Asymptomatic COVID-19 testing pushes forward, with lessons from cancer
As broad COVID-19 screening gains momentum, early cancer detection provides a playbook
The growing demand for asymptomatic COVID-19 testing is shifting the conversation from one-time diagnosis to ongoing screening, and from single results to population outcomes. Companies moving the ball forward for early detection of cancer, some of which have jumped into COVID-19 testing, offer clues to what will be needed for broad uptake.
While diagnostic testing aims to classify disease in sick patients, screening seeks to uncover issues in seemingly healthy individuals. The same test could be used either way or tuned to serve one purpose over the other.
So far, SARS-CoV-2 detection has primarily followed the diagnostics playbook, particularly in the U.S.
But rising case numbers, pressure to open schools and workplaces and the imminent arrival of flu season are increasing calls for tests that can identify presymptomatic or asymptomatic individuals capable of infecting others.
In its July testing plan report, The Rockefeller Foundation argued for a paradigm shift that distinguishes screening tests from diagnostics, saying the former should prioritize testing speed, volume, cost efficiency and frequency over sensitivity, and called on FDA to “re-think” its sensitivity requirements for screening tests “to make sure the perfect is not the enemy of the good.”
Because their target populations are large and healthy, screening tests are subject to more rigorous regulatory scrutiny than diagnostic tests in terms performance, utility, risks and cost-effectiveness. In the case of COVID-19, one of FDA’s goals is to minimize false negatives by setting a high bar for sensitivity, since in broad asymptomatic populations, even a small percentage of misses could give a large number of people a false sense of security.
“The alternative to testing with low sensitivity is no testing, which has 0% sensitivity.”
Although asymptomatic individuals have been able to access FDA-authorized COVID-19 tests at doctors’ discretion and via research studies, only one of the more than 100 molecular tests on the market is cleared for use in people not suspected to have the disease. On July 24, the agency gave its first Emergency Use Authorization (EUA) for COVID-19 testing of asymptomatic individuals to a molecular test from Laboratory Corp. of America Holdings (NYSE:LH), but only if they have a prescription from a physician.
Asymptomatic screening has also been limited by the slowness of the dominant testing technology, reverse transcription polymerase chain reaction (RT-PCR) -- a sensitive molecular test that detects viral RNA and is primarily run in centralized labs. Strained testing infrastructure has stretched the turnaround time for RT-PCR results by several days and even weeks, which renders results irrelevant, and samples from asymptomatic individuals are generally deprioritized.
Given the backlog on RT-PCR, NIH and HHS are seeking to test broader swaths of the population by embracing tests that are more efficient, but typically less sensitive. These include rapid point-of-care technologies, which return results within an hour and cost less than lab-based assays, and pooled RT-PCR testing, which combines multiple samples in the same tube to boost volume and cut costs (see “NIH, HHS Fast-Tracking Point-of-Care, Pooled Tech”).
And on Tuesday, seven states announced a partnership with The Rockefeller Foundation to purchase and deploy more than three million rapid point-of-care antigen tests to help detect outbreaks more quickly and expand long-term testing in schools, workplaces and nursing homes (see “States Bet on Antigen Tests for Asymptomatic Screening”).
Given the strains on testing resources, the argument to lower the sensitivity bar for SARS-CoV-2 screening tests is resonating even among those developing highly sensitive assays.
“The alternative to testing with low sensitivity is no testing, which has 0% sensitivity,” said Guardant Health Inc. (NASDAQ:GH) CEO Helmy Eltoukhy. The cancer testing company is seeking EUA for an NGS-based COVID-19 test it developed to screen its workforce; according to Eltoukhy, the test is more than 50 times more sensitive than widely used RT-PCR assays.
A similar logic underlies the growing field of early cancer detection, in which companies like Guardant are developing routine, non-invasive tests to catch tumors before they progress.
“The economics are going to have to support that kind of use.”
Since avoiding false alarms is a top priority in cancer screening, minimizing that risk requires setting stringent detection cutoffs. Although that means some cancers will be overlooked, more will be caught early than they would have been otherwise (see “Sensitivity, Specificity Tradeoffs for Cancer Screening”).
Companies in that space told BioCentury screening tests have a better chance of taking off if developers position them as complementary to more established approaches, build an evidence base to support their actionability, and initially gear the tests toward high-risk groups.
“The economics are going to have to support that kind of use,” said Lance Baldo, CMO at Adaptive Biotechnologies Corp. (NASDAQ:ADPT), a TCR sequencing company also developing a COVID-19 test. “You can have the best test in the world, but if it’s not being applied in a practical way, it’s not going to help the individual or society.”
Test early and often
There is growing consensus that consideration of test performance should go beyond the sensitivity and specificity of individual assays, and take into account the frequency of testing.
The idea has been picking up steam following the publication of a June medRxiv modeling study, which argues that if a test is accessible and fast enough to routinely screen large numbers of people, it’s acceptable to miss some cases in a given batch because they’re likely to be caught by the next test.
“When you think about clinical testing, you really care about this particular test on this individual,” said Roy Parker, who coauthored the study and is director of the BioFrontiers Institute at University of Colorado Boulder. For screening, he said, “it’s about how am I using this test, or combination of tests, in order to catch infections before they spread.”
The case for less sensitive tests is also bolstered by the kinetics of SARS-CoV-2 infections, he said.
