Factors Associated With MRI Screening in Women With a Personal History of Invasive, Unilateral Breast Cancer Predominately Encompass Detectability Concerns Rather Than Future Breast Cancer Risk

Abstract

For women with a personal history of invasive breast cancer, no validated mechanisms exist to calculate the risk of developing a future contralateral breast cancer (CBC). Accurately predicting future CBC risk would help refine MRI surveillance screening recommendations. The Manchester risk stratification model was recently developed to evaluate CBC risk in this population, primarily for surgical decision making. We hypothesized that this model may be informative for the use of MRI surveillance screening as CBC risk is an assumed consideration for high-risk surveillance. Two hundred fifty one women with newly-diagnosed, non-metastatic, unilateral invasive breast cancer who were seen in our institutional multidisciplinary breast clinic (MDC) and underwent unilateral surgery (breast conserving surgery or mastectomy) between June 2012 and November 2015 comprise the study cohort. We calculated lifetime CBC risk using the Manchester model, which incorporates major known factors for CBC, including age, family history, genetic mutation status, estrogen receptor positivity, and endocrine therapy use. Patients were then categorized as having a low (<10%), above average (10-20%), moderate (>20-30%), or high (>30%) risk for developing a future CBC. Univariate and multivariate logistic regression analyses (UVA/MVAs) were performed to evaluate whether the CBC risk was predictive of MRI surveillance while adjusting for other factors used to support MRI surveillance. For the 22% (n=55) undergoing MRI surveillance, 66% were low-, 23% were above average-, and 11% were moderate-/high-risk. Factors associated with MRI surveillance on MVA included previous mammography-occult breast cancer (OR 18.95, p<0.0001), endocrine therapy use (OR 3.89, p=0.009), dense breast tissue (OR 3.69, p=0.0007), mastectomy vs. lumpectomy (OR 3.12, p=0.0041), and CBC risk as calculated by the Manchester model (OR 3.17 for every 10% increase, p=0.0002). No pathologic factors increasing the risk of ipsilateral breast cancer recurrence were significant on MVA. However, if CBC risk was removed as a predictor, age becomes associated with the use of MRI surveillance on MVA (OR 1.77 for every 10 years decrease, p=0.0006). Although CBC risk was predictive for MRI surveillance, 89% of the invasive cohort undergoing MRI had < 20% calculated risk for CBC. Other concerns related to future breast cancer detectability (dense breasts and/or previous mammography-occult disease) predominate decision making. Pathologic factors important for determining ipsilateral recurrence risk, aside from age, do not appear associated with selecting MRI surveillance.