A novel quantitative prognostic model for initially diagnosed non-small cell lung cancer with brain metastases

Abstract

Background: The prognosis of non-small cell lung cancer (NSCLC) with brain metastases (BMs) had been researched in some researches, but the combination of clinical characteristics and serum inflammatory indexes as a noninvasive and more accurate model has not been described. Methods: We retrospectively screened patients with BMs at the initial diagnosis of NSCLC at Sun Yat-Sen University Cancer Center. LASSO-Cox regression analysis was used to establish a novel prognostic model for predicting OS based on blood biomarkers. The predictive accuracy and discriminative ability of the prognostic model was compared to Adjusted prognostic Analysis (APA), Recursive Partition Analysis (RPA), and Graded Prognostic Assessment (GPA) using concordance index (C-index), time-dependent receiver operating characteristic (td-ROC) curve, Decision Curve Analysis(DCA), net reclassification improvement index (NRI), and integrated discrimination improvement index (IDI). Results: 10-parameter signature's predictive model for the NSCLC patients with BMs was established according to the results of LASSO-Cox regression analysis. The C-index of the prognostic model to predict OS was 0.672 (95% CI = 0.609 ~ 0.736) which was significantly higher than APA,RPA and GPA. The td-ROC curve and DCA of the predictive model also demonstrated good predictive accuracy of OS compared to APA, RPA and GPA. Moreover, NRI and IDI analysis indicated that the prognostic model had improved prediction ability compared with APA, RPA and GPA. Conclusion: The novel prognostic model demonstrated favorable performance than APA, RPA, and GPA for predicting OS in NSCLC patients with BMs.