Factors affecting surgical decision-making in patients undergoing BRCA1/2 testing at breast cancer diagnosis.

Abstract

Abstract Abstract #1095 Background: The preferred local treatment for women with early stage breast cancer (BC) without a genetic predisposition is breast conservation (BCS). However, in BRCA1/2 mutation carriers, bilateral mastectomy (BLM) at time of diagnosis may be a reasonable alternative. We sought to evaluate the impact that “upfront” genetic counseling and testing (GCT) may have on surgical decision-making in at-risk newly diagnosed BC patients and determine factors that may predict for BLM.
 Methods: 225 consecutive stage 0-3 primary unilateral BC patients underwent GCT for BRCA1/2 germline mutations through the Cancer Genetics service at Dana-Farber Cancer Institute prior to definitive surgical treatment (defined as unilateral mastectomy, BLM or BCS with radiation therapy) from 2004-07. Demographics, BC characteristics, genetic counselor-obtained family cancer history, referring MD, definitive surgery and GCT results were extracted from medical records. Use of contralateral breast MRI as part of diagnostic work-up was assessed; results recorded as normal or abnormal if further imaging or biopsy were recommended (no additional cancers found). Myriad model was used to estimate pre-test probability of BRCA1/2 mutation. Bivariate predictors of surgical treatment were assessed; logistic regression was used to identify independent predictors of BLM.
 Results: Mean age at BC diagnosis was 43.5 years (range 26-68) with TNM stage: 0 in 11.1%, I in 35.1%, II in 38.7% and III in 15.1%. Mean pre-test probability of a deleterious BRCA1/2 mutation was 12.3% (range 2.9%-51.4%). Results revealed BRCA1/2 mutation in 34 (15.1%), variant of unknown significance (VUS) in 14 (6.2%) and no mutation in 177 (78.7%) patients. BRCA1/2 positives were more likely to undergo BLM than patients testing negative (82.4% vs 19.8%; p<0.0001) or VUS (21.4%; p<0.0001). Overall, the strongest independent predictor of BLM was a positive genetic test result (OR 24.4; 95%CI 8.3-71.2). In patients testing negative, BLM was more common in those diagnosed age <40 (30.3% vs 13.5%; p<0.01) and in those with high Myriad scores (36% vs 17.1%; p<0.05). Referral for GCT by surgeon vs non-surgeon was associated with BLM (27.9% vs 15.5%; p<0.05). While having a contralateral breast MRI did not alter rates of BLM (MRI 16.7% vs No MRI 16.7%), an abnormal MRI finding was associated with BLM (42.9%; p=0.001). In multivariate logistic regression, abnormal MRI result was the strongest predictor of BLM in test negative patients (OR 7.5; 95%CI 2.2-25.8).
 Discussion: While a positive genetic test was the strongest predictor of BLM in patients referred for GCT, high rates of BLM were observed even in women with negative results. Since patients are undoubtedly selected for GCT referral, the impact of age at BC diagnosis and family cancer history (Myriad score) may reflect issues of risk perception or communication by patients and providers. The influence of breast MRI on contralateral prophylactic mastectomy decisions in BC patients at diagnosis warrants further investigation. Citation Information: Cancer Res 2009;69(2 Suppl):Abstract nr 1095.