Repeated sampling offers insight into glioblastoma therapy response

A multi-institutional study from the Accelerating GBM Therapies Through Serial Biopsies TeamLab, led by investigators from the Mass General Brigham Cancer Institute, found that serially testing tumor samples can help detect when a cancer treatment is activating the immune system in recurrent glioblastoma (GBM), even when traditional imaging measures cannot. Their results are published in Science Translational Medicine.  

GBM is the most aggressive type of brain cancer, known for growing and spreading quickly. It is challenging to treat and almost always comes back. But it can be hard to understand what's happening inside the tumor during treatment.

Getting tissue from GBM patients is difficult because the brain is sensitive, the procedures are risky, and the tumors themselves are complex and change over time. But studying the cancerous tissue itself is also the best way to understand how the tumor reacts to treatment."

E. Antonio Chiocca, MD, PhD, executive director of the Center for Tumors of the Nervous System at the Mass General Brigham Cancer Institute

The study involved over 100 brain tumor scientists and clinicians from multiple hospitals, cancer institutes and universities across the United States and was conducted through a multi-institutional collaboration funded by Break Through Cancer. In the study, researchers collected 96 samples over four months from two patients with recurrent GBM in a clinical trial of a new immunotherapy, CAN-3110, which is a form of immunotherapy known as an oncolytic virus, specially engineered to selectively infect and kill tumor cells.

"Standard practice is to not serially sample a patient's brain tumor as they undergo treatment but instead to take a sample only once before a treatment and then follow a patient's response using MRI," said Chiocca. "But this study's findings suggest that this thinking and practice may need to change to revolutionize how patients can monitor their disease."

The researchers used multi-omic analysis with data integration supported by Break Through Cancer's Data Science Hub (DASH). They integrated data from many sources - genetic material, peptides in and around the tumor, metabolites, immune changes and protein signaling factors, and AI-enabled digital pathology, among others.

The serial samples showed that, over time, the drug changed the environment inside and around the tumor, even though the tumor appeared to be progressing on magnetic resonance imaging (MRI) scans. This may happen because the immune system causes swelling and inflammation, which can look like new or enlarging areas of contrast on a scan even when a tumor hasn't grown, a phenomenon known as pseudoprogression.

If CAN-3110 therapy can reshape the microenvironment and activate the immune system, it may improve patient outcomes, according to the researchers. Of the two patients treated during the study, one showed evidence that the tumor was responding to therapy, while the other's disease remained stable.

"The breadth and depth of data we generated from repeated tumor biopsies really underscore the value of this approach for studying how therapies work," said Chiocca, who emphasized the collaborative work of the entire team of brain specialists from across the country. "These results give us strong reason to adopt a new paradigm in GBM drug development - one that builds longitudinal sampling into clinical trials to capture real-time snapshots of how tumors respond to treatment over time. Although we reported results on just the first two patients from this trial, we are accruing 12 patients to further solidify our findings. We plan to adopt this clinical trial platform now for two additional and distinct vaccine immunotherapies.