NEW YORK (GenomeWeb) – Guardant Health this week said that it has made new improvements to its Guardant360 comprehensive liquid biopsy test.
Although it has not yet made public the validation data supporting these numbers, the company now lists a reportable range for its assay down to 0.04 percent mutant allele fraction for SNPs and 0.02 MAF for indels.
The firm also updated the Guardant360 panel to 73 genes from 70 in its most recent iteration. Guardant CEO Helmy Eltoukhy told GenomeWeb this week that the content improvements also include a more comprehensive representation of gatekeeper, or acquired resistance, mutations, expanding the power of the test to inform patient management at different time points in their disease progression and care.
"We are very aggressive about learning from every sample we sequence so we can continue to have best-in-class performance in liquid biopsies," Eltoukhy said.
According to Guardant, the newly launched enhancements are the direct result of having access to nearly 30,000 commercial samples since first introducing Guardant360 in 2014.
As it has processed these samples, Guardant has also learned from them, Eltoukhy said.
"What we find is, when you first launch an assay, you have studies on hundreds, or maybe low digits of thousands of samples. If you think about detecting something at extremely low frequency, if you have [a variant] that occurs at a rate of 1 in 100, you're likely not going to see that in a validation of 200 samples, [but] as you expand to thousands of samples you start powering the analysis for things that occur at 1 in 100 or even 1 in 1,000 frequency," he said.
Over the last two years of operation, he explained, the company has used advanced analytic techniques to look at the data it has produced below its reportable range.
"We essentially start categorizing things: Are these errors from the process? Is this DNA damage? Are there artifacts from bioinformatics in the back end? And with nearly 30,000 commercial samples and many more research samples, we've been able to lower the limits of where we go," he said.
According to Eltoukhy, the improvements will hopefully help to continue to distinguish Guardant360 from emerging and existing competitors, as the blood-based cancer sequencing field has grown considerably since Guardant's launch in 2014.
Among others, Resolution Biosciences announced its test in mid-2015, Genomic Health launched its liquid biopsy panel, Oncotype SEQ, in June, Foundation Medicine launched its FoundationACT in May of this year, and Personal Genome Diagnostics moved into clinical liquid biopsy testing this October.
Representatives from several of these firms joined a workshop organized by the US Food and Drug Administration and the American Association for Cancer Research earlier this year, where they grappled with questions about how best to demonstrate the validity and utility of this new class of test.
As more and more blood-based cancer sequencing assays have been launched, more attention is being paid to the analytical and clinical performance claims of these tests, as well as to how they compare to one another.
Genomic Health recently shared the first analytic validation data for its liquid biopsy panel, Oncotype SEQ, in a poster at the European Society for Medical Oncology annual meeting, reporting that the test had a per-sample sensitivity down to 0.1 percent mutant allele fraction for single nucleotide variants, or below three copies for CNVs, with more than 99 percent specificity.
According to Genomic Health, Oncotype SEQ was able to maintain a 95 percent detection rate as long as tumor DNA was present with a frequency of at least 0.19 to 0.56 percent compared to background DNA, depending on the type of genetic alteration in question.
For its FoundationACT test, Foundation Medicine reports 99 percent sensitivity for base substitutions at 0.5 percent MAF or higher; and insertions, deletions, fusions, and rearrangements down to 1 percent MAF.
Academic collaborators have published data on Resolution Biosciences' lung cancer panel showing that it could identify mutations present in DNA dilutions down to 0.4 percent MAF with 100 percent sensitivity and specificity. The company also reports that its lower level of detection is 0.1 percent.
When Guardant first published the analytic validation of Guardant360 in PLoS One last year, it didn't compare sensitivity numbers for all categories of variants in a single chart, but the company's spec sheets reflecting its newest updates do include this type of granularity.
The company said its reportable range is down to 0.04 percent MAF for SNVs and fusions, to 0.02 percent for indels, and to 2.12 copies for copy number variations.
The test has at least 99.9 percent sensitivity as long as SNVs and indels are present at more than .25 percent, and fusions at more than 0.3 percent.
If SNVs or indels are present at between 0.05 and 0.25 percent, analytical sensitivity drops to 63.8 percent for SNVs or 67.8 percent for indels. It also has 83 percent sensitivity in detecting fusions below 0.3 percent MAF.
The validation data from the studies Guardant performed to produce these numbers have not been made public, but Eltoukhy said that the company intends to publish it in the future.
In terms of differentiation from other tests, Eltoukhy argued that there is another important data point coming out of Guardant's 30,000 samples, which is that the company has determined that about 50 percent of all the alterations it has reported have been present at frequencies below 0.4 percent.
This suggests that tests that don't detect variants below this threshold could mislead the field in terms of the overall clinical utility of liquid biopsy. In other words, lack of concordance between other liquid biopsy tests and tissue results might be due to the fact that the tests are missing what appears (based on Guardant's data) to be a significant proportion of actionable alterations.
"With tests with reportable ranges only down to only 0.5 percent or 1 percent, [our data suggests] that that doesn't mean much clinically, and I worry about tests out there creating disillusionment," Eltoukhy said.
As Guardant moves forward with promoting the new improvements to its test, Eltoukhy also highlighted the fact that through its longer history on the market, not only has Guardant been able to collect data that allows it to push the limits of sensitivity now significantly farther than other companies, but it's also been able to collect and release more data on its test, including not only analytical validity numbers, but also clinical utility evidence.
"We are the only company with outcomes studies … showing that where we find a biomarker with an associated targeted therapy, that patients actually do respond at rates that are comparable or even slightly higher than some of the tissue-based studies that were done," Eltoukhy said.
For example, Guardant began a therapy matching trial right after it launched its test, focused on lung and gastric cancers. In data presented at the American Society of Clinical Oncology annual meeting, researchers from Samsung Medical Center and Guardant reported that lung cancer patients in this NEXT-2 trial had an 88 percent response rate and 94 percent disease control rate when matched to a targeted therapy by their Guardant360 results. Gastric cancer patients had a 60 percent response rate and 90 percent disease control rate on average.
In addition to the updates that Guardant said increase the test's sensitivity and breadth, the company also said that it has simplified its report interpretation and expanded access to online support services for patients and doctors with an iPhone app.