External verification and improvement of the Neo-Bioscore staging system in a Chinese cohort

BACKGROUND: Accurately predicting outcomes for patients with breast cancer receiving neoadjuvant chemotherapy (NAC) is critical for clinical decisions. Prognostic models applicable to the Chinese population remain limited. The Neo-Bioscore staging system has been utilized as a predictive model for survival of breast cancer patients after NAC. This study aimed to validate the applicability of Neo-Bioscore in a Chinese population and develop an improved staging system based on it to predict prognosis of Chinese patients more accurately. METHODS: This study retrospectively collected clinicopathological and survival data in patients receiving NAC from February 2005 to August 2018 in PLA General Hospital. Discrimination, calibration and clinical usefulness were used to assess model performance. Univariate and multivariate analyses assessed relationships between clinicopathological factors and disease-specific survival. For model modification, postoperative pathological staging in the Neo-Bioscore was substituted with the posttreatment pathological tumor (ypT) stage and posttreatment pathological lymph node (ypN) stage. Neo-Bioscore and Modified Neo-Bioscore were compared with the American Joint Committee on Cancer (AJCC) staging system. RESULTS: A total of 436 patients with a median follow-up of 67 months were included. Five-year disease-specific survival (DSS), overall survival, and disease-free survival rates were 88.0%, 87.9%, and 76.8%, respectively. The concordance index (C-index) of the Neo-Bioscore staging system, posttreatment pathological stage (PS), and pretreatment clinical stage (CS) for DSS were 0.78 [95% confidence interval (CI): 0.72–0.83], 0.75 (95% CI: 0.69–0.82), and 0.68 (95% CI: 0.62–0.74), respectively. No significant difference between the Neo-Bioscore and PS was observed in the C-index (P=0.399). ypT and ypN were included in Neo-Bioscore to replace PS and create a modified staging system named MNeo-Bioscore. The C-index for DSS of the MNeo-Bioscore was 0.82 (95% CI: 0.78–0.87), and the calibration curve and decision curve analysis (DCA) curve performed well in internal validation. CONCLUSIONS: The Neo-Bioscore staging system provided precise prognostic stratification for Chinese breast cancer patients receiving NAC; ypN and ypT stage may be substituted for PS to add significant prognostic value for Neo-Bioscore.