NEW YORK (GenomeWeb) – A number of research groups and companies have recently begun using cell-free tumor DNA as a biomarker to monitor cancer.
However, studies have thus far been less successful in finding cell-free DNA from brain and spinal cord tumors in patients' blood. Now, researchers from Johns Hopkins University have shown that cell-free DNA can instead be detected in the cerebrospinal fluid of such patients.
Publishing their results today in the Proceedings of the National Academy of Sciences, the researchers analyzed 35 patients with central nervous system cancers, including six with medulloblastomas and 29 with gliomas.
First, the group sought to identify somatic mutations from the primary tumor tissue, which was collected during surgery. They used targeted sequencing initially, which found mutations in 13 tumors, followed by exome sequencing to identify at least one mutation in each of the remaining tumors.
Then, using a next-generation sequencing approach the lab previously developed called SafeSeq-S, which enables the detection of mutations with allele fractions as low as .01 percent, they searched for the mutations identified from the primary tumor tissue in cerebrospinal fluid (CSF) collected from each patient.
Twenty-six patients had detectable levels of CSF tumor DNA at highly variable mutant allele fractions — ranging from .1 percent to 77 percent. The authors theorized that the high variability between samples might be due to an "anatomical or biological factor."
Further analysis found that anatomical location correlated with the ability to detect CSF tumor DNA. Patients with tumors near a cerebrospinal fluid reservoir were much more likely to have detectable tumor DNA than those whose tumors were further away from reservoirs.
None of the five patients whose tumors were entirely enclosed by the brain or spinal cord had detectable CSF tumor DNA. Meantime, 24 out of the 28 cases in which tumors were adjacent to a reservoir had detectable levels of CSF tumor DNA.
Aside from anatomical location, the researchers found that high-grade tumors were more likely than low-grade tumors to have detectable CSF tumor DNA. The four cases with tumors near a cerebrospinal fluid reservoir but that did not have detectable CSF tumor DNA were all low-grade gliomas.
Four patients also had tumors located in places that would make surgery "treacherous." For such cases, it would be ideal to be able to detect CSF tumor DNA without prior knowledge of the tumor's genotype.
To test this, the researchers performed whole-exome sequencing on the CSF in those four cases to see whether tumor DNA could be identified. In two of the four cases, exome sequencing identified comparable mutations to those identified from the SafeSeq-S approach. The two cases where exome sequencing did not identify CSF tumor DNA had mutant allele fractions less than 1 percent.
There is a great need for more sensitive noninvasive techniques to assess brain and spinal cord tumors, the authors wrote. For instance, in nearly 30 percent of glioblastoma patients that have surgery because disease recurrence is suspected, the surgery turns out to be unnecessary, and rather than a tumor, imaging had picked up tissue scarring or another treatment-related effect.
"Although the rate of detection observed in this study was not 100 percent, its sensitivity was comparable with or superior to other noninvasive tests for malignanices in general," the authors wrote, and was "particularly sensitive for tumors that abutted a CSF reservoir or cortical surface."
The researchers cautioned that although the study was "exploratory," it sets the stage for a "longitudinal study of the clinical use" of CSF tumor DNA as a biomarker.