June 2007: Volume 33, Number 6
A New Era in Cervical Cancer Screening
What Changes Will New HPV Tests and Vaccines Bring?
By Deborah Levenson
Cervical cancer was once the deadliest malignancy among American women. But since adoption of the Papanicolaou (Pap) test 50 years ago, rates of cervical cancer and related mortality have plummeted. In the mid-1990s, the introduction of liquid-based cytology techniques improved cytologists’ ability to find cancerous cells in a cervical sample, but even these newer techniques miss some cancers and their precursor, cervical intraepithelial neoplasia (CIN). Over the last two decades, the desire to improve screening has led to the development of more sensitive and complex molecular tools—including hybrid capture and polymerase chain reaction (PCR)-based tests—that identify the DNA of certain high-risk versions of human papilloma virus (HPV) subtypes that are now recognized as the primary causal factor in cervical cancer. With the overall HPV infection rate in U.S. women ages 14 through 59 at 26.8% (JAMA 2007; 297: 813–819), DNA tests are valuable tools that can identify some of these dangerous subtypes.
Today, the American College of Obstetrics and Gynecologists and joint guidelines from the American Cancer Society and the American Society for Colposcopy and Cervical Pathology (ASCCP) recommend a combination of cervical cytology and HPV screening for women age 30 and older and HPV tests for women with certain types of abnormal cytology. New ASCCP guidelines expected this summer will likely expand those recommendations. Just one company, Digene (Gaithersburg, Md.), has an FDA-approved HPV test, although Roche Molecular Diagnostics (Pleasanton, Calif.), Third Wave Technologies (Madison, Wisc.), Gen-Probe (San Diego, Calif.), and SensiGen (Ann Arbor, Mich.) all have forthcoming assays. These developments, as well as the introduction of a controversial new HPV vaccine (See Sidebar), could lead to a dramatic reduction in use—or even the demise—of the Pap test, according to some laboratorians and physicians experienced with HPV tests.
“I think the Pap’s days are numbered as a routine screening test. If I were running a cytology or histology lab, I would prepare for an era where the tests are based much more on immunological or nucleic acid-based staining, with more expert interpretation and image analysis. In other words, get ready for much more sophisticated results,” asserted Atila Lorincz, PhD, who recently left his position as Senior Vice President of Research & Development at Digene and is now Professor of Molecular Epidemiology at Wolfson Institute of Preventive Medicine (St. Barts, U.K.) and University of London (U.K). But widespread HPV testing faces several hurdles, including validation concerns and longstanding clinician experience and comfort with the Pap.
More Than Just a Shot in the Arm
HPV Vaccine Becomes a Political Issue
While more than 20 states are considering legislation aimed at immunizing preteen girls against HPV subtypes 16 and 18—and such bills are already scheduled to become law in Virginia and the District of Columbia—controversy surrounding a lobbying campaign by the manufacturer of the only FDA-approved vaccine has thwarted an immunization program in at least one state.
In April, the Texas legislature blocked an executive order by Governor Rick Perry (R) that would have required all 11 and 12 year-old girls entering sixth grade to get the vaccine beginning in 2008, with provisions for opting out. Bypassing the legislature and press reports linking the governor’s order to a Merck (Whitehouse Station, N.J.) lobbyist who is a former chief of staff for Perry raised the ire of lawmakers and many others in Texas. Meanwhile, Merck has ceased a national lobbying campaign that has prompted legislation with similar goals in other states, although the manufacturer is continuing a television advertising campaign, according to Merck spokeswoman Jennifer Allen.
The Merck campaign was premature not only from a political standpoint, but also from a public health perspective, maintained Neal Halsey, MD, Professor of Pediatrics at Johns Hopkins School of Public Health. “Before widespread use of any new vaccine, you usually want to wait a couple of years to gain wider safety data and to make sure that manufacturers can make enough. Inadequate supply is a common problem with other vaccines,” he explained. Right now the nation as a whole also lacks an effective infrastructure to deliver the vaccine to adolescents and young adults, who in general have poor immunization rates, he added.
Politics isn’t the only factor holding up widespread use of the vaccine. At a cost of $360 for a three-shot series, and reimbursement reportedly at about $50 or less, many pediatricians are refusing to stock it. In addition to poor reimbursement, pediatricians find managing inventory of the vaccine cumbersome and expensive.
With time and greater use, however, the vaccine is likely to gain widespread acceptance. “This vaccine appears to be very safe in the clinical trials and more experience with several million people in the next year or so will be valuable in proving safety,” said Halsey, who added that he would “welcome and support” HPV vaccination at that point. “Primary prevention by vaccine is far more effective than secondary prevention through testing,” he pointed out.
