Research Spotlight: External Controls Show Promise for Use in Glioblastoma Clinical Trials

Q: What challenges or unmet needs make this study important?

Glioblastoma (GBM) is a highly aggressive primary brain tumor that has a poor prognosis despite standard of care surgery, radiation therapy and chemotherapy. Many ongoing research efforts aim to develop more effective therapies for GBM, but there have been very few advances over recent decades. Because inefficient clinical trial designs have been identified as one of the challenges for the field, the idea of “external control data” could offer a solution.

External control data means comparing outcomes for people enrolled in a clinical trial to people external to the study. Patients have expressed support for these approaches, as many patients enroll in GBM clinical trials in hopes of receiving a promising experimental therapy and do not want to receive standard of care therapy on the control arm of a trial.

Q: What methods or approach did you use?

We conducted a validation study that re-analyzed a completed randomized controlled trial with external control data. We did this to see if external control data could work for future clinical trials and early-phase testing of promising therapeutics for patients newly diagnosed with GBM.

To serve as possible external control arms for GBM trials, we derived patient-level data from multiple real-world and clinical trial datasets, including data from our institution and others. We did this over several years, as it’s very challenging to assemble this type of data.

We then re-analyzed the experimental arms from the INSIGhT randomized trial (NCT02977780), swapping out the internal control arm from the trial with our external control dataset. We applied statistical techniques to adjust for differences between the external control patients and INSIGhT trial participants. This included estimating treatment effects with the external controls.

As part of our study, we evaluated several novel clinical trial designs—specifically, a single arm trial with external controls, and hybrid trials with randomization and external controls. We also simulated scenarios to evaluate trial designs that integrate external controls.

Q: What did you find?

When we re-analyzed the data from the INSIGhT trial, we obtained the same result when using our external control datasets as the original randomized trial. This supports the use of external control datasets for future clinical trials.

Q: What are the real-world implications, particularly for patients?

For GBM, where the current standard of care control therapies lead to poor outcomes, patients often have a strong preference to receive promising experimental therapies. Our study provides a pathway to novel clinical trials where more patients receive the experimental therapy and fewer patients are assigned to the control arm. Not only would these trials be more patient-friendly, but they would also help accelerate the discovery of new promising therapies for GBM.