A Propensity Score Matching Study Based On A High-risk Prostate Cancer Population

Objective:To compare survival outcomes between radical prostatectomy and external beam radiation therapy for high-risk prostate cancer in order to provide guidance for treatment decisions in this population.Methods:We identified a cohort of men diagnosed with high-risk prostate cancer(HRPCa)between 2007 and 2015 in the Surveillance,Epidemiology,and End Results(SEER)database,compared overall survival(OS)and cancer-specific survival(CSS)rates using Kaplan-Meier survival curves and log-rank tests,calculated relative risks(HRs)in univariate and multivariate Cox regression models,and identified factors affecting the risk of all-cause and cancer-specific death.A new matched cohort was generated after balancing covariates using propensity score matching(PSM)for the corps column,and survival curves and log-rank tests were refitted to evaluate OS and CSS more rigorously.Results:A total of 31,755 patients were identified in the corps,of whom 17,835 received RP only and 13,920 received EBRT only.The OS curve showed that RP was associated with a better OS outcome,with an HR of 0.243(95% CI: 0.229-0.257)in the RP group compared with the EBRT group.The CSS curve showed that the EBRT group was associated with worse CSM outcomes,with an HR of 0.200(95% CI: 0.173 – 0.232)in the RP group using the EBRT group as the control.Univariate and multivariate Cox regression analysis identified treatment modality,age,ethnicity,marital status,T stage,GG,and PSA level as factors affecting the risk of death.A new matched cohort was generated after 1:1 PSM of the coronavirus column,in which 3596 patients were treated with RP only and 3596 patients were treated with EBRT only.The OS curve for the matched cohort showed that the RP group was associated with better OS,with an HR of 0.480(95% CI: 0.441 – 0.523)for the RP group compared with the EBRT group.The CSS curve showed that the EBRT group was associated with worse CSM outcomes,with an HR of 0.407(95% CI: 0.339 – 0.494)in the RP group using the EBRT group as the control.Conclusion:Three main conclusions emerged from this study.First,there is a clear bias in the choice of treatment modality for HRPCa: relatively younger patients,those with higher T stage,those with lower GG,and those with lower PSA levels are more biased towards choosing RP as the initial treatment.Second,age,race,marital status,T stage,GG,and PSA level were strongly associated with the risk of death: advanced age,black race,divorce or widowhood,high T stage,high GG,and high PSA level were contributing factors to the death event,while relatively young age,Asian race,married,low T stage,low GG,and low PSA level were protective factors.Third,after balancing the above factors associated with the risk of death,there was still a significant survival benefit in the RP group compared with the EBRT group,with an approximately 62% reduction in the risk of all-cause mortality and an approximately 70% reduction in the risk of cancer-specific mortality.