NEW YORK – Although it has been selling clinical ctDNA analysis for several years, this month marks the first time that liquid biopsy firm Biocept has published details of its proprietary background-blocking PCR method or data on its analytical and clinical validity in the scientific press.
The move comes at an opportune time, as Biocept is hoping to expand its business from running tests through its CLIA lab to offering its circulating tumor DNA detection technology as research use-only kits that other labs could validate and utilize.
In the study in PLoS One earlier this month, company researchers presented a detailed description of the company's Target Selector approach, including a wild-type suppression method the firm has come to call "switch blocker." They also reported strong reproducibility, and analytical sensitivity and specificity across five genes, matching closely some of the other high-sensitivity PCR methods that are also being used in the clinic and/or marketed as kit products.
The study data reflects a standard analytical validation, using contrived samples in which a certain, known amount of mutant DNA is spiked into a background of wild-type DNA. Overall, the firm tested more than 3,000 samples as part of this process, focusing on five "important and clinically actionable targets" in three genes that are among the menu of the tests it offers clinically.
More than 600 samples were used to demonstrate analytical sensitivity, and each of the five targets studied — EGFR exon 19 deletions, L858, and T790 mutations, BRAF V60, and KRAS G12/G13 — demonstrated an analytical sensitivity of 0.02 percent or better, the authors wrote. And only a single false negative, from a BRAF assay, was observed in 667 tests performed.
Another 560 tests were performed to calculate analytical specificity, which Biocept reported was greater than 99 percent for all the target alterations. The company had two false positives across the assays after performing its full clinical testing protocol of qPCR, followed by Sanger sequencing and melt curve temperature cutoffs, which it uses to cull any non-specific amplification products.
"We hope this highlights the strength of the Target Selector technology," Jason Poole, Biocept's VP of research and development and first author of the study, said this week. "Once you have an optimized assay, it's a very strong test," he explained.
In the study, Biocept also performed a small clinical specificity experiment, running each of the assays on 20 samples from healthy individuals. Hearteningly, no mutations were detected in this group.
Although 20 controls may seem small, Poole argued that there are diminishing returns, at least in the context of a mainly technical validation in which the goal is to double check that a test isn't being confounded by the size constraints of cell-free DNA.
"You could probably find a reason to test it in several hundred patients. But with 20 patients, you get a good idea of whether or not your assay performs in [the fragment size range of] cell-free DNA, which is that 170-base-pair range," he said.
The current study did not include any clinical validity data, something the authors caution in the paper itself. Poole and coauthors did, however, reference other clinical performance data in the current study — offering a more indirect take on the accuracy of Biocept's testing in the clinic over the last few years.
"For me, this is a major part of the paper," Poole said. "It's not a true clinical validation ... but [it is some] validation, for me, that the test is working as it should."
Overall, the study authors wrote that "collective results from CLIA lab testing performed on over 2,000 patient samples using the described Target Selector ctDNA assays … have similar detection rates compared to the prevalence of the corresponding mutations found in the literature."
Amongst other details, Poole and colleagues wrote that Target Selector KRAS detection rates of 20.16 percent in Biocept's clinical testing history fit well within the reported prevalence of KRAS mutations in tissue, which ranges from 20 percent to 25 percent in the Western Hemisphere.
Similarly, "for BRAF V600E, the frequency is 1–2 percent," the authors added, "and the Target Selector BRAF detection rate is observed at 1.69 percent."
Poole also highlighted data the company has been able to collect in ethnic subpopulations, further cementing that Biocept has been successfully detecting the expected rates of mutations as it has tested patients clinically.
"People of different backgrounds will have different levels of activating mutations for EGFR," Poole explained. "In Asia [or Latin America] you might have 35 to 40 percent rate … but if you have European Americans, it may only be 10 to 15 percent."
"With North America, and the US in particular being a melting pot, it's nearly impossible to know, unless you have the ethnicities of all your patients, what your EGFR activating mutation rate would be," Poole added. Nevertheless, "we do match what you would expect to find in North America," based on the large studies that have been done.
