July 2012
Feature Story

William Check, PhD

What if you built a better mousetrap and people ignored it? What if it was so much better that it captured two-thirds more murine pests, but people still were unwilling to change their approach to mouse catching, even though, over the long run, the new mousetrap could save money? If it were just mice at stake, you might shake your head and move on to another invention. But if you were talking about a better method for preventing cancer and saving human lives, you would work even harder to get clinicians to adopt the improved test.

That’s the prevailing situation with cervical cancer screening in the United States. Based on research showing that 90 percent or more of cervical cancer cases are due to persistent infection with high-risk genotypes of human papillomavirus (hrHPV), clinical tests have been devised to detect this virus in cervical samples with high sensitivity. Four commercial hrHPV tests have been approved in the United States. Many public health officials, cytopathologists, and oncologists have urged clinicians to incorporate hrHPV tests into cervical cancer screening. And yet the use of these assays appears to be growing slowly.

Perhaps this situation will change with new guidelines issued earlier this year by a committee working under the auspices of the American Cancer Society, American Society for Clinical Pathology, and American Society for Colposcopy and Cervical Pathology (Saslow D, et al. Am J Clin Pathol. 2012;137:516–542). Committee member and cytopathologist Mark H. Stoler, MD, professor of pathology and clinical gynecology at the University of Virginia, told CAP TODAY that this continued and more emphatic effort to increase the use of hrHPV testing in cervical cancer screening is due to a number of advances: several large randomized clinical trials comparing hrHPV testing with cytology that were published since the last guidelines in 2002, the availability of the four FDA-cleared hrHPV assays, and the introduction of HPV vaccines in 2006. But one can’t help but believe it also represents the determination of those who see the advantages of hrHPV testing to get physicians to recognize and adopt the better mousetrap.

“Development of the guidelines presented an opportunity to evaluate how different screening tools work in populations with low cervical cancer prevalence,” committee member and cytopathologist Ann T. Moriarty, MD, technical supervisor at AmeriPath Indiana, said in an interview. (Dr. Moriarty noted that her comments are her personal views, and do not necessarily represent the views of her employer.) “We’re not going to do the Pap test differently,” Dr. Mori-arty continued, “but the women of the U.S. will be screen-ed differently. Intervals of screening and how you screen will change.”

These changes will require adaptation. “With the combination of the vaccine and increased use of type-specific HPV testing, patients and physicians will have to be more comfortable with longer screening intervals,” Elizabeth R. Unger, MD, PhD, told CAP TODAY. “You can very easily overscreen, particularly when you use HPV screening with the Pap smear,” cautions Dr. Unger, chief of the chronic viral diseases branch, Centers for Disease Control and Prevention.

Paradoxically, improved screening technology is needed because the traditional method is so good. “We’ve done a wonderful job with the Pap test since the 1950s in reducing incidence and mortality from cervical cancer,” Dr. Moriarty says. “But if we continue to use an insensitive test, which the Pap is, we will not be as successful in detecting the small precursor lesions or the rare cancers that will occur.”

“We are victims of our own success,” Dr. Stoler told CAP TODAY. “It’s all about disease prevalence and test sensitivity. Before screening, the cancers and precancers were bigger and easier to find and 10 times more common. But as we have screened out much of the relatively easy cases, what’s left are many cases that yield more equivocal Paps and that may also be harder to see on colposcopy.” Screening has lowered the prevalence of disease so far that the Pap test’s limitations are being reached. “We need to adopt more sensitive methods like HPV assays,” Dr. Stoler says.

Major recommendations of the new guidelines include: no annual screening at any age; no screening before age 21; cytology alone at three-year intervals for women ages 21–29; and no screening with any modality for women older than 65 who have “evidence of adequate negative prior screening and no history of CIN2 within the last 20 years.”

Where the guidelines make a historic break with past U.S. recommendations is in women ages 30 to 65. For the first time, co-testing with cytology and hrHPV is the “preferred” approach, with cytology alone being rated “acceptable.” Also, the interval for co-testing is longer, five years, compared with three years for cytology alone.

In an online talk in May, sponsored by the Association for Molecular Pathology, Dr. Stoler presented the evidence from large European randomized trials that compelled these changes. Data from three of these trials, which the committee rated of “high” quality, demonstrated that co-testing achieved a 20 percent to 30 percent increase in detection of CIN3 in the first round of screening compared with cytology alone (Naucler P, et al. N Engl J Med. 2007;357:1589–1597; Ronco G, et al. Lancet Oncol. 2010;11:249–257; for a summary, see: Arbyn M, et al. Lancet Oncol. 2009; 10:935–936). For CIN2/3, co-testing was 64 percent more sensitive.

