NEW YORK (GenomeWeb) – Guardant Health and collaborators at the John Wayne Cancer Institute presented data this week on the firm's targeted sequencing-based circulating tumor DNA assay at Cambridge Healthtech Institute's Molecular Medicine Tri-Conference in San Francisco.
Guardant Health launched its assay, Guardant360, last year as a laboratory-developed test from the firm's CLIA certified and CAP accredited laboratory in Redwood City, Calif. Guardant360 initially screened 54 oncogenes, but has since been expanded to include 68 genes and four gene rearrangements. It can detect a range of mutation types, including SNVs, indels, copy number variants, and fusions.
In a presentation at Tri-Con, AmirAli Talasaz, president and COO of Guardant Health, discussed details of the assay including its analytical validation; data from its 510-patient multi-center and multi-cancer clinical validation study and its first consecutive 1,500 commercial cases; as well as some specific patient examples. Talasaz said that the firm is currently preparing some manuscripts for its internal validation studies and that its partners will also be submitting papers for publication in a peer-reviewed journal this year.
Guardant recently completed a study showing that its test impacted patients' enrollment into clinical trials, their eligibility, and time to treatment. It is now in the midst of starting a clinical outcome study, which will be a multi-drug, multi-arm cell-free DNA biomarker-driven trial.
In an interview with GenomeWeb following his presentation, Talasaz said that the company now processes hundreds of samples per month from over 100 hospitals and cancer centers across the US. Talasaz said that although Guardant is not currently pursuing clearance from the US Food and Drug Administration, the company "welcomes oversight" from the FDA. "Quality is very important, and while we haven't had any discussions with FDA yet, we haven't ignored its expressed intention to regulate LDTs," he said.
The Guardant360 test has a list price of $5,400 and Talasaz said that the firm is being reimbursed from private payors at "similar levels" to companies that offer tissue-based tumor profiling tests. Foundation Medicine, for instance, has said it is reimbursed an average of $3,600 per test.
Researchers have long known that tumors shed DNA into the bloodstream, and that those fragments of DNA are detectable in patients' blood. But only recently has the idea of analyzing that DNA via next-generation sequencing seemed feasible for a clinical setting, due to improvements in the technology's accuracy as well as its drop in price.
At the JP Morgan Healthcare Conference in San Francisco last month, several companies including Illumina, Thermo Fisher, Sequenom, and Foundation Medicine announced their intention to develop NGS-based ctDNA assays, highlighting the fact that such tests are viable for patient care.
So far though, Guardant Health and Personal Genome Diagnostics, a spinout from Johns Hopkins University, are the only two firms to successfully commercialize ctDNA tests that run on NGS platforms.
Guardant Health's test relies on proprietary technology it calls digital sequencing. In short, after it extracts the fragments of cell-free DNA from the blood, it uses unique barcodes to tag each single strand of a DNA fragment prior to enrichment and sequencing.
Bioinformatics are a key component to the assay. Talasaz explained that while a significant amount of noise is introduced during sample prep and sequencing, the company is able to model that noise and account for it, yielding higher accuracy and a lower limit of detection of 0.1 percent.
During his Tri-Con presentation, Talasaz noted that in 150 samples across five cancer centers, the firm compared cfDNA to tissue-based sequencing, showing that cfDNA had a sensitivity, specific, and diagnostic accuracy of 86 percent, 98 percent, and 97 percent, respectively. However, upon manual evaluation, Talasaz said that it appears that the cfDNA test was correct in the majority of discordant cases.
Next, the firm conducted a clinical validation study in 510 patient samples, of which 504 met the quality criteria necessary to be tested. Talasaz said that this is a much higher success rate than for tissue-based tumor sequencing tests, in which 50 percent to 90 percent of samples pass criteria and are able to be analyzed.
Of those 504 samples, 86 percent had at least one alteration and 75 percent of patients had at an actionable alteration.
In the 1,500 commercial cases, Talasaz said that these metrics have held up, with 77 percent of patients having at least one alteration and 75 percent having an actionable alteration. None of the identified alterations were found in 150 healthy controls.
Talasaz said that the majority of customers are using the assay to assess tumors in advanced cancer patients in order to determine a potential treatment. For example, the Guardant360 test identified a druggable BRAFv600e mutation in a 62 year-old male with late-stage metastatic cancer. The patient had previously been screened for that mutation, but it had been missed. He was able to receive targeted therapy, and showed a "huge response," Talasaz said.
In another case, a 66-year-old woman who had originally been determined to have HER2 negative breast cancer was screened via Guardant360 after she continued to progress after standard-of-care treatments. The assay found that her cancer had in fact changed to HER2 positive and she began targeted therapy with Herceptin. She responded well to the drug and subsequent blood samples showed the tumor level decreasing.
Currently, customers are primarily use the Guardant360 test on patient samples from a single time point, however, Talasaz said that another application is to serially monitor patients' response to drugs or check for residual disease following surgery or treatment.
Dave Hoon, director and professor of molecular oncology at the John Wayne Cancer Institute and an early adopter of the Guardant360, presented on such applications at Tri-Con.
Hoon and his colleagues have tested the assay in melanoma patients retrospectively at different time points throughout patients' disease. For instance, they analyzed ctDNA from the blood of stage III melanoma patients collected at multiple time points following surgical resection, demonstrating that analysis of cell-free DNA correlates with patients' disease progression and may even be able to identify disease recurrence before standard imaging methods.
In one patient, plasma samples were collected at surgery, and then at seven, 10, 13, 18, 20, 22, and 27 months post-surgery. The Guardant360 assay was run retrospectively at each of those time points and the evolution of that patient's disease could clearly be seen, Hoon said.
For instance, prior to surgery, mutations to the BRAF and EZH2 gene could be seen at relatively high levels in the blood. Immediately after surgery, those mutations dropped to zero, where they remained for 10 months after surgery. Just after 10 months, the levels of one mutation started to rise. Soon after, the patient was diagnosed with stage IV recurrence and another surgical resection was performed at 18 months.
Hoon said that imaging was not sensitive enough to detect disease at 10 months, when the signal in the ctDNA first started to rise. Then at 27 months, the two original mutations as well as new mutations began to rise quickly. "Everything goes crazy," Hoon said.
Being able to detect residual melanoma after surgery is critical to patients' survival, and cfDNA may outperform better current imaging methods in diagnosing residual disease and progression, Hoon said. Cell-free DNA is providing "important information as a surrogate to disease progression events." In the future, he added, cell-free DNA monitoring could "allow clinicians to know if disease has been fully resected" and could "identify in real time new druggable gene targets during tumor progression."