Abstract
OBJECTIVE: Breast cancer (BC) biomarkers, such as hormone receptors expression, are crucial to guide therapy in BC patients. Antiandrogens have been studied in BC; however, limited data are available on androgen receptor (AR) expression test methodology. We aim to report the core needle biopsy (CNB) accuracy for AR expression in BC. METHODS: Patients diagnosed with stage I-III invasive BC from a single institution were included. Androgen receptor expression was evaluated by immunohistochemistry (IHC) using 1 and 10% cutoff and the AR expression in surgical specimens (SS) was the gold standard. Kappa coefficients were used to evaluate the intraprocedural agreement. RESULTS: A total of 72 patients were included, with a mean age of 61 years old and 84% were Luminal A or B tumors. The prevalence of AR expression in all BC samples was 87.5% using a cutoff ≥ 10% in SS. With a cutoff value ≥ 1%, CNB had an accuracy of 95.8% (Kappa value = 0.645; 95% confidence interval [CI]: 0.272-1.000; p < 0.001) and 86.1% (Kappa value = 0.365; 95% CI: 0.052-0.679; p < 0.001) when ≥ 10% cutoff was used for AR positivity. Androgen receptor expression in CNB (cutoff ≥ 1%) had a sensitivity of 98.5%, specificity of 60%, positive predictive value of 97.0%, and a negative predictive value of 76.9% in the detection of AR expression in SS. CONCLUSION: Core needle biopsy has good accuracy in evaluating AR expression in BC. The accuracy of CNB decreases with higher cutoff values for AR positivity.