NEW YORK (GenomeWeb) – A recent evaluation of FDA-approved molecular HPV screening assays from Roche, Hologic, and Qiagen has found that the three tests showed variations in which samples tested positive or negative. These disparate results point to an obvious question: would it be worse to receive a falsely positive result or a negative result when an individual is actually infected with a high-risk viral strain?
It appears the answer to that question will remain a matter of debate among the makers of those tests and researchers.
In a study published this month in The Journal of Clinical Microbiology, a group at the Mayo Clinic compared three FDA-approved PCR-based assays for HPV in samples already showing cytological hallmarks of cervical disease. The study also assayed whether testing for two high-risk HPV types thought to account for about 70 percent of all cervical cancers, genotypes 16 and 18, could help in triaging patients.
A similar head-to-head study published in PLoS One in January also compared these three tests, plus one from Genomica. These data provided "an unexpected challenge for the choice of an assay in primary cervical screening," as substantially different subsets of women tested positive in the various assays, a result best expressed in a blossoming four-way Venn diagram.
Interpretations of these study results understandably differ.
The lead author on the JCM study, Matthew Binnicker, an associate professor of laboratory medicine and pathology at Mayo Clinic, told PCR Insider in an email that the three tests were equally sensitive in cases where a women had cytological evidence of cervical cancer.
However, the apparent higher sensitivity of the Qiagen digene HC2 test versus the Roche cobas and Hologic Aptima assays in other cases is actually a tendency to "overcall" viral infections. "The HC2 assay called 40 more samples positive for high-risk HPV compared to the Aptima HPV test, and these 40 patients did not show evidence of cervical cancer by tissue biopsy," he noted.
Qiagen, on the other hand, maintains that the study confirms HC2's superior sensitivity and should ensure the test's continued use as a "gold standard" HPV screening.
Tadd Lazarus, chief medical officer at Qiagen, explained in an interview that the HC2 test measures the entire HPV genome, while the Roche test only measures the L1 portion and will give a false negative result in cases of L1 dropout.
The Hologic test, meanwhile, measures mRNA "elaborated when there is a long-lasting progressive infection," Lazarus said. "When you're using a test like that for screening, it's not positive for early infection, even if the high-risk HPV types are present; it wasn't designed ever to be used for a primary screen, it was designed to look for progressed HPV infection that had become more stable," he said.
According to Binnicker, the two Venn diagrams in his study illustrating discordance in positive and negative samples between the tests are most revealing.
"The take home message ... is that the Qiagen HC2 assay calls more samples 'positive' for high-risk HPV than the other tests," he said.
"This could be interpreted as the Qiagen test being more sensitive; however, it can also be interpreted as the Qiagen test being less specific for cervical … which is what we demonstrated when we compared the results of the three high-risk HPV tests to the results of cervical biopsy."
Meanwhile, Binnicker found that the Hologic Aptima HPV test calls more samples 'negative' for high-risk HPV than the other two tests, and "this most likely means that the Hologic test is more specific," meaning it doesn’t overcall the presence of high-risk HPV, he said.
The Qiagen test also does not report genotype data, the study noted. "Having the ability to not only determine whether high-risk HPV is present, but also that HPV-16 and/or HPV-18 is present, can help triage care for women that are being screened," Binnicker said. "If the HPV-16 or HPV-18 genotyping results are negative, the provider can choose to see the patient again in 12 months for repeat screening. However, if HPV-16 and/or HPV-18 are detected, the patient should be referred directly to [cervical biopsy] to determine if cancer is present," he said. Thus, the Qiagen test was only included in the study because it was the routine test used in the lab at the time, Binnicker said.
However, Qiagen's Lazarus begs to differ about the importance of genotyping HPV.
"Does that mean that we should not also be focusing on the 30 percent of women who have the 12 other high-risk types?," he asked. Indeed, one recent study suggested the 16 and 18 genotypes may lead to cancers with a better prognosis than other strains.
