Purpose: This study was designed to examine the relationship between different methodologies for response evaluation and long-term survival estimation in patients underwent neoadjuvant chemotherapy (NCT) for breast cancer.
Methods: We retrospectively analyzed 569 patients who were diagnosed with LABC and received NCT followed by breast and axilla surgery. The RECIST 1.1 criteria and Miller-Payne (MP) grading scale were used to evaluate patient responses to NCT. Univariate and multivariate survival analyses were performed to investigate the correlation between treatment response and long-term patient survival.
Results: Clinical response (RFS [P < 0.001]; OS [P = 0.003]), pathological response evaluated by pCR (RFS [P < 0.001]; OS [P < 0.001]), and MP grade (RFS [P < 0.001]; OS [P < 0.001]) were significant predictors of risks of relapse and survival. However, in hormone receptor-positive (ER and/or PR+) subtypes, the clinical response (P = 0.004 for Luminal-A and P = 0.038 for Luminal-B) and MP grade (P = 0.002 for Luminal-A and P < 0.001 for Luminal-B) significantly predicted RFS independently according to multivariate Cox regression model. MP grade (P = 0.015 for Luminal-A and P = 0.009 for Luminal-B) also was an independent predictor of patients' OS. However, these two methods failed to predict patient survival in hormone receptor-negative (ER and PR-) subtypes.
Conclusions: Our findings indicate that the value of response evaluation methods varies for different breast cancer subtypes. Conceiving of further prospective approaches for new individualized response-evaluation models are needed in the neoadjuvant setting.