标题：Development of a predictive model utilizing the neutrophil to lymphocyte ratio to predict neoadjuvant chemotherapy efficacy in early breast cancer patients
摘要：Neutrophils and lymphocytes are key regulators of breast cancer (BC) development and progression. Neutrophil to lymphocyte ratio (NLR) values have been found to offer clear prognostic utility when evaluating BC patients. In this study, we sought to determine whether BC patient baseline NLR values are correlated with pathological complete response (pCR) following neoadjuvant chemotherapy (NCT) treatment. In total, 346 BC patients underwent NCT at our hospital from January 1, 2014 to October 31, 2019, and data pertaining to these patients were retrospectively analyzed. Correlations between clinicopathological characteristics and pCR rates were assessed via multivariate logistic regression analyses. A predictive scoring model was used to gauge the likelihood of pCR based upon regression coefficient (β) values for each significant variable identified through these analyses. NLR cut-off values suitable for identifying patients likely to achieve pCR following NCT treatment were calculated using receiver operating characteristic (ROC) curves. All patients in the present study were females with a median age of 48 years old (range 22–77). An optimal NLR cut-off value of 1.695 was identified and was associated with respective sensitivity and specificity values of 63.6% and 45.5%. We found that higher NLR values were significantly associated with younger age, premenopausal status, and non-pCR status. Logistic regression analyses indicated that NLR, tumor size, hormone receptor (HR) status, and Ki-67 expression were all independent predictors of pCR. The area under the curve (AUC) for the resultant predictive scoring model was 0.705, and this model was assessed via K-fold cross-validation (k = 10) and bootstrapping validation, yielding respective AUC values of 0.68 and 0.694. Moreover, the incorporation of NLR into this predictive model incrementally improved its overall prognostic value relative to that of a model not incorporating NLR (AUC = 0.674). BC patients with a lower baseline NLR are more likely to exhibit pCR following NCT treatment, indicating that NLR may be a valuable biomarker for BC patient prognostic evaluation and treatment planning. Overall, our results demonstrate that this NLR-based predictive model can efficiently predict NCT efficacy in early BC patients with a high degree of accuracy.
