Analysis Of Prognostic Factors And Establishment Of Prognostic Model For Small Cell Lung Cancer

Background and aim: Small cell lung cancer(SCLC)is a pathological subtype of lung cancer,and has the characteristics of short doubling time and fast growth rate.Only a few patients can receive surgical treatment,and the vast majority of SCLC patients need to be treated by non-surgical treatment.The purpose of this study was to analyze the prognostic factors of SCLC patients,establish a predictive model based on COX regression coefficient,and visually display the results with nomogram,in order to predict the survival of SCLC patients individually.Methods: A total of 30799 patients diagnosed with SCLC as training cohort were selected using the Surveillance,Epidemiology,and End Results(SEER)database between January 2010 and December 2016.The following information was collected from the database: age,gender,race,degree of tumor differentiation,T stage,N stage,radiotherapy,chemotherapy,site of tumor metastasis,overall survival(OS).Kaplan-Meier method was used to conduct single factor survival analysis and determine the factors influencing the prognosis.Multivariate analysis was performed by COX regression,and independent prognostic factors affecting OS were determined.R3.6.0 software was used to establish the predictive nomogram model for the selected variables,and the prediction accuracy and discriminant ability of nomogram were evaluated by concordance index and calibration curve.We also compared the accuracy of the nomogram model with the 7th edition American Joint Committe on cancer(AJCC)TNM staging system.In addition,the verification cohort of SCLC patients from the first hospital of Jilin university and China-japan union hospital was used to further verify the nomogram model.Results: 1The results of single-factor Kaplan-Meier analysis showed that gender,age,race,degree of tumor differentiation,T stage,N stage,metastatic sites,whether to receive surgical treatment,whether to undergo chemotherapy or radiotherapywere associated with OS of SCLC patients.2 Multivariate COX regression analysis showed that male patients,elderly patients,the white race,late T and N stages,distant metastasis(especially liver metastases or multiple organic metastasis),without surgical treatment,without radiotherapy,without chemotherapy were independent prognostic factors for death risk in SCLC patients.3 The nomogram model established in this study was superior to the 7th edition AJCC TNM staging system(the concordance index of training cohort,0.687 for nomogram versus 0.650 for TNM staging system,P<0.01;the concordance index of validation cohort,0.726 for nomogram versus 0.693 for TNM staging system,P<0.01).The calibration plots presented a good agreement in both cohorts between the nomogram prediction and actual observation for 1-,3-,and 5-year OS.Conclusions: 1.Multivariate analysis can predict prognosis more accurately after removing confounding factors.Male patients,elderly patients,the white race,late T and N stages,distant metastasis(especially liver metastases or multiple organic metastasis),without surgical treatment,without radiotherapy,without chemotherapy were independent prognostic factors for death risk in SCLC patients.2 The prognostic model of SCLC established in this study is superior to the traditional TNM staging model,which can provide more accurate individual prediction for survival rate of SCLC.This practical prognostic model may help clinicians in decision making and design of clinical studies.