Objectives:
Competing risk models were used in this study. The purpose of this study was to assess the predictive usefulness of lymph node characteristics in elderly patients with stage III serous ovarian cancer.
Methods:
We conducted a retrospective analysis on 148,598 patients from 2010 to 2016 using the surveillance, epidemiology, and end results database. Lymph node characteristics were collected and examined, including the number of lymph nodes retrieved the number of lymph nodes examined (ELN) and the number of positive lymph nodes (PN). Using competing risk models, we evaluated the connection between these variables and overall survival (OS) and disease-specific survival (DSS).
Results:
This study included a total of 3457 ovarian cancer patients. Multivariate analysis using the COX proportional hazards model found that ELN>22 was an independent predictive factor for both OS (hazard ratio [HR] [95% CI]=0.688 [0.553 to 0.856], P<0.05) and DSS (HR [95% CI]=0.65 [0.512 to 0.826], P8 was identified as a significant risk factor for both OS (HR [95% CI]=0.908 [0.688 to 1.199], P=0.497) and DSS (HR [95% CI]=0.926 [0.684 to 1.254], P=0.62). Subsequently, using the competing risk model, ELN>22 was found to be an independent protective factor for DSS (HR [95% CI]=0.738 [0.574 to 0.949], P=0.018), while PN>8 was identified as a risk factor for DSS (HR [95% CI]=0.999 [0.731 to 1.366], P=1).
Conclusions:
Our findings demonstrate the robustness of the competing risk model to evaluate the results of the COX proportional hazards model analysis.