Abstract P3-02-07: Predictors associated with MRI surveillance screening in women with a personal history of unilateral breast cancer but without a genetic predisposition for future contralateral breast cancer

Abstract Purpose Women with a personal history of breast cancer (BC), in the absence of genetic predisposition or significant family history, are generally categorized as having an intermediate (15-20%) risk of developing future BC. Screening MRI is recommended as a supplement to mammography for women with a ≥20% lifetime risk. However, models available to calculate risk exclude women with a personal history of BC. Thus, there has existed no reliable mechanism to calculate future risk in these women to refine MRI surveillance recommendations. The Manchester score was recently developed and intended to inform contralateral breast cancer (CBC) risk for surgical decision making. We hypothesized that the score could be informative to guide follow-up imaging recommendations among women with a personal history of BC. Patients & Methods 322 women with newly-diagnosed, non-metastatic, unilateral BC were seen in multidisciplinary breast clinic and underwent unilateral surgery (either breast conserving surgery or mastectomy) at our institution between June 2012 and November 2015. Using life expectancy, family history, genetic mutation status, and endocrine therapy use, we calculated the Manchester score, i.e. CBC risk, for all women. Patients were categorized as low- (<10% CBC lifetime risk), above average- (10-20%), moderate- (20-30%), and high-risk (>30%). We also reviewed the rationale that treating physicians noted for recommending MRI surveillance. Univariate logistic regression analysis (UVA) was used to assess if Manchester score was predictive of MRI surveillance in addition to other known factors. Results In the entire cohort, 75.8% (n=244) were low-risk for CBC, 18.6% (n=60) were above average-risk, 4.3% (n=14) were moderate-risk, and 1.2% (n=4) were high-risk. Using a 20% CBC risk as a threshold for MRI justification, 5.6% (n=18) met indications for MRI surveillance. Among the 21.7% (n=70) undergoing MRI surveillance, 57.1% (n=40) were low-risk, 32.9% (n=23) were above average-risk, 7.1% (n=5) were moderate-risk, and 2.9% (n=2) were high-risk. There was a significant trend for higher rates of MRI surveillance as the risk score increased (odds ratio, OR 1.1, p<0.0001). On UVA, the highest odds for MRI surveillance were in women with a mammographically-occult BC history (OR 9.35, p<0.0001), pre-operative breast MRI use (OR 8.41, p<0.0001), and dense breast tissue (OR 4.88, p<0.0001). The top clinician-endorsed reasons for MRI surveillance were dense breast tissue (61.4%), young age at diagnosis (28.6%), and mammographically-occult BC history (25.7%). Conclusions Although Manchester score was significantly predictive for MRI surveillance, we identified a large subset of women with <20% calculated CBC risk who underwent MRI surveillance (90.0% of the cohort undergoing MRI) and a small subset with >20% CBC risk who continue with mammography alone (4.4%). Overall, concern about poor detection of a future CBC appeared to dominate the selection of surveillance MRI use, even within a largely low-risk population. We believe this calculation could be informative prospectively to select surveillance strategies in women at a high future risk of CBC. Citation Format: Hegde JV, Wang P-C, Kusske A, Hoyt AC, Hurvitz S, McCloskey SA. Predictors associated with MRI surveillance screening in women with a personal history of unilateral breast cancer but without a genetic predisposition for future contralateral breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-02-07.