NEW YORK – Genomic sequencing of low-grade gliomas uncovered hidden, more aggressive glioblastomas in patients treated at the University of California, San Francisco Brain Tumor Center, allowing them to receive appropriate, more aggressive treatment that led to better outcomes compared to historical controls.
In a study published in Neuro-Oncology, researchers led by David Solomon, assistant professor at the UCSF department of pathology, described the experience of 38 patients who were seen at the Brain Tumor Center between 2015 and 2021. Their pathology results were consistent with IDH-wild-type diffuse astrocytic glioma, but with results from the UCSF500 Cancer Gene Panel they were diagnosed with grade 4 glioblastoma.
This type of molecular testing is routine at UCSF for patients who present with brain tumors. The molecularly integrated diagnosis was based on identification of DNA alterations known to drive tumor growth. The patients were then offered a standard course of therapy for glioblastoma, and a chance to join a clinical trial.
"Because we diagnosed these patients as having glioblastoma, they were able to receive more aggressive treatment with radiation and chemotherapy and enroll in clinical trials for glioblastoma that they wouldn't have been eligible for based on microscopic diagnosis," said Solomon.
Patients with low-grade glioma have a more favorable outlook than those with high-grade tumors. Patients with low-grade tumors, who are typically younger, can be followed without immediate radiation and chemotherapy until progression of the cancer. These patients live for around seven years after diagnosis.
The 38 patients who were diagnosed as having glioblastoma based on the molecular features of their tumors and received appropriate treatment had median overall survival of 24 months. In comparison, patients in a biologically matched retrospective cohort who were diagnosed only with conventional microscopic grading and treated less aggressively, had median overall survival of 16 months.
The study authors said that the reason for the significant difference in survival is that some patients diagnosed as having low-grade tumors by microscopic diagnosis actually have more aggressive disease that's not being adequately treated. In the UCSF study, the researchers hoped to demonstrate the usefulness of next-generation sequencing toward this end.
The results support routine use of genomic profiling to aid in accurate classification and treatment of brain cancers, Solomon and colleagues said.
UCSF's analysis was inspired by a 2015 study from The Cancer Genome Atlas Research Network that identified a subset of patients with grade 2 or grade 3 gliomas who lacked a mutation in IDH1, which is associated with better outcomes. These patients were five to 10 years older than patients with IDH-mutant gliomas, or about 50 years old, and followed a clinical course more similar to grade 4 glioblastoma than to a lower-grade glioma. The UCSF cohort identified by molecular profiling resembled these patients.
In 2018, cIMPACT-NOW, a consortium that helps oncologists stay up to date on the treatment of central nervous system tumors, proposed three molecular features for classifying an IDH-wild-type diffuse astrocytoma as a "diffuse astrocytic glioma, IDH-wild type, with molecular features of glioblastoma WHO grade IV." The features were combined whole chromosome gain/trisomy of chromosome 7 and loss/monosomy of chromosome 10, focal EGFR gene amplification, and a TERT promoter hotspot mutation. These criteria were included in the fifth edition of the WHO Classification of Tumors of the Central Nervous System in 2021.
Even though the patients in the UCSF study received NGS profiling on a panel that tests for alterations in many cancer-linked genes, Solomon noted that matching actionable biomarkers with targeted therapies is a bit more challenging in brain cancer because many approved precision oncology products do not cross the blood-brain barrier. Even so, Solomon said, "brain tumors are one of the spaces in the oncology field where molecular medicine has really been leading the way."
Luis Parada, director of the Brain Tumor Center at Memorial Sloan Kettering is doing work that dovetails with the research conducted by Solomon's group. Parada studies the origins of brain tumors in engineered mouse models.
"We've identified multiple cells of origin for these tumors," said Parada. "[The tumors] don't arise in any old cell in the brain. They arise in very rare cells in the brain, distinct ones, and the genomic information held by the cell relates to the cell that they came from."
That finding is consistent with the published results from Solomon's group, showing that molecular features of the reclassified low-grade gliomas resembled those of aggressive tumors much more than nonaggressive tumors. "It turns out that the molecular features are a better and more accurate way of evaluating the tumor, and predicting its prognosis and its course" than simply how the tumors look, said Parada.
According to Parada, this type of molecular reclassification of brain tumors is becoming very common. Instead of pathologists looking only at tumor morphology, sequencing is now routinely incorporated into the diagnostic workup for patients after imaging and surgery. "That will become standard of care," he said.
UCSF launched its UCSF500 Cancer Gene Panel for clinical use in 2015. Solomon said that currently all UCSF patients who undergo brain tumor resection at UCSF also undergo molecular profiling of their tumor samples on the gene panel. Patients receive a final diagnosis incorporating their molecular findings about three weeks after initial testing.
Around 5 to 10 percent of patients have had their brain cancer diagnosis amended based on the results of the UCSF500 Cancer Gene Panel since its implementation, including six out of the first 31 pediatric cancers treated there.
Although the UCSF study focused on patients whose brain tumors were reclassified to a more aggressive form of cancer, there are also cases where the tumors are downgraded. For example, a child's tumor, which was classified as grade 4 based on pathology, was reclassified to grade 1 after genomic testing. Although that can happen, it's "rare," Parada said, because most high-grade brain tumors are correctly classified by mitotic index, a measure of active cell division.
"Most tumors that look high grade are high grade," said Parada.
In Solomon's view, genomic testing is instrumental to ensure the most accurate diagnosis, grade, and prognosis for brain cancer patients, to rule out an inherited tumor predisposition syndrome, and to identify any actionable genetic vulnerabilities that can be targeted with off-label therapies.
Even though all brain cancer patients have their tumors sequenced at the time of diagnosis, reimbursement for the test can be tricky, Solomon acknowledged. He is hopeful insurers are starting to recognize genomic profiling as a standard and routine part of managing and treating brain cancer patients, and "that [molecular analysis] will continue to be reimbursed by insurance payors more and more moving forward."