated in later years, and had fewer close/positive margins. At median follow-up of 57 months there were 184 IBTRs; 5-year LRR rates were 8.5% (MRI) and 7.2% (no MRI) (p = 0.52), and 8-year rates were 14.6% and 10.2%, respectively. MRI was not associated with lower LRR rates after adjustment for age, menopausal status, family history, presentation, adjuvant therapy, margin status, number of excisions, and year of surgery in both the entire cohort and in the subgroup who did not receive RT. Select factors from multivariate analysis for patients with all covariates available are shown in the Table. Conclusions: We observed no association between perioperative MRI and LRR rates for patients with DCIS, even when RT was not given. The benefit of perioperative MRI for DCIS remains uncertain. [Table: see text]