A long-term survival prediction model for non-small cell lung cancer based on blood inflammatory biomarkers

Lung cancer with the highest incidence and mortality rates of malignant tumors in the world, has become one of the greatest economic burden diseases in China. To construct a simple, efficient, and practical prognostic model for lung cancer can accurately guide the development of treatment plans for patients without additional expenses, which is beneficial for reducing medical costs and improving patients’ quality of life. Clinical characteristics and blood inflammatory biomarkers (BIBs) of 501 patients diagnosed with NSCLC for the first time between June 2004 and December 2016 were collected. Group analyses were conducted on clinical characteristics and BIBs. Relevant factors for overall survival were selected using LASSO regression, and a prognostic model based on BIBs was established. Kaplan–Meier analysis indicated statistically significant differences in lung cancer survival among subgroups of gender, smoking, pathological type, TNM stage, surgical status, and Performance Status (PS) score (all P < 0.05), whereas no significant differences were observed among age and Charlson Comorbidity Index (CCI) subgroups (all P > 0.05). Statistically significant differences in lung cancer survival were also observed among subgroups of various BIBs (all P < 0.05). The BIBs prognostic model demonstrated good predictive value with AUCs of 0.811, 0.891, 0.856, and 0.841 for 1-year, 3-year, 5-year, and 10-year survival, respectively, effectively predicting NSCLC prognosis. The prognostic model based on clinical characteristics and BIBs exhibit simplicity, efficiency, and broad application prospects.