Refined efficacy and outcome estimates of surgical treatment in oldest-old patients with glioblastomas based on competing risk model and conditional survival analysis: A surveillance, epidemiology, and end results population-based study

Abstract

Background and purposeTo determine the refined estimates of the surgical effects on the short- and long-term prognoses of oldest-old patients (aged ≥80 years) with glioblastomas. Materials and methodsUsing the Surveillance, Epidemiology, and End Results registry, we identified the oldest-old patients with glioblastomas between 2005 and 2016. Propensity score matching, Kaplan–Meier analysis, Cox regression analysis, and competing risk model were used to assess the curative efficacy of the surgical treatments. Stratification and interaction analysis were performed to explore the potential interaction effects. The conditional survival rates were calculated to explore the longitudinal change in the survival probability over time. ResultsThis study enrolled 3309 patients with a median overall survival of 3 months. The overall survival differed significantly among the different surgical groups. Considering the gross total resection as reference, subtotal resection presented adjusted subdistribution hazard ratio (95% confidence interval) of 1.197 (1.052–1.362; p < 0.001); biopsy/partial resection, 1.242 (1.083–1.424; p = 0.002); and no surgery, 1.309 (1.145–1.497; p < 0.006). Age ≥ 83 years, widowed/other status, tumor size 4–5 cm, and temporal tumors showed significant interaction effects. The adjusted subdistribution hazard ratio of radical resection was 0.729 (0.645–0.825; p < 0.001). Based on disease-specific survival, the 1-year survival rate was 18% and 26% for the non-radical and radical surgery groups, respectively. The 1-year conditional survival rates in the second year were 54% and 39% in the non-radical and radical surgery groups, respectively. The 3-year survival rates were 10% in both the groups. ConclusionsRadical surgery may have short-term benefits in the oldest-old patients with glioblastoma, with a significant increase in the 1-year survival rate. However, its contribution in the long-term outcome is limited due to decreased conditional survival rates from the second year after surgery. Prudent patient selection and improved postoperative management may be needed to promote the therapeutic efficacy of tumor resection synergistically.