NEW YORK (GenomeWeb) – Fresh off a $225,000 grant from the Mark Foundation for Cancer Research, a group led by Weill Cornell Medicine and the New York Genome Center researchers has developed a whole-genome sequencing assay to track cumulative patterns of genetic mutations in a patient's bloodstream. The researchers believe the assay could eventually be used to assess minimal residual disease status in patients after tumor resection.
The team has also begun to commercialize the test to look at cell-free DNA (cfDNA) in patients as part of a spinout led by Weill Cornell and NYGC.
Researchers have in recent years found that they can use cfDNA to track somatic tumor mutations in a patient's blood but have struggled to sensitively detect certain cancer types because of lower tumor fractions and burdens in patients, leading to challenges with current deep targeted sequencing tools.
"There is sort of an absolute ceiling on sensitivity that is imposed by the number of genomic equivalents available in a standard plasma sample," said Dan Landau, principal investigator on the grant and an assistant professor at Weill Cornell and core faculty member at the New York Genome Center. "That means that even if the molecular method for deep targeted sequencing has higher sensitivity, we'd be limited by the number of fragments that cover the targeted genomic site. In most cases of low burden disease, this limits the sensitivity to a fraction of circulating tumor DNA of about one to a thousand."
Landau's team therefore developed a method that focuses on a broad approach by examining the whole genome in a patient's tumor sample and using an algorithm to identify cumulative mutation patterns and region-specific changes in cfDNA copy numbers. He explained the algorithm initially identifies mutations from a patient's tumor after they undergo resection. Researchers then sequence about a 5-ml blood sample using Illumina's HiSeq X10 or NovaSeq systems, and attempt to match any mutations in the blood sample that were also in the tumor sample.
"Instead of looking at a few sites with very deep sequencing, we look at tens of thousands of mutations across the cancer genome with more shallow sequencing, and we integrate the signal across the entire genome," Landau said. "That allows us to radically transform the detection sensitivity, we [have shown] that using this method we can get up to two orders of magnitude enhanced sensitivity."
Landau claims that by removing certain data limitations, his team's assay detects the frequency of tumor DNA in cfDNA down to a limit of detection of 1 in 100,000.
"Essentially what we're doing is measuring the dilution of the mutation signal in the sample of the plasma cell-free DNA," Landau explained. "So the more signal you have, either in the form of high mutation rate or high copy number rate, the more likely we're able to extract that signal, even in ultra-low circulating tumor DNA fractions."
At the 2018 American Association for Cancer Research Meeting in Chicago, Landau and his team presented a poster reporting the assay's use on non-small cell lung cancer samples (NSCLC), integrating genome-wide mutation signals to obtain a tumor fraction estimate.
"Benchmarking on artificial plasma showed tumor fraction detection sensitivity as low as 1:100,000, two orders of magnitude more sensitive than currently available methods," the authors noted.
In the study, Landau and his team performed WGS on resected NSCLC samples and matched germline samples of eight NSCLC patients, as well as on matched pre-and post-surgery cfDNA.
The researchers identified pre-surgery circulating tumor DNA (ctDNA) in all the early-stage pre-operative samples, as well as in about 40 percent of post-operative patients, which the group correlated with post-operative disease progression.
Landau and his team then trained a custom machine learning framework to distinguish between cancer-altered sequencing reads and reads altered by sequencing errors. The team performed genome-wide pattern matching to a specific genomic signature that marked lung cancer mutations, which they noted indicated the presence of ctDNA in the patients' plasma.
According to Landau, the overall process — from sample collection to sequencing results — takes between two and five days to produce results.
In addition to NSCLC, Landau said that his team has been gearing the technology to monitor tumors that are "challenging to established methods " and have "a high mutation rate or a high copy number variation rate," including melanoma, colon, and other unspecified cancer types.
Landau said that his team has filed a series of patents related to the sequencing technology and will use the funding from the Mark Foundation for supporting large-scale clinical trials. Weill Cornel and the NYGC are jointly pursuing efforts to establish a startup to commercialize the liquid biopsy platform.
Landau's team at the NYGC also previously received a grant from the American Lung Association in 2018, which it will use to further improve the assay's limit of detection.
Other commercial and academic entities have launched sequencing-based assays to monitor tumor mutations in a patient's bloodstream. Examples include Natera with its Signatera platform, Guardant Health with its Lunar assay, and well as a group of Johns Hopkins researchers who have developed a detection method called the Safe-Sequencing System.
In contrast to other methods for monitoring recurrence in patients, Landau argued that his team's tool defines the "personalized compendium of mutations" found in the patient's tumor. He said that the assay then searches these specific mutations in cfDNA across the entire genome, removing the cap of sensitivity imposed by the genomic equivalent limit.
"Targeted methods are highly successful in certain settings, such as understanding spatial diversity of clones in a patient with metastatic disease or understanding emerging resistance variants," Landau noted. "In other contexts, where there is a low burden of disease, [however], we would need methods that provide greater sensitivity to be able to use them for continuous monitoring of disease."
Landau and his team also plan to work on a similar method for early detection of cancer rather than monitoring the condition after resection. In addition to working with Weill Cornell Medicine to use the technology in different studies, Landau's group is collaborating with multiple academic partners at Memorial Sloan Kettering Center, Massachusetts General Hospital, and the Francis Crick Institute in London.
Overall, Landau believes the assay will "arm the clinicians and patients with real-time information to guide decisions about adjuvant therapy or ongoing therapy."