Tumor size, HER-2 status, CA125, CEA, SII, and PNI: key predictors of pathological complete response in LABC patients

The objective of this study was to identify characteristic factors for pathological complete response (pCR) in patients with locally advanced breast cancer (LABC) undergoing surgery and neoadjuvant chemotherapy (NACT). We retrospectively collected pathological data from 237 LABC patients treated in Affiliated Fuzhou First Hospital of Fujian Medical University from January 2010 to June 2021 and divided them into a training group (n = 166) and a validation group (n = 71) in a 7:3 ratio. A predictive model for pCR was established through logistic regression analysis and evaluated using the receiver operating characteristic (ROC) curve and the area under the curve (AUC). Significant differences between the pCR and non-pCR groups were observed in tumor size (P = 0.001), T stage (P = 0.003), estrogen receptor (ER) (P = 0.031), progesterone receptor (PR) (P = 0.013), human epidermal growth factor receptor 2 (HER-2) (P = 0.001), and molecular type (P = 0.001). The pCR group also had lower levels of carbohydrate antigen 19-9 (P = 0.013), cancer antigen 125 (P = 0.011), carcinoembryonic antigen (CEA) (P = 0.001), and systemic inflammatory index (SII) (P = 0.006), but a higher prognostic nutritional index (PNI) (P = 0.001) compared to the non-pCR group. There were no statistical differences in baseline data between the training and validation groups (P>0.05). Multivariate logistic regression analysis identified tumor size (P = 0.001), HER-2 (P = 0.010), CA125 (P = 0.005), CEA (P = 0.001), SII (P = 0.010), and PNI (P = 0.001) as independent risk factors for pCR. We constructed and visualized a nomogram model that included these 6 factors and developed a dynamic prediction model using the Dynamic Nomogram (DynNom) package. In a random sample of 6 patients, the probability of non-pCR reached 98.8%. The model's AUC was 0.881 in the training group, with a clinical benefit rate of 71.68% and a concordance index (C-index) of 0.881, indicating a good fit. In the validation group, the AUC was 0.722, with a clinical benefit rate of 70.2% and a C-index of 0.722, also indicating a good fit. The Delong test showed a significant difference in AUC between the two groups (P = 0.027). In conclusion, this study constructed and validated a Nomogram model based on clinical pathological features and hematological indicators, finding that higher pCR rates were associated with smaller tumor size, HER-2 positivity, lower levels of CA125 and CEA, lower SII, and higher PNI, significantly enhancing breast cancer management and offering important clinical implications.