Measuring epigenetic signatures found in blood plasma could help classify brain tumors, according to a Northwestern Medicine study published in Nature Medicine.
Diagnosing brain tumors via imaging can be difficult, and clinicians usually must resort to obtaining tissue specimens via invasive surgery. Using epigenetic markers — in this case, DNA methylation — to identify the tumors could improve diagnosis, classification and treatment of the tumors, according to Craig Horbinski, MD, PhD, director of Neuropathology in the Department of Pathology and a co-author of the study.
“Upfront screening of every patient with a brain lesion could be transformative,” said Horbinski, who is also a professor of Neurological Surgery and director of the Nervous System Tumor Bank and the Mouse Histology and Phenotyping Laboratory at the Robert H. Lurie Comprehensive Cancer Center of Northwestern University. “It could help guide the neurosurgeon’s approach to the patient.”
One major challenge to diagnosing and treating brain tumors is ensuring accurate diagnosis of lesions found on imaging. The lesions can be tremendously varied, ranging from low-grade tumors to aggressive cancers, but distinguishing between them can sometimes be impossible via imaging alone.
Subsequently, many patients must endure invasive neurosurgery to obtain tissue for diagnosis and genetic analysis. This informs treatment and patient outlook, but can be complicated for clinicians and taxing on the patient, according to Horbinski.
“The brain is more delicate than other organs, and is completely encased in bone,” Horbinski said. “Even a relatively straightforward neurosurgical biopsy or resection takes hours.”
Previous studies have shown that cancer cells have distinct alterations in DNA methylation from cancer to cancer, giving clinicians a broad idea of the characteristics of that particular tumor or lesion. These alterations can be measured in free-floating DNA fragments found in blood plasma, but previously, it was unknown if these two concepts could be combined to classify brain tumors.
To test if measuring these fragments in blood plasma could identify tumors, Horbinski and his collaborators employed this diagnostic tool in nearly 450 patient blood plasma samples. Applying a machine-learning program, the investigators found that the tool could identify different kinds of brain tumors, and distinguish them from metastatic cancer or healthy controls.
The tool could also distinguish between intracranial glioblastoma with different mutations, a remarkable level of granularity that can help clinicians plan their diagnostic and treatment strategy.
According to Horbinski, this could help pathologists prepare to analyze tissues that are eventually resected, or help the treating oncologist determine how well a tumor is responding to therapy.
“It might even someday prompt earlier treatment of certain tumors, before they are ever operated on,” Horbinski said.
Further, this method wouldn’t require hospitals to have expert epigenetic testing and analysis on-site, Horbinski said, broadening the potential impact of this diagnostic tool.
“All a hospital or clinic would have to do is obtain the patient’s blood, freeze it, and ship it to a central lab that has the requisite testing equipment, capacity and expertise,” Horbinski said.
This study was supported in part by the Northwestern Nervous System Tumor Bank, funded by a P50CA221747 SPORE for Translational Approaches to Brain Cancer, and National Institutes of Health grant R01NS102669.