A More Direct Route to Detection
With the Pap test, cytologists view cervical cell samples through a microscope to look for CIN and atypical squamous cells of undetermined significance (ASCUS). Liquid-based versions that reduce cell clumping and improve sampling representation can make such abnormal cells easier to spot. In contrast, HPV tests detect viral DNA and can provide quicker, more objective answers about a woman’s health status than the Pap. “The Pap is based on a human looking down the microscope to find the occasional abnormal-looking cell. It’s like finding a needle in a haystack,” Lorincz commented. Digene’s FDA-approved Hybrid Capture 2 High-Risk HPV DNA Test and Roche’s Amplicor HPV Test, the next assay expected to emerge from the FDA pipeline, directly identify DNA of 13 cancer-causing subtypes of HPV. Both companies also plan to introduce assays that would identify which of these dangerous subtypes are present in a sample.
The Digene HPV Test, which uses the company’s proprietary hybrid capture 2 (hc2) technology, was approved in 2000 for ASCUS reflex testing and received expanded approval in 2003 for screening women ages 30 and older, in conjunction with the Pap. Also known to laboratories as the high-risk HPV DNA test or as the DNA with Pap Test, the assay works by signal amplification. After HPV DNA from a cervical sample hybridizes to RNA probes in the assay, a chemiluminescent system captures and detects the hybrids. Digene’s test can detect about 50,000 copies of HPV in a 1 mL sample, and labs can purchase additional reagents to differentiate between high- and low-risk HPV subtypes. Lorincz said that the test’s sensitivity is 95% and the specificity is 90% to 95% among women ages 30 and over in screening settings. Labs can manually process up to 88 samples on one microtiter plate. Using a robotic instrument, one technician can perform 354 tests in one batch in an 8-hour shift, according to Lorincz. Mostly larger hospitals and reference labs run the Digene test, but Lorincz maintains that properly trained technologists in medium-sized labs labs can also perform the tests. “But if you do only five or ten specimens a week, it’s not cost effective,” he pointed out.
Digene is preparing to provide more opportunity to use the HPV DNA test with liquid-based cytology. While its technology has been approved for use in conjunction with the Cytyc (Marlbororough, Mass.) ThinPrep System, in August 2006 Digene also sought approval to use the TriPath Imaging (Burlington, N.C.) SurePath specimen-collection medium. Approval would allow labs to perform the test directly on cervical cell specimen collected in SurePath vials.
While Digene has had the only FDA-approved HPV test in the market for several years, Roche Molecular Diagnostics submitted its much anticipated Amplicor HPV Test and Linear Array HPV Genotyping Test to the FDA in March. The Amplicor HPV test uses PCR and nucleic acid hybridization to detect the dangerous HPV subtypes. According to Walter H. Koch, PhD, Roche’s Vice President and Head of Research, the new test offers labs several advantages. It requires only a 250–500 microliter sample, compared to the 4 mL needed for the Digene test. The Roche test also includes an internal control which guards against false negatives that can result from insufficient cellular material for sample amplification. Koch also pointed out that PCR amplification carries no risk of cross-hybridization between HPV subtypes.
The version of Roche’s Amplicor HPV Test now under review at FDA uses manual sample preparation, but the test has also been designed for use with an automated sample preparation system now in development. Roche plans to develop future versions of both Amplicor and the forthcoming genotyping test that offer increased real-time PCR automation and automated sample preparation, Koch noted. Automation is a key element in reducing error and increasing productivity that other companies, including Gen-Probe, are also reportedly working on.
Mucosotropic HPV Types
Both the FDA-approved and forthcoming HPV DNA tests target 13 high-risk subtypes, while the HPV vaccines target only two of them. This chart shows these and other high-risk subtypes, as well as more common low-risk subtypes that do not cause disease and how they are related.
Used with permission of Mark Stoler, MD
The Question of Performance
If Roche’s new test receives FDA approval, for the first time, labs will have the option to choose between two commercial assays for HPV testing. European studies comparing the two assays have shown comparable performance. Italian researchers at Centro per lo Studio e la Prevenzione Oncologica in Florence examined 1,032 samples and found that the Amplicor test “provides a viable alternative to Digene’s assay,” with good agreement for samples with high-grade cytology. They also reported that the tests had similar sensitivity in detecting CIN lesions. Clinical sensitivity for the Roche and Digene tests were 96.7% and 97.8%, respectively, while clinical specificity was 54.9% and 51.6%, respectively. (Journal of Clinical Microbiology 2007; 45: 364–369).
French researchers at Alphabio Laboratory in Marseille tested 271 women with ASCUS for the 13 high-risk HPV subtypes and 234 controls with both the Roche and Digene tests. For ASCUS samples, the negative predictive value for the Digene and Roche tests were 92.8% and 87.8%, respectively. (Journal of Clinical Biology 2007; 45: 313–16). “The accuracy of these two assays is good and is compatible with routine clinical use in the triage of ASCUS cases,” the researchers wrote.