According to Nall, Biocept has been lucky to be able to conduct significant numbers of tests in Latin American populations under an agreement with a large pharma partner, whereby Biocept's testing is being used to qualify patients there for targeted therapies.
In the study, Poole and his coauthors compared EGFR mutation frequencies that the company saw over a 12-month period using Target Selector for patients in Mexico and Colombia to rates reported in a 2015 paper in the Journal of Thoracic Oncology, which included data from several thousand tumor tissue tests.
Biocept's rates of 39.87 percent for EGFR activating mutations in Mexico and 26.56 percent in Colombia, correlated strongly with the tissue prevalence described in the cited study: 36.7 percent in Mexico and 25 percent in Columbia.
According to Nall, part of the reason Biocept decided to make the move to developing and selling kit versions of its Target Selector tests was because it believes that it is CLIA lab experience has really cemented the validity of its technology, giving it high confidence of how it will perform in the hands of others.
"The beauty of having the CLIA lab and having all this experience in the clinical world [is that you can make sure] that the test is pressure tested before you put it in the hands of somebody else. So now that we have the kit, I think we all feel very solid that this is ready to be used routinely throughout the world for clinical use," he said.
As a liquid biopsy kit provider, Biocept will be competing against a different slate of companies than it faced in the CLIA testing market. The most analogous competitors include Bio-Rad, whose Droplet Digital PCR (ddPCR) has already been taken up for clinical liquid biopsy testing by institutions like Mayo Clinic, as well as other high-sensitivity tools like Sysmex Inostics' BEAMing — a magnetic bead-based digital PCR platform — and Biocartis' quantitative real-time PCR technology, Idylla.
With signals from the clinic that liquid biopsy may now be poised for broader adoption, some different products have now begun to be tested against one another head to head. A study this June, for example, compared BEAMing and Idylla KRAS kits, finding that they were largely concordant, though with differences at low allele fractions that could have clinical implications.
Another paper last month put BEAMing up against ddPCR, showing good agreement between the two and "suggesting sufficient reproducibility for clinical use."
So far Biocept's Target Selector assays haven't been part of this kind of direct technology comparison Poole said. "As one of the original folks out with liquid biopsy, a lot of our work to date has been comparing to tissue. But now that it's gotten to be where … people are starting to use [liquid biopsy] more routinely, I think the timing is right."
One interesting thing coming out of other platform comparisons, Poole added, is that PCR assays look to be hewing fairly closely to one another in recent head to head comparisons, while comparisons of next-gen sequencing liquid biopsy tests have sometimes found troubling discordance. "That's one of the things that I appreciate about the Target Selector is that it is more depth versus breadth," he said.
Concordance aside, Poole argued that there are other reasons that users may consider in picking one platform over another. In the study, for example, he and his coauthors highlight that the Biocept approach offers a middle ground between the breadth of NGS and the acute sensitivity of tools like ddPCR.
"While ddPCR is an extremely sensitive methodology it necessitates design and testing for each specific mutation within a hot spot region. For instance, evaluating the most common seven mutations in KRAS exon 2 (codons 12 and 13) requires running at least seven separate assays. Input DNA is required for each SNV, which may exhaust the available sample. In contrast, a single Target Selector assay covers the same two-codon hotspot with … sensitivity for all SNVs," the group wrote.
"It's a good [middle ground] … between the single alleles of droplet digital PCR, where if you have a target like KRAS you need to do testing on 20 different markers [and] you run out of sample ... versus NGS where you can test 70 genes but you don't get the sensitivity you need," Poole added.
As Biocept seeks to make the case for adoption of its kits, especially as it expands that menu beyond EGFR, Nall agreed that clinical validity data for the company's in-house CLIA testing could also help give it a boost.
"As a small company ... we don't have just a team dedicated towards this like a larger firm does. Our clinical and scientific affairs are pretty much the same people who do all the work," Nall said. "But I would expect that we'll be able to get that data out in a peer reviewed publication in the relatively near term," he added.