Background: The aim of neoadjuvant chemotherapy is to increase the likelihood of successful breast conservation surgery (BCS). Accurate identification of BCS candidates is a diagnostic challenge. Breast Cancer Index (BCI) predicts recurrence risk in estrogen receptor+lymph node-breast cancer. Performance of BCI to predict chemosensitivity based on pathological complete response (pCR) and BCS was assessed.
Methods: Real-time RT-PCR BCI assay was conducted using tumor samples from 150 breast cancer patients treated with neoadjuvant chemotherapy. Logistical regression and c-index were used to assess predictive strength and additive accuracy of BCI beyond clinicopathologic factors.
Results: BCI classified 42% of patients as low, 35% as intermediate and 23% as high risk. Low BCI risk group had 98.4% negative predictive value (NPV) for pCR and 86% NPV for BCS. High versus low BCI group had a 34 and 5.8 greater likelihood of achieving pCR and BCS, respectively (P=0.0055; P=0.0022). BCI increased c-index for pCR (0.875-0.924; P=0.017) and BCS prediction (0.788-0.843; P=0.027) beyond clinicopathologic factors.
Conclusions: BCI significantly predicted pCR and BCS beyond clinicopathologic factors. High NPVs indicate that BCI could be a useful tool to identify breast cancer patients who are not eligible for neoadjuvant chemotherapy. These results suggest that BCI could be used to assess both chemosensitivity and eligibility for BCS.