A second major advantage: Co-testing reduced the rate of invasive cancer in the second round of screening. In the trial by Ronco, et al., the number of invasive cancers in the second round in the cytology-only group was seven, while none were found in the co-testing group. (Each cohort had about 34,000 subjects.) Based on the overall data, the committee concluded: “A negative co-test has a high negative predictive value for CIN3+ and cancer in the subsequent 5 to 6 years.”

Evidence that is less stringent suggested a decrease in the lifetime incidence of cancer with co-testing and, in a modeling study, a 17 percent decrease in lifetime risk of colposcopy.

For women who are hrHPV-positive but cytology-negative, who are estimated to make up four percent to five percent of a screening cohort, direct referral to colposcopy is discouraged. Instead, the guideline offers two triage options: Either repeat co-testing at 12 months or HPV genotyping for types 16 and 18. (This is the only situation in which HPV-genotype-specific testing is recommended.) Colposcopy is recommended in the first option if either test is positive at 12 months and in the second option if the specimen is positive for either genotype. In the U.S. ATHENA trial, women who were cytology-negative but positive for HPV types 16 or 18 had almost a 10 percent risk of having CIN3+, compared with 0.3 percent for those who were negative on both cytology and hrHPV testing (Wright TC Jr, et al. Am J Clin Pathol. 2011;136:578–586).

Women with ASC-US cytology and a negative HPV test should continue with routine screening, the guideline recommends. Their five-year risk for CIN3+ is not significantly different than for women who are double negative.

In arriving at these recommendations, the committee balanced benefit versus harm, with colposcopies standing in as a surrogate for potential harm. “Colposcopy is the most accessible measure we have in the literature,” Dr. Stoler explained in an interview, “and it sums up potential harms—time away from work, biopsy, risk of being referred for surgery, surgery itself, and pathology and other downstream impacts.” Co-testing will lead to more colposcopies in the initial screening round because the specificity of hrHPV testing is five percent to 10 percent lower than that of cytology. However, with a five-year rescreening interval, “The increased interval dominates the equation,” Dr. Stoler says. With the management strategies proposed in the guideline, “You get close to an ideal referral rate,” he notes, about two percent.

When Dr. Stoler completed his online talk, an attendee asked about laboratory-developed tests for hrHPV. “I don’t see any utility to laboratory-developed tests for HPV,” he answered. “I know that may sound harsh, but [clinical validation] is beyond the capacity of the average laboratory.” Expanding on his remarks in an interview, Dr. Stoler noted the four FDA-approved tests. “Trial data supporting those tests required tens of millions of dollars,” he says. “Companies would not spend that kind of money if it were not required to support the validity of this application of molecular diagnostics. Laboratories can’t do that kind of validation where the endpoint is disease, not molecules of HPV. This is different from trying to measure a biomarker in a patient with cancer.”

Dr. Unger agrees that this is an important issue. She uses HPV testing for epidemiologic research. Still, she says, “The same basic principles apply for clinical testing. For clinical testing I think it’s even more important to standardize the whole package—specimen collection, processing, and testing. Change any one of those and you change the result.” The package insert for an FDA-cleared test defines a standardized protocol.

Wider adoption of the guidelines will negatively affect cytopathology laboratories, Dr. Stoler concedes. “For co-testing to make any sense, you have to go to five-year intervals,” he told CAP TODAY. “If everyone followed the guidelines and all women 30 to 65 got tested only once every five years, that’s already a lot fewer cytology exams.” Dr. Stoler cites the projection of Diane Solomon, MD, of the National Cancer Institute, who attempted to quantitate the effect: 20 to 30 million Pap tests annually. “We have to adapt,” Dr. Stoler said. “And we will adapt.”

Just as European countries have led in evaluating and implementing the addition of hrHPV testing to cytology, so too are they ahead in the next logical step—using hrHPV testing as the primary cervical cancer screening modality. Denmark and the Netherlands are making this change now, and other countries will no doubt follow.

In the new U.S. guideline, screening with hrHPV alone is not recommended. “Even though most data suggest it should work,” Dr. Stoler said in his online talk, “there is no FDA-approved test for this indication and no [published] U.S. primary screening trial.” In an interview, however, Dr. Stoler added further thoughts. “[In the talk] I was giving what was in the guidelines, not editorializing,” he clarifies. “Do you think the six RCTs done in Canada and Europe are adequate data? If you do, that is the argument for HPV primary screening, especially if you bring in the cost-efficiency argument.” Dr. Stoler also noted that pending for the trials supporting the two most recently approved hrHPV tests—Roche cobas 4800 and Gen-Probe Aptima—are three-year followup data. “Those data,” he speculates, “might lead to an FDA submission” for primary screening.

