Abstract 3363: Clinicopathologic factors associated with surgical margins in primary invasive breast cancer

Background: For breast cancer patients undergoing lumpectomy, margin status remains an important risk factor for local recurrence. Positive margins mandate re-excision, and some patients may require multiple resections and mastectomy to obtain negative margins. In this study, association of various clinicopathologic factors with positive surgical margins for primary invasive breast cancer was assessed using both public and in-house datasets. Method: The datasets were obtained from The Cancer Genome Atlas-Breast Cancer (TCGA-BC) and Clinical Breast Care Project (CBCP). The proportion of positive to negative margins in both TCGA-BC (3.86%, 38/984) and CBCP (3.78%, 15/397) was similar. The role of each clinicopathologic factor for margin prediction was evaluated using logistic regression, Fisher’s exact test and Cox proportional hazards models. Results: The univariable and multivariable logistic regression using both TCGA-BC and CBCP data sets showed that higher tumor stage, tumor size and positive lymph nodes significantly contributed (p < 0.05) to positive surgical margins. In agreement with the results of logistic regression, Fisher’s exact test also showed the association of tumor stage with margin status. Furthermore, PAM50 subtype normal-like was significantly associated (p = 0.044) with positive margins while Her2 subtype was close to significance (p= 0.062) in TCGA-BC. However, they were not significant in the multivariable model. In addition, young patients also showed close significance to positive margins (p = 0.054) in the multivariable model. The bivariable cox hazards model demonstrated that margin status along with PAM50 subtype significantly associated with progression of the disease in TCGA-BC. On the other hand, margin status along with stage or TNM (T:Tumor size, N: Lymph Node Status, M: Metastasis) was insignificant. This shows that margin status was surrogate to tumor stage or TNM. Conclusion: This characterization study showed that chances of obtaining positive margins increases with higher tumor stage, tumor size and positive lymph nodes. Subtype and age also trended to affect margin status. If validated in a dataset of a larger number of cases with positive margin, our results may provide an additional perspective of how the margin status is associated with the clinical outcome of breast cancer patients which may further impact care provider’s decision making for re-excisions. Disclaimer: The contents of this publication are the sole r