Cumulative Survival And Conditional Survival Estimates Of Five Treatment Strategies For StageⅠNon-Small Cell Lung Cancer Older Than 70-years

Purpose: With the longevity prolonging,the risk of death changes over time.Comparing with traditional cumulative survival rate,CS(conditional survival),namely the possibility of patients who had already survived for x years to survive extra y years,considering this factor,is used to describe the dynamic possibility of survival.This study aimed to estimate CS of the different strategies for stage I non-small cell lung cancer(NSCLC)older than 70-years old based on estimating cumulative survival rate of the different strategies for those patients.Methods: This study involved 16,322 patients after 70-years old with stage I NSCLC(8th edition)from the Surveillance,Epidemiology and End Results database from 2004 to 2015.Student’s t-test was used to compare the difference of two continuous variables.One-way analysis of variance was used to compare the difference of three continuous variables,and if varance is unequal,Kruskal-Wallis test was applied.Pearson’s χ2 test was used to compare the difference of categorical variables.Survival time was culculated from the month of diagnosis to the month of death or the last time to follow up.Kaplan-Meier method and Log-rank test were used to compare the overall survival(OS)and disease-specific survival(DSS)of different strategies.The hazard function based on OS or DSS was estimated and drawed by kernel method.CS estimates was calculated as CS=S(x + y)/S(x),where S is the OS or DSS at a particular point in time.The Cox regression model was used to adjust the influence of clinical baseline characteristics to prognosis,to identify whether treatment strategies is the independent prognostic factor.Propensity-score matching(PSM)was used to match clinical baseline characteristics among surgical groups and among nonsurgical groups respectively,to decrease the bias to prognostic analysis caused by baseline characteristics.For all analyses,the significance test was based on a two-tailed hypothesis of 0.05 levels.All analyses were performed using IBM SPSS 23.0(SPSS,Inc,Chicago,IL)and R program(version 3.2.2,R Foundation for Statistical Computing,Vienna,Austria;available at: https://www.r-project.org/).Results: A total of 16,322 patients were identified in the study.Among surgical groups,the amount of patients who received lobectomy,segmentectomy,wedge resection were 9574(75.10%),688(5.40%),2492(19.50%).Among non-surgical groups,the amount of patients who received radiotherapy or supported therapy were 2307(64.70%)and 1261(35.30%).The risk of death based on OS peaked at the 7th,8th and 87-95 th months after diagnosis.the risk of death based on DSS peaked between 7th months and 18 th months after diagnosis and then declined.Treatment strategies,the year of diagnosis,race,gender,age,T classification,pathological grade,and pathological type were significantly associated with OS.Treatment strategies,year of diagnosis,gender,age,T classification,pathological grade,and pathologic type were significantly associated with DSS.Among the three surgical options,the OS and DSS of the lobectomy were superior to the sublobar resection.The OS of the segmentectomy was superior to the wedge resection,but there was no statistically significant difference with wedge resection in DSS.After pair-matching the three surgical treatments through PSM,the OS and DSS of the lobectomy and segmentectomy were similar,and the p values were close to 0.05.The sample size should be expanded or a randomized controlled study should be conducted to further clarify whether lobectomy has a similar effect with segmentectomy.Lobectomy and segmentectomy were superior to wedge resection in OS.In DSS,lobectomy was significantly better than wedge resection.Segmentectomy was superior to wedge resection,but no significant difference was seen.Regardless of matching or not,in both non-surgical treatment strategies,the OS and DSS of the radiotherapy were superior to the alternative therapy.In terms of 3-year conditional survival rate(3y-CS),lobectomy was the highest,followed by segmentectomy,then wedge resection,radiotherapy and alternative therapy.After the three surgical procedures were paired-matched,the OS-3y-CS and DSS-3y-CS of the lobectomy were consistently higher than the sublobar resection.The OS-3y-CS of the segmentectomy was higher than the wedge resection,with the similar DSS-3y-CS.Although the OS-3y-CS of supported therapy is much lower than radiotherapy in the early stage,it gradually converges with radiotherapy after the third year.DSS-3y-CS of both non-surgical strategies gradually increased,however,DSS-3y-CS for radiation therapy was consistently higher than alternative therapy.Conclusion: This study is based on the SEER database,and compared the cumulative survival rate of patients with stage I NSCLC older than 70-years old who have received different treatment strategies since 2004.Furthermore,based on the cumulative survival rate,the conditional survival rate was evaluated.Conditional survival rates provide a more meaningful dynamic prognostic assessment for stage I NSCLC in elderly who have survived for a period of time after diagnosis.Our results suggest that lobectomy is still the best treatment option,and only advanced age is not sufficient as an indication for predisposition to sublobar resection or non-surgical treatment.If the patient’s physical condition does not allow for lobectomy,segmentectomy is preferred over wedge resection.If it is not suitable for surgical treatment,radiotherapy is more beneficial to prognosis than alternative treatment.This study may be helpful in the selection of treatment strategies for elderly patients with stage I NSCLC and provide dynamic information for the prognosis of patients with advanced stage I NSCLC who received different treatment decisions.