By Deborah Borfitz

May 9, 2022 | Evidence is emerging that cerebrospinal fluid (CSF) testing might be useful in clinical settings for monitoring tumor growth and treatment response in pediatric patients with diffuse midline glioma (DMG), a lethal high-grade brain tumor that is inoperable and unlikely to be cured. Only about 500 new cases get diagnosed in the U.S. in any given year, but over two-thirds of them carry the exact same mutation (H3 K27M) allowing development of an assay that analyzes cell-free tumor DNA samples using digital droplet polymerase chain reaction (ddPCR), according to Carl Koschmann, M.D., pediatric neuro-oncologist at University of Michigan Health C.S. Mott Children’s Hospital who specializes in treating DMG.

The assay is currently being used in a phase 2 clinical trial of an experimental therapy called ONC201 (Oncoceutics, recently acquired by Chimerix), in combination with different drugs (panobinostat or paxalisib), which is being conducted by the Pacific Neuro-Oncology Consortium. But the methods used for CSF collection and to generate the test could be applied more broadly for the benefit of other high-grade gliomas if they likewise share a common genetic cause, Koschmann says.

Results of the industry-sponsored phase 1 trial of ONC201 suggest CSF sampling can provide additional information about tumor markers before changes appear on an MRI, the traditional way to assess response to therapy for gliomas. In cases with tumor growth on ONC201, the experimental assay detected a spike in the allele fraction in the H3 K27M tumor DNA an average of three to four months before the tumor grows, as recently reported in Neuro-Oncology (DOI: 10.1093/neuonc/noac030).

This apparent window of opportunity for adjusting treatment is one of the most important findings of the study, says Koschmann, since the expectation was that DNA spilling from the tumor would track closely with change in tumor size. It was only one of the surprising discoveries.

The other, he adds, is that tumor DNA was detectable in most of the plasma samples as well, only the signal was “more subtle” than what was coming from the spinal fluid. “There is a story in the plasma of patients with brain tumors, but it takes using substantial wet lab techniques [serial amplification and concentrations of the sample] to get it to show its head.” 

Brain tumor liquid biopsy researchers who have abandoned their search for genetic clues in plasma may want to circle back using the described methodology, Koschmann says. It’s what helped the Michigan Medicine team increase the sensitivity of their assay to over 90% even in the plasma samples—and without creating false-positives or disrupting the variant allele fraction of the mutated DNA.

In spinal fluid samples from brain tumor patients, but even more so in plasma samples, tumor DNA is otherwise a “needle signal in the haystack,” he continues. The research team worked through multiple ways to amplify that signal until they settled on one where they could trust the percentage they were finding.

Joining the University of Michigan in the exploratory Arm D of the phase 1 clinical trial, where the assay was used on 24 patients, were Seattle Children’s Hospital, the Miami Cancer Institute, and the University of California, San Francisco. With the phase 2 study—involving more than 20 sites across the U.S. and Europe collectively enrolling 200-plus participants up to 39 years of age—they now hope to show the assay yields valuable information from plasma and CSF in a larger patient cohort, says Koschmann.

“We want to see that it’s feasible, it’s safe, and [produces] similar signals as we saw in the pilot study,” he adds. If so, Koschmann believes the liquid biopsy test will be ready for adoption in clinical settings outside of a trial.

As with the initial study, the spinal fluid and plasma tests in the phase 2 trial will not affect patient care in real time, says Koschmann. Patient samples will be run in batches later this year and, if findings look “credible and consistent,” they could be shared with families. “But we’re not yet ready to change treatment based on results.” 

Changes Ahead 

This is believed to be the first clinical trial involving serial CSF collection from glioma patients of any kind, says Koschmann. Serial lumbar puncture is not routinely performed in this patient population, unlike leukemia and other types of pediatric brain cancers that are more likely to spread into the spinal fluid. 

Consequently, children being treated at pediatric cancer centers often get spinal taps but generally don’t experience the procedure. For the DMG studies, CSF collection happens at the same time patients are getting an MRI and are being anesthetized anyway, he says, at baseline and again at two and six months.

The ddPCR machine used in the study is still found mostly in research labs, Koschmann adds. But in clinical labs where ddPCR is available, the new assay is poised to become a reimbursable, CLIA-certified test for real-time patient monitoring.

Theoretically, Koschmann continues, the test could hop over now to any CLIA-certified testing lab interested in offering it based on the published data. The Neuro-Oncology paper includes all the necessary information to recreate the assay and run the test. 

Only a half dozen labs anywhere are working on liquid biopsy tests for pediatric glioma, and perhaps another one to two dozen in the adult space, and they have “all started to collaborate and learn from each other,” he says. “It is really just a question of what is the best way to move this over to the clinic so patients can get the test run if they’re interested… [and] clinicians and families have [that] extra piece of information. 

The harder part, says Koschmann, may be finding the right language to use with families about adding spinal taps to their child’s clinical care regimen. “I think it’s only a matter of time before either our blood tests get so good that we don’t need to do the spinal test or the spinal fluid tests are so useful that it will become standard practice.”

While diffuse midline glioma remains a “tough prognosis” for kids—they typically have a 12-to-18-month survival rate—more scientists are working on treatments than ever before, says Koschmann. Several drug candidates appear to improve overall outcomes and, as the combinatorial phase 2 trial with ONC201 will help determine, perhaps they’ll prove synergistic in a therapeutic cocktail. “I think there is room for some optimism.”