NEW YORK – Having officially launched its Shield colorectal cancer screening assay last week as a laboratory-developed test, Guardant Health has become the first of several companies expected to compete in a new field of blood-based cancer detection.
As such, the firm stands to benefit from a head start in marketing and other aspects of clinical establishment. But it has also become a target for skepticism from competitors on its heels.
On a call discussing the company's first quarter financial results last week, co-CEO AmirAli Talasaz addressed questions raised recently by one of its main competitors, Exact Sciences.
"Throughout the history of Guardant, we have dealt with skeptics who claim that a blood test cannot add value in oncology. And every time they have been proven wrong," Talasaz said.
"We believe our Shield blood test has the potential to be much more effective in reducing CRC mortality and adding more life year gains than … stool-based screening," he added.
On its own Q1 call in April, Exact Sciences, which currently markets the field's leading stool-based CRC screening test, Cologuard, and is developing its own blood-based competitor to Guardant's Shield, took aim at Guardant, questioning the likelihood of blood-based tests competing directly with Cologuard.
Exact CEO Kevin Conroy argued that regulatory and reimbursement hurdles could hamper clinical adoption of blood-based testing. He suggested that low sensitivity for advanced adenomas (a colorectal cancer precursor) would mean blood tests could fail to be endorsed for use every three years, requiring use every year instead. This would then translate to significantly lower reimbursement rates, which would undercut gross margins.
He also predicted that persuading physicians to order a "less accurate" test would be a challenge. Using internal modeling, Exact estimated that blood-based cancer screening tests with an overall sensitivity of 80 percent would have stage I sensitivity of 70 percent and advanced adenoma (AA) sensitivity of just 15 percent. In contrast, Cologuard is 90 percent sensitive for stage I cancers, and 45 percent for AAs.
Guardant last week painted a much rosier picture, claiming that internal studies have demonstrated overall test sensitivity of 91 percent and stage I sensitivity of 90 percent — equivalent to Cologuard. According to the company, AA sensitivity for the assay is 20 percent, which still lags the stool test but is better than suggested by Exact's modeling.
Guardant hadn't previously assessed its test in AAs, and Talasaz said the new addition is the result of ramping up R&D over the past year to add a proteomics component to Shield's core of epigenetics and fragmentomics.
The firm conceded that these numbers are from a case-control study in a set of just 309 samples, rather than a prospective trial — a question Exact raised in its own call, arguing that test performance is all but guaranteed to drop in a prospective setting compared to retrospective data.
But Talasaz argued that Guardant has a good chance of keeping its numbers up when it reports data from its ECLIPSE trial later this year.
"We continue to see reproducibility over time and over different reagent batches, all the things that basically could contribute to technical variation," he said. In addition, the way samples were collected as part of the ECLIPSE trial suggests current specificity is likely to hold up.
Another of Exact's qualms focused on US Preventive Services Task Force procedures for issuing recommendations on cancer screening tests. Conroy described the group's cost/benefit modeling strategy as being "largely dependent on advanced adenoma detection." Assuming 100 percent compliance, which USPSTF has done in previous assessments, a test like Shield would be predicted to avert fewer cancer cases and prevent fewer deaths than stool testing.
According to Talasaz though, it's not impossible that USPSTF would consider compliance or adherence in future assessments of blood-based tests.
"Lack of compliance is on top of many people's minds, so it's something that doesn't have a high bar of education or awareness around it," he said, citing a study from Harvard Medical School in 2019 which compared effectiveness of different CRC screening tests modeled at 100 percent adherence rates compared to reported real-world adherence rates.
While DNA-based stool testing (Cologuard) outperformed a blood test called SEPT9 in preventing cancers and deaths when assuming 100 percent compliance, the reverse was true when real-world adherence was modeled. And this was despite SEPT9's low performance compared to what Guardant expects from Shield — just 68 percent CRC sensitivity and 22 percent AA sensitivity, with only 80 percent specificity.
In light of this, "we believe our Shield blood test has the potential to have a higher benefit assessment by USPSTF than … stool-based screening tests," Talasaz said.
"Now we need to do our work to make sure we continue to publish papers on the modeling and show the importance of adherence, which is the reality of the business we are running. That's the way to save lives — to make sure patients comply and adhere to [testing] year after year," he added.
In his discussion of the evolving space, Exact Sciences' Conroy also suggested that failure to demonstrate high enough benefit for an every-three-year testing schedule would mean a severe drop in reimbursement compared to Cologuard.
But Talasaz said Guardant remains confident it will be reimbursed equivalently.
In a memo issued in January 2021, the US Centers for Medicare and Medicaid Services laid out a national coverage determination proposing to reimburse any US Food and Drug Administration-approved blood-based CRC screening test at an interval of every three years, as long as it reached sensitivity and specificity of at least 74 percent and 90 percent, respectively.
Under that rubric, which doesn't necessitate sensitivity metrics for advanced adenomas, Guardant's assay would be expected to meet the coverage criteria, despite Exact's protestation that without strong adenoma detection a test would only be paid for as a yearly screen.
Guardant appears to have little doubt that CMS's position on this will change. "An NCD is an NCD," Talasaz said. "With the specificity of 90 percent, our understanding is [that] they want to make sure that the specificity is better than Cologuard."
He added that the company doesn't expect the FDA approval to convey specific mandates regarding test frequency, saying that in its experience, the agency's considerations hew tightly to safety and efficacy.
Guardant initially projected it would be able to share readout from the ECLIPSE trial by the end of the first half of this year, but, because of some delays in pathology reporting, final results have been delayed until what the firm hopes will be this September or October.
That said, the company has already begun a modular submission to the FDA and is confident it can file for premarket approval by the end of this year.
Post-approval, Guardant will then have the chance to seek pricing of its test as an Advanced Diagnostic Laboratory Test under the Protecting Access to Medicare Act of 2014.
"With PAMA, we are confident that we will be able to secure Medicare pricing of over $500," Talasaz said.
To support that price-setting process and as it works to secure Medicare coverage and convince private payors to follow in CMS's footsteps, Guardant is offering patients a self-pay price of $900.
Talasaz added that in Guardant's current LDT "launch phase," the firm has about 100 employees in its commercial team for Shield. As it gets closer to USPSTF guideline inclusion, it expects to scale that to about 700 people.