Real-world analysis of treatment patterns and survival outcome of glioblastoma patients in a German single-center study: Can survival rates of randomized controlled trials be achieved?

Background: The reported survival data for glioblastoma patients vary strongly between different studies. In our study, we therefore examined which data are applicable in a real-world population in a German center and how these real-world data perform in comparison to survival data presented in randomized controlled trials (RCTs).

Methods: Data of all patients treated with newly diagnosed glioblastoma in a single German center between 2010 and 2019 were analyzed and treatment patterns plus survival rates were matched to existing real-world data and results of RCTs.

Results: Two hundred thirty-three patients were analyzed. Median age was 63 years, f/m ratio was 1:162, and 73% of patients underwent surgery, while 27% had biopsy only. The extent of resection had a significant impact on overall survival (OS; P <.001), as well as age (P <.001), methylguanine methyltransferase methylation status (P <.001), and eastern cooperative oncology group performance status (P <.001). The median OS of our whole study population was 10.55 months. While a fictitious Stupp study cohort (built by using eligibility criteria of the EORTC-22981-26981 trial) with an OS of 14.3 months nearly achieved the survival results of the presented data from the EORTC-22981-26981 trial, the OS of the patients who did not fulfill the eligibility criteria was only 6.9 months.

Conclusion: Survival of patients with unfavorable prognostic factors is still poor and these patients are not represented in recent RCTs. Outcome data of RCTs can be transferred to real world cohorts, if in- and exclusion criteria are fulfilled, while outcome is significantly inferior in cohorts that do not fulfill these criteria.