The concentration of viral RNA in the respiratory system is thought to shoot up exponentially and then decline gradually. This creates a short window at the beginning of the infection in which a more sensitive test will catch a case that a less sensitive test misses.
Highly sensitive tests will catch more patients at the tail end of infection, but since viral RNA can persist long after symptoms resolve and the patient is no longer infectious, labeling someone COVID-19 positive at that point is less useful, he said.
A key question is whether the viral kinetics that form the basis for the model also hold true for asymptomatic patients, on whom there is much less data.
The only way to answer that question is through more asymptomatic testing, said Sherlock Biosciences Inc. CEO Rahul Dhanda, whose company has developed a CRISPR-based test for the virus.
While FDA’s standards for asymptomatic screening tests set a sensitivity cutoff of 95% percent positive agreement (PPA) with an authorized RT-PCR assay, the agency acknowledges that testing frequency can compensate for lower sensitivity.
“If screening using a highly sensitive test is not feasible, serial testing on different days or with different tests could be considered,” FDA wrote on its website for frequently asked questions about SARS-CoV-2 testing.
The compact between seven U.S. states and The Rockefeller foundation has put antigen tests at the center of the debate about using cheaper, lower-sensitivity tests to screen asymptomatic populations. The partners are in discussions with the two companies with FDA authorized antigen tests, Becton Dickinson and Co. (NYSE:BDX) and Quidel Corp. (NASDAQ:QDEL), and hope the demand demonstrated by their cooperative purchasing agreement will incentivize more test developers to bring the technology to the market.
Unlike molecular testing technologies like RT-PCR, antigen tests are immunoassays that detect viral proteins instead of RNA. They can be run on rapid lateral flow “dipstick” formats by non-experts, similarly to pregnancy tests (see “Diagnostic Tech Tableau”).
The first data reported for BD and Quidel’s SARS-CoV-2 antigen tests suggested they were both less sensitive than molecular tests on the market (see “A Test for the Utility of Antigen Testing”).
But Quidel CEO Doug Bryant told BioCentury his company’s test closed that gap in subsequent studies, once FDA lifted its requirement that manufacturers evaluate tests using samples diluted in viral transport media, and more dry swabs became available for test validation.
The clinical sensitivity for Quidel’s antigen test, originally reported as 80%, is now listed as 96.7%, above FDA’s cutoff for asymptomatic screening. The test’s limit of detection (LoD), a measure of analytical sensitivity, is still about 1,000 times higher than those of FDA authorized molecular tests reporting values in the same units (see “Limits of Detection for New Coronavirus”).
Guardant’s Eltoukhy thinks there won’t be a single ideal product profile for COVID-19 screening tests, and practical factors like the how many tests are feasible to run in a week will play a role. For example, a daily saliva test can afford to be less sensitive than a weekly test using an uncomfortable nasal swab, he said.
Lessons from cancer detection
Companies developing blood-based screening tests for cancer are still building their case to regulators and payers, a process that could inform the mainstream adoption of COVID-19 screening.
Liquid biopsy technologies are opening the door for oncology to move past its “era of under-diagnosis,” said Seema Singh Bhan, SVP of public policy and external affairs at Thrive Earlier Detection Corp. On July 29, Thrive announced a $257 million series B round to support a registrational trial for its CancerSEEK test (see “Thrive to Make Case for Broad Access to CancerSEEK”).
Only a handful of cancers have standard-of-care screening tests, and these are organ-specific, such as mammograms; other tumors are typically only found once the cancer has progressed enough for symptoms to emerge. Blood-based tests that detect circulating tumor DNA (ctDNA), cancer-specific epigenetic patterns or tell-tale immune signatures offer the opportunity to intercept tumors early, from multiple tissues of origin (see “Wave of Enthusiasm in Disease Detection”).
A key part of incorporating these tests into medical practice is positioning them as a complement to standard-of-care tests and checkups, not a replacement; a corollary for COVID-19 testing would be making sure patients know to seek clinical diagnoses if they feel sick but received a negative screening result at work.
“We can’t underscore enough how tests like Thrive’s CancerSEEK are not intended to be looked at in a vacuum,” said Singh Bhan.
Eltoukhy said the risk/benefit calculation of a cancer screening test, and the weight placed on its performance specs, depends on the downstream consequences of a false result. For example, a screening test for pancreatic cancer, which risks triggering an invasive biopsy with a high risk of adverse events, may have more stringent specificity requirements than one for colorectal cancer, where a positive result would lead to an extra colonoscopy.
Similarly, institutions conducting COVID-19 screens must take the consequences of a false positive into account when selecting a testing technology, and incorporate a confirmatory test when those consequences would bring substantial hardship.
Adaptive's Baldo said it’s key to show regulators and payers the tests have utility in clinical practice through prospective and retrospective studies, coupled with real-world data. “We’re hearing a lot from the agency about the need for real-world data to support these intended uses, but also to support post-marketing efforts, and show how these tests actually fit into the continuum.”
Baldo said COVID-19 screening tests could build that evidence base by becoming embedded in larger ongoing studies, such as vaccine trials or studies evaluating infection rates, and thinks public-private partnerships could play an important role.
Singh Bahn said focusing a screening test’s use on populations at higher risk of contracting disease and incurring downstream healthcare costs can make a more straightforward case for its value proposition. She said Thrive is considering targeting its test toward populations over the age of 50, which have a much higher risk of developing cancer; for COVID-19, that could mean focusing on nursing homes or incarcerated populations.