# Conditional survival in glioblastoma: The evolution of prognostic factors over time

## TL;DR

This study examines how survival rates and prognostic factors for glioblastoma change over time, showing that different factors influence outcomes at different stages after diagnosis.

## Contribution

The study introduces conditional survival analysis to show how prognostic factors for glioblastoma evolve over time, offering new insights for risk stratification in clinical trials.

## Key findings

- Conditional survival estimates decrease initially but show improvement at 24 months post-diagnosis.
- Residual tumor volume is a key early predictor, while MGMT methylation and age become more significant later.
- Age, MGMT status, and postoperative tumor volume are baseline prognostic markers for glioblastoma.

## Abstract

Conditional survival provides insights into the evolution of prognosis over time and reveals changing associations of prognostic factors during disease progression. Data on the temporal evolution of prognostic factors in glioblastoma remain scarce. We analyzed 315 patients with IDH‐wildtype glioblastoma from a prospectively collected registry (01/2008–06/2017). Our primary outcome was 12‐month conditional survival (CS), defined as the probability of surviving the next 12 months given survival for “s” months. This analysis was conducted at five landmarks (s = 0, 6, 12, 18, 24) for baseline prognostic factors, including tumor volume compartments. 12‐month conditional survival estimates at s = 0, 6, 12, 18, and 24 months from diagnosis were 0.51 (95% CI 0.45–0.56), 0.46 (95% CI 0.39–0.52), 0.41 (95% CI 0.33–0.49), 0.43 (95% CI 0.33–0.52), and 0.56 (95% CI 0.42–0.67), respectively. At diagnosis (s = 0), 12‐month survival estimates varied significantly with age >60 at diagnosis, preoperative tumor rim volume >20 cm3, absence of O6‐methylguanine‐DNA methyltransferase (MGMT) promoter methylation, postoperative KPS ≥70, residual postoperative tumor >1 cm3, or biopsy only. Residual tumor volume mainly influences survival in the initial months following surgery, while MGMT promoter methylation and age remain significant predictors beyond this period. These findings may refine stratification strategies in recurrent glioblastoma trials.

Neurooncologists rely on glioblastoma risk stratification systems for incorporating known disease‐ and patient‐specific factors at diagnosis. Although useful for general comparison and disease monitoring, these estimates are less informative for patients who survive beyond various time intervals post‐diagnosis. Here, the authors analyzed data on isocitrate dehydrogenase (IDH)‐wildtype glioblastoma patients to estimate conditional survival from 6 to 24 months following diagnosis. Age, MGMT methylation status, postoperative Karnofsky performance status, and postoperative enhancing tumor volume were identified as baseline prognostic markers. In the months after surgery, residual tumor volume dominated prognosis, whereas age at diagnosis and favorable MGMT status determined prognosis thereafter.