No OS or PFS Advantage With Targeted Agents or Immunotherapy in Glioblastoma

In an effort to quantify the comparative efficacy of immunotherapeutic and targeted agents against standard of care in glioblastoma, investigators conducted a systematic review and Bayesian random-effects network meta-analysis (NMA) adhering to PRISMA-NMA guidelines. Randomized controlled trials were identified through searches of PubMed and ClinicalTrials.gov, and analyses were conducted using gemtc in R interfaced with JAGS. Between-study heterogeneity was summarized using tau-squared, I-squared, and Cochran's Q. Node-splitting was planned but was not estimable because the network geometry lacked comparisons informed by both direct and indirect evidence.

The meta-analysis incorporated 24 unique randomized controlled trials. Across these, 26 randomized contrasts from 23 trials contributed to the overall survival (OS) network, while 28 randomized contrasts from 24 trials contributed to the progression-free survival (PFS) network; because trials could contribute to more than one outcome network, these overlapping counts were not additive. No investigational regimen showed a pooled OS or PFS advantage over harmonized control regimens, with all pooled relative-effect 95% credible intervals including 1.0, and treatment rankings were sensitive to sparse network structure. Assessment of objective response rate (ORR) was too sparse for treatment ranking and was reported descriptively across two trials contributing five arms and three treatment–control contrasts.