A. Roy1, Z. Zeng3, X. Li6, S. Espino4, Y. Luo3, H. Jiang5, S. Khan4 1Northwestern University,Feinberg School Of Medicine,Chicago, IL, USA 3Northwestern University,Department Of Preventative Medicine, Feinberg School Of Medicine,Chicago, IL, USA 4Northwestern University,Department Of Surgery, Feinberg School Of Medicine,Chicago, IL, USA 5Northwestern University,Department Of Statistics,Evanston, IL, USA 6Harvard School Of Public Health,Department Of Social & Behavioral Sciences,Boston, MA, USA
Introduction: The most sensitive imaging modality currently used to detect multifocal and metacentric breast cancer is magnetic resonance imaging (MRI), yet its role in preoperative evaluation of disease extent remains controversial. It was initially hoped that preoperative MRI use would allow more complete resection of disease and would improve outcomes, but a prospective study has shown that it does not reduce re-excisions and retrospective analyses do not support an improvement in cancer outcomes. However, the number of local recurrences and patients in these studies is small. In an attempt to clarify the potential benefit of pre-operative breast MRI for long-term breast cancer outcomes, we report a retrospective review of data on 3902 women diagnosed with primary breast cancer at the Lynn Sage Breast Center of Northwestern Medicine.
Methods: The Enterprise Data Warehouse of Northwestern Medicine was searched for women diagnosed with ductal carcinoma in situ (DCIS) or invasive breast cancer who underwent breast conservation therapy (BCT) between 2000-2016. The use of preoperative MRI was extracted along with clinical and therapeutic details. The frequencies of local recurrence (LR) and distant recurrence (DR) were evaluated with Cox proportional hazards model, adjusting for age, race, tumor size, tumor grade, lymph node status, ER status, PR status, HER2 status, P53 status, Ki67 status, systemic therapy status, and radiation therapy status.
Results: Among 3902 women with primary breast cancer, 1,303 had preoperative MRI and 2,599 did not. Compared to the women who did not have MRI, women with MRI were younger (55 vs 59 years, p<0.0001), had larger tumor size (1.64 cm vs 1.55 cm, p=0.03), and underwent systemic therapy more frequently (p<0.0001). Median follow-up time for the MRI group was 75 months, and for the non-MRI group was 125 months (p<.0001). Ipsilateral LR was experienced by 224 women (5.74%), and DR occurred in 227 women (5.82%). In univariable Cox regression models, the hazard ratio (HR) with use of MRI was 0.94 (95% CI 0.71 to 1.24; p=0.65) for LR; and 0.84 (95% CI 0.62 to 1.15; p=0.29) for DR. In multivariable Cox regression models, the HR with use of MRI was 0.88 (95% CI 0.65 to 1.17; p=0.37) for LR; and 0.77 (95% CI 0.56 to 1.06; p=0.11) for DR.
Conclusion: Women who received preoperative MRI differed significantly from those who did not, but cancer outcomes for either local or distant recurrence following BCT were not significantly different by MRI use. However, this and other retrospective analyses are likely subject to bias given the factors that appear to drive the use of MRI.