While Binnicker asserted that the detection of HPV-16 and/or HPV-18 showed the highest specificity when compared to the results of tissue biopsy, Lazarus suggested that since the study was performed only with patients having abnormal Pap test there was in fact no clinical relevance for genotyping at all. Current guidelines recommend genotyping only in HPV-positive, but cytology-negative, cases.
And a false negative result is a terrible thing, in Lazarus' opinion. "If you're co-testing, and a woman comes back and she is HPV negative and [negative for intraepithelial lesion and malignancy by cytology], then she can go as far out as three to five years [before follow-up]," Lazarus said.
"Those are the women we worry about the most. If that's a false negative, and they do indeed have high-risk HPV, they're being sent out thinking that the screening interval is being extended, and extending it without proper security for the woman; three to five years is a nice long period of time if it's a false negative," he said.
Screening tests are best when they're highly sensitive, Lazarus asserted. "What this Mayo Clinic paper underscores is the reason why our test remains the referential gold standard .. [it] reveals that our sensitivity for [moderately abnormal] cytology or greater is 97.5 percent, whereas the Aptima test from Hologic and the cobas test from Roche have the same sensitivity [of] 91.4 percent. That's a large statistical difference, with non-overlapping confidence intervals. That's a terribly important difference," he said.
The PLoS One study also seems to argue as Qiagen does — the majority of women who test positive for HPV will show no cytological abnormality, so limiting a study to women who already have lesions or misshapen cells constitutes a "selected population."
As long as co-testing is the standard of care, the Pap test provides the necessary specificity, Qiagen argues. "The purpose of adding an HPV test is to enhance the overall sensitivity. ... As the study was performed in ... patients with abnormal Pap test results – the needed specificity was given anyway. Therefore, the sensitivity of the HPV test made the difference, and even under this test pool environment, where basically sensitivity will always be higher for an HPV test because of the removal of Pap negatives, HC2 still had the best sensitivity ... compared to other FDA approved tests," the company noted in a follow-up email.
How are these debates about the relative importance of false positives and negatives ultimately resolved? Both outcomes are clearly bad, and balancing them "is very complicated and requires an understanding of risk aversion, societal preferences, etc," Matejka Rebolj, author of the PLoS One study and an associate professor in the centre for epidemiology and screening at the University of Copenhagen, told PCR Insider in an email.
Binnicker emphasized that a screening test should have high sensitivity, but that should be balanced with an appropriate level of specificity to clinical disease. "The HC2 assay showed the highest sensitivity in patients with a biopsy result of [moderately abnormal cells]; however, that was countered by the poorest specificity," he said. "Ideally, a healthcare provider is seeking a screening test that can detect the majority (>90 percent) of pre-cancerous lesions, while offering a high enough specificity so that women are not unnecessarily referred to further testing or invasive procedures," he said.
Meanwhile, Qiagen asserted, "Only HC2 has the long-term safety data to demonstrate that a single positive HC2 result increases cumulative incidence of [moderately abnormal cells] over a 10-year period to nearly seven percent. Furthermore, a single negative HPV result lowers that incidence to less than one percent. No other high-risk HPV assay has data to show that a single negative result lowers the 10-year risk of [moderately abnormal cells] to these acceptable standards," the firm said in its email to PCR Insider. Until these data are available, a negative result has to be interpreted with caution."
Competition in this important molecular diagnostic market will likely continue to increase, even as clinicians debate whether molecular diagnostics should replace cytological testing. While Qiagen has recently seen a slowing in it's HPV business, new HPV tests are in the works from other companies, such as Advanced Cell Diagnostics, Becton Dickinson, and Cepheid, and Abbott has a test currently available outside the US.
None of the companies whose tests were evaluated in the JCM study provided funding, equipment, or reagents, Binnicker noted. However, Hologic and Roche did provide technical support to the Mayo Clinic lab, "as we did not have experience with their methods. The Qiagen HC2 assay was our routine test at the time, so no technical support was required," he said.