At least one study, however, raises questions about analytical specificity. Researchers at Institute of Microbiology and Immunology in Ljubljana, Slovenia prospectively tested 312 specimens for the 13 high-risk subtypes with both the Roche and Digene assays. Researchers received the same result in 85.9% of these samples, but noted that the Roche test had higher analytical specificity in their hands because of cross-reactivity with HPV subtypes other than the 13 targeted by the assay. “Prospective studies with clinical endpoints are urgently needed to assess the clinical utility of the higher analytical sensitivity of the Roche test,” they wrote.
Concerns about Future Tests
Although ASCCP would not publicly discuss specifics of its forthcoming HPV recommendations, the guidelines are expected to both expand clinical indications for HPV testing and note the lack of well-validated assays. A recent paper by researchers at the University of Virginia in Charlottesville and the National Cancer Institute (American Journal of Clinical Pathology 2007; 127: 335–337) highlights questions about analytic sensitivity and suggests parameters that HPV assays should meet before they are used routinely in clinical practice. According to the authors, the complex interplay between clinical and analytic sensitivity is one reason more HPV tests aren’t on the market. “HPV tests are not like those for HIV and hepatitis. You need to compare performance not just in terms of how many molecules you detect, but also by how well your test correlates with presence or absence of HPV-related disease,” explained one of the paper’s authors, Mark H. Stoler, MD, Professor of Pathology, Cytology and Gynecology at University of Virginia, and Associate Director of Surgical Pathology and Cytology at University of Virginia Health System. Stoler is a consultant for Merck (Whitehouse Station, N.J.) the company that developed the FDA-approved Gardisil vaccine, Cytyc, Digene, Ventana (Tucson, Ariz.), and Roche Molecular Systems. “It’s a testimony to how high the bar is that so few companies have tests on the market because it takes quite a commitment to do a proper clinical trial to get this right,” Stoler explained. Noting that Digene’s analytic cutoff is 10 to 50 times more sensitive than its clinical cutoff, he added, “It’s important to get the right point of clinically meaningful cutoff. A test that’s too sensitive will pick up lots of HPV of no clinical consequence. You want to titer the sensitivity of a test so it picks up everyone who has the disease but not the vast majority of people who are positive for HPV but don’t have the disease. That’s a complex situation.”
In their paper, Stoler and his colleagues propose a working definition for clinical and analytical validity of HPV tests. As a minimal standard before widespread implementation of any new test, they recommend that HPV tests have a clinical sensitivity of ≥92% for CIN3 in order to provide very high negative predictive value, and a clinical specificity of ≥85% in the screening setting. Stoler’s suggested parameters also call upon labs to document how well they can reproduce reported results to ensure reliable performance. All current HPV tests are limited by a lack of internal standards to evaluate specimen adequacy, the authors maintain.
Part of the impetus behind the recommendations was Stoler’s concern about companies that are marketing laboratory developed tests—also known as homebrew tests—for HPV. While devising a homebrew test with products available from several companies “isn’t that hard,” managing their complexity and quality control is more problematic. “An HPV homebrew test requires several steps and has many variables, such as how you collect the sample, process it, use cutoffs, and the populations you use to validate it. Labs also need to do their own ROC curve,” Stoler asserted. “There’s no reason to use a homebrew test if you have a valid, FDA-approved test, given the complexity we cite in the article.”
“The parameters set forth by the article are a reasonable place to start,” agreed Frederick Nolte, PhD, Director of Clinical Microbiology and Molecular Diagnostic Laboratories, Emory University Hospital (Atlanta, Ga.) and Professor of Pathology and Laboratory Medicine. Nolte participated in the FDA’s 2003 expanded approval of Digene’s tests, and is on Roche’s Scientific Advisory Board.
The problems—and solutions—the University of Virginia group identifies are especially important because the volume and extent of data submitted by Digene for the initial FDA approval are unlikely to be replicated by other companies, Nolte predicted. “Premarket studies similar to Digene’s would take decades. It’s unrealistic to expect companies to come up with that volume of data. So there will always be a little doubt about their performance characteristics,” he explained. FDA’s expanded approval of Digene’s assay for screening women over 30 was based on much less data than initial approval for reflex testing, he pointed out. Digene’s product information also notes that no clinical sampling was performed specifically to support the use of the test as an adjunct to the Pap. “Right now, there’s just one test that has data to demonstrate its value. What are we going to do when there are other tests to choose from?” he asked.
The HPV Vaccine: Another Prevention Tool
Meanwhile, a new HPV vaccine has the potential to drastically reduce the number of dangerous subtypes that tests aim to identify. In July 2006, the FDA approved Gardasil to protect women from HPV subtypes 16