A U.S. population-based study of more than 330,000 women enrolled in co-testing at Kaiser Permanente Northern California provided indirect evidence for the value of hrHPV primary testing. Among these women, the five-year risk with a negative HPV test, 0.17 percent, was almost identical to that with a double negative test, 0.16 percent, raising a question about how much protection the cytology component is providing (Katki HA, et al. Lancet Oncol. 2011;12:663–667). The authors concluded: “Testing for HPV without adjunctive cytology might be sufficiently sensitive for primary screening for cervical cancer.”

“I was totally against primary HPV screening initially,” Dr. Unger says. “But evidence is increasing that HPV is very good for primary screening because it has a very low false-negative rate. If the HPV test is negative, you can stop.” Dr. Unger adds that HPV assays are much easier to standardize than cytology, “so they would be particularly good in developing countries.”

However, she notes, HPV primary screening does not mean HPV testing only. It means HPV followed by some appropriate triage. “Just as in co-testing, you don’t automatically take all positives to colposcopy.” The reason is that hrHPV testing has lower specificity than cytology, so too many women would be referred. “People are proposing cytology as a triage method, and it does work reasonably well,” Dr. Unger says. Cytology is being proposed in the European countries that are adopting hrHPV for primary screening. A report from Sweden concluded: “Primary HPV DNA-based screening with cytology triage and repeat HPV DNA testing of cytology-negative women appears to be the most feasible cervical screening strategy” (Naucler P, et al. J Natl Cancer Inst. 2009;101:88–99). However, Dr. Unger cautions, “We are not there yet. We still need further investigation of other methods as a second test.”

One proposed approach is to test for persistence of hrHPV genotypes. “Most active HPV infections regress spontaneously,” Dr. Moriarty says. “Cervical cancer can’t occur without persistent infection and integration and transformation of cells.” Tests that measure E6/E7 HPV gene expression, which is essential to development of cancer, are also being evaluated. A small study in Norway found that such an assay was more specific than cytology in a post-colposcopy setting (Arbyn M, et al. Int J Cancer. 2012 May 24, Epub ahead of print).

Dr. Moriarty was chair of the working group that recommended not adopting hrHPV for primary screening right now. “In the U.S. we have an opportunistic screening program,” she explains. “That is very different from programs outside the U.S., especially Europe, where people are invited to screening, then followed up. In the U.S. it is ‘catch-as-catch-can’ and we don’t have centralized registries for followup.” Dr. Moriarty calls hrHPV primary screening “promising,” and adds: “It will probably be in place eventually in some places in the U.S.” However, she also emphasizes the need for research into methods to increase specificity.

HPV vaccination is no reason to change recommended screening practice, according to the guideline. “With HPV vaccination, there should be a great decrease in cervical cancer,” Dr. Stoler says. However, in the U.S. coverage is still low: Vaccination rates are increasing, but as recently as 2010 only one-third of females ages 13–17 got three doses.

“There is a major difference between developed countries where we embarrassingly lag in the adoption of vaccination,” Dr. Stoler says. “Theoretically, if we had high coverage, all women who enter the screening system between ages 21 and 29 will never get disease related to HPV types 16/18, which account for 70 percent of women who come to LEEP [loop electrosurgical excision procedure]. Vaccine protection lasts at least 10 years with no sign of breakthrough. But in the U.S. we can’t change [the screening algorithm for vaccinated women]. We don’t even have a good registry of who has been vaccinated.”

Getting clinicians to adopt the new guidelines is only the first step in the battle to reduce the incidence of cervical cancer even further. Higher vaccination rates and possible hrHPV primary screening are logical next steps. However, all of these measures benefit women who are already in the health care system. What about the large proportion of cervical cancer cases in women who have had no recent screen?

“From surveys we know that most women are getting screened,” Dr. Unger says. “We have a continuing effort to try to reach unscreened women through CDC’s Breast and Cervical Screening Program. That’s one reason why the vaccine has such appeal: It could reach unscreened women and could be getting down to the small numbers of cancers that are not always screen-detected, such as some endocervical cancers.”

Even these efforts reach only women in developed countries. Many other campaigns are trying to bring screening to developing countries. Dr. Stoler ended his talk with a slide showing a global map of cervical cancer incidence. Many countries in Africa and South America showed very high rates, as did Russia. Overlaid on the map was the admonition: “We need to do better.”