NEW YORK (GenomeWeb) – Researchers have developed a way to look for signals of a tumor in blood-borne DNA fragments by tracking the end-coordinates of molecules, rather than looking for cancer-associated mutations.
In a study, published today in PNAS, a team from the Chinese University of Hong Kong reported on its analysis of plasma samples from cohorts of liver cancer, liver transplant, and hepatitis patients, concluding that tracing the ends of DNA fragments can be used as a marker of cancer in the blood.
Led by CUHK researcher Dennis Lo — who has pioneered a variety of other liquid biopsy methods and is commercializing a test for virus-associated head and neck cancers in collaboration with US firm Grail — the team set out to look at different ways that they could glean a cancer-associated signal from circulating DNA fragments, outside of the obvious use of cancer mutations.
In their study, the Hong Kong investigators first described the approach they have developed to detect the total pool of cancer-associated somatic mutations — both drivers and passengers — in plasma, using a variety of informatics and error-correction tools to be able to distinguish cancer-associated changes from germline mutations or other non-tumor sources of DNA alteration.
They then turned their attention to trying to identify other sources of cancer signal aside from mutations.
According to the authors, as liquid biopsy efforts are turning increasingly toward early detection in asymptomatic people, various approaches are being examined to try to hone in on more powerful and specific signals that can work across a genetically diverse landscape of tumor types.
Apart from looking for tumor mutations, other methods being studied have included DNA methylation, nucleosome positioning, and in certain cancer types, the detection of DNA from tumor-generating viruses.
Cell-free DNA fragmentation has recently been shown to be a nonrandom process, Lo and his coauthors explained, and as such, offers what they saw as a promising source for patterns that are specific to tumor DNA and could be used to identify the presence of cancer.
Using plasma samples obtained from a small group of liver transplant recipients, the team first showed that cell-free DNA molecules from the liver tended to end more frequently at certain genomic coordinates than non-liver-derived molecules.
Investigators then analyzed sequencing data from a cohort of patients with hepatocellular carcinoma and compared them to samples from individuals with chronic hepatitis B, finding what they described as millions of tumor-specific plasma DNA end coordinates.
According to the authors, the abundance of these circulating DNA molecules with cancer-specific end coordinates also correlated closely with patients' tumor DNA fractions, even in those whose samples were only sequenced shallowly. This suggests that the fragment-end approach might offer a cheaper way to look for signs of cancer in the blood than mutation-based methods.
"Many of the strategies formerly developed for the detection of circulating somatic mutations may be adapted for the detection of tumor-associated cell-free DNA preferred ends," Lo and his colleagues wrote.
As a result, the combination of somatic mutation detection and tumor-associated DNA end coordinates could offer the potential to greatly enhance the sensitivity of liquid biopsy and "it would be interesting to assess the value of deploying these end coordinates for the noninvasive detection of early cancers," the group concluded.
Lo and colleagues reported that they have filed patent applications based on the data reported in the study. Multiple authors of the study, Lo included, disclosed that they receive funding from, hold equity in, and/or consult for Grail.
The company has not disclosed whether or how it might incorporate analysis of DNA fragment end coordinates into its development of early cancer detection tests, but has reported on analyses of cancer-associated mutations, and has said it is also exploring DNA methylation and gene expression signals.