Abstract
BACKGROUND: The reliability of core needle biopsy (CNB) for HER2-positive breast cancer is well established. However, data on HER2-low and the potential for inconsistencies with surgical samples are limited. MATERIALS AND METHODS: Concordance between CNB and surgical samples was assessed using the unweighted Cohen kappa statistic (Kc) in a consecutive series of 776 treatment-naïve early-stage breast cancer patients. Logistic regression models were used to evaluate the association between concordance and clinico-pathological features. RESULTS: The agreement for HER2-positive status between CNB and surgical specimens was high at 95%, with a Kc value of 0.86 indicating almost perfect agreement. However, 65 of 123 (53%) cases initially classified as HER2-0 were reclassified as HER2 1 + or 2 + /ISH-negative, and 89 of 374 (24%) cases initially classified as HER2 1 + /2 + were HER2-0 in surgical samples. This resulted in a Kc value of 0.22, indicating fair agreement in classifying HER2-0 versus HER2-low breast cancer. Tumor size was a significant factor influencing discordance, with tumors larger than 2 cm having double the risk of misclassification. CONCLUSION: These findings suggest that HER2 status should be retested, particularly for large tumors initially diagnosed as HER2-0, in light of new effective therapies for HER2-low breast cancer, such as antibody-drug conjugates.