Construction Of Survival Prediction Model For Metastatic Non-small Cell Lung Cancer Based On SEER Database

Research purpose:Lung cancer is the malignant tumor with the highest morbidity and mortality worldwide.Although there have been major breakthroughs in new treatment technologies such as immunotherapy or molecular targeted therapy in recent years,the 5-year survival rate of patients with metastatic NSCLC is only about 5%-7%.Therefore,this study intends to construct a survival probability prediction model for patients with metastatic NSCLC to predict the survival rate of patients.Research methods:1.Extract patient information from SEER database,retrieve NSCLC patients diagnosed between 2015 and 2018,determine the number of patients after screening according to the inclusion criteria,and randomly divide them into modeling queue and validation queue in proportion.Compare the variable differences between the two groups through chi-square test.2.SPSS 26.0 was used for statistical analysis of the data,single factor and multi-factor COX regression analysis was used to screen the possible prognostic factors,the factors with statistical significance were included in the model construction,and the forest map was drawn by Graphpad software,and the Kaplan-Meier survival curve analysis was performed by logarithmic rank sum test.3.The R 4.1.2 software was used to construct the nomogram,and the Receiver Operating Characteristic Curve(ROC),calibration curve and clinical decision analysis curve were used to evaluate the model.The internal validation cohort was used to internally validate the prediction model.Results of the study:Basic clinical featuresA total of 2514 patients were included in the study and randomly assigned at 7:3.There were 1759 patients in the modeling queue and 755 patients in the validation queue,respectively.In all analysis variables,there was no statistical difference between the modeling queue and the internal validation queue(P>0.05).Risk factors in overall survival of patients with metastatic NSCLCUnivariate analysis showed that age,gender,marital status,histological grade,T stage,lymph node metastasis,organ metastasis(bone,liver,brain),and treatment(chemotherapy,radiotherapy,surgery)were potential prognostic factors for overall survival(OS)in patients with metastatic NSCLC(P < 0.05).Multivariate analysis showed that age,gender,histological grade,T stage,lymph node metastasis,bone metastasis,liver metastasis,brain metastasis,chemotherapy,radiotherapy,and surgery were independent prognostic factors for OS(P < 0.05).The hazard ratio(HR)showed that advanced age,male,poor differentiation,T stage,lymph node metastasis,organ metastasis(bone,liver,brain)were independent risk factors for OS in patients with metastatic NSCLC(HR > 1),among which lymph node metastasis and brain metastasis had the greatest impact on OS.Chemotherapy,radiotherapy and surgery were independent protective factors for OS in patients with metastatic NSCLC(HR <1).Prediction model buildingBased on 11 independent prognostic factors affecting OS of patients with metastatic NSCLC,a prediction model of survival probability of patients with metastatic NSCLC was constructed.Evaluation of predictive modelsThe area under the curve(AUC)of nomograms predicting the survival rates of patients with metastatic NSCLC at 12,24,and 36 months was 0.768,0.760,and 0.770,respectively.The AUC scores of internal validation cohorts predicting the survival rates of patients with metastatic NSCLC at 12,24,and 36 months were 0.796,0.803,and 0.807,all greater than 0.7,with good predictive accuracy.The predicted 12 month,24 month,and 36 month OS calibration curves in the modeling and internal validation queues fluctuate both above and below the standard line,with good accuracy.DCA displays a nomogram that predicts survival probabilities at 12,24,and 36 months in patients with metastatic NSCLC,with a wide range of thresholds.Performance of predictive models in patient risk stratificationIn the modeling cohort and the internal validation cohort,there was a significant difference between the Kaplan-Meier survival curve of the high-risk group and the low-risk group(P<0.01).Conclusion:1.Elderly,male,poorly differentiated,T3-T4 stage,lymph node metastasis,bone metastasis,liver metastasis and brain metastasis are independent risk factors for OS in patients with metastatic NSCLC.Chemotherapy,radiotherapy and operation are independent protective factors for OS in patients with metastatic NSCLC.2.The prediction model based on SEER database has good accuracy,accuracy and clinical applicability,and can well predict the survival probability of metastatic NSCLC at 12 months,24 months and 36 months.