Reanalysis of Glioblastoma Vaccine Trial Suggests Larger Survival Benefit

Individual patient-level analyses of a phase 3 trial testing an autologous tumor lysate-loaded dendritic cell vaccine (DCVax-L) in glioblastoma suggest that the survival benefit associated with the therapy may be larger than originally reported, according to data presented July 8 at the 2026 Annual Meeting of the British Neuro-Oncology Society.¹

The new analyses, conducted by independent statisticians on behalf of Northwest Biotherapeutics, the vaccine's manufacturer, applied propensity score matching (PSM) and inverse probability weighting (IPW) to compare outcomes in trial participants against individual patient data (IPD) obtained from several randomized controlled trials in newly diagnosed glioblastoma. The company had previously relied on cohort-level data for external comparators, as IPD access was not yet available when the trial's primary results were published.

Across 2 large comparator trials, PSM analyses estimated a median overall survival advantage of 4.9 to 6.3 months with the vaccine; a third comparator trial yielded an estimated advantage of 3.4 to 3.7 months. Hazard ratios across these analyses ranged from 0.69 to 0.77, with P values from .004 to .027. IPW analyses, using a different statistical approach, produced comparable estimates of 3.4 to 4.3 months. Marnix Bosch, PhD, chief technical officer of Northwest Biotherapeutics, presented the findings. These figures exceed the 2.8-month median survival advantage reported in the trial's original cohort-level analysis.

To assess whether the observed associations could be explained by unmeasured confounding rather than treatment effect, the statisticians additionally conducted E-value and Rosenbaum's gamma sensitivity analyses.

"These new PSM and IPW analyses, together with the E-value and Rosenbaum's gamma sensitivity analyses, provide strong evidence and reinforcement of the survival benefit with DCVax-L treatment, based on my review of the statisticians' analyses," said Bosch in a news release. "The key is that all of these analyses…have produced results that are all directionally consistent and all in the same range of magnitude. The statisticians' findings indicate that the median survival benefit reported from the original analysis of the DCVax-L phase 3 trial, based on cohort-level data, was likely substantially underestimated."

Trial Background

The phase 3 trial (NCT00045968) enrolled 331 patients with glioblastoma and evaluated DCVax-L, an autologous vaccine prepared by pulsing patients' own dendritic cells with lysate derived from their resected tumor tissue, in combination with standard therapy. Primary results were published in JAMA Oncology in 2023 and reported a median overall survival advantage in the range of several months for vaccine-treated patients compared with external control populations;² a marketing authorization application based on those results is currently under review by UK regulators.¹

Because the trial used external controls rather than a concurrent randomized comparator arm, external validity depends heavily on how closely those controls resemble the treated population. The original analysis matched trial participants to controls using cohort-level summary statistics. The newly presented analyses instead used IPD, which allows matching on individual patient characteristics such as age, performance status, and extent of resection, a method the trial's statistical analysis plan had anticipated using once such data became accessible.

Clinical Context and Caveats

PSM and IPW are established methods for estimating treatment effects in nonrandomized comparisons, and sensitivity analyses such as E-values and Rosenbaum's gamma are standard tools for gauging susceptibility to unmeasured confounding. However, none of these techniques can fully substitute for a concurrently randomized control group, and residual confounding from factors not captured in either dataset cannot be entirely excluded. The updated findings have been presented at a scientific meeting and have not yet undergone independent peer review or publication in a journal.

Clinicians evaluating these data should consider them alongside the original peer-reviewed trial results and await further validation before drawing conclusions about the magnitude of the treatment effect.

REFERENCES

1. Northwest Biotherapeutics Presents Updated Survival Data from Phase III Trial of DCVax®-L For Glioblastoma Using Individual Patient Level Data in Multiple Independent Analyses. News release. Northwest Biotherapeutics. July 8, 2026. Accessed July 10, 2026. https://tinyurl.com/m7txxpej

2. Liau LM, Ashkan K, Brem S, et al. Association of autologous tumor lysate-loaded dendritic cell vaccination with extension of survival among patients with newly diagnosed and recurrent glioblastoma: a phase 3 prospective externally controlled cohort trial. JAMA Oncol. 2023;9(1):112-121. doi:10.1001/jamaoncol.2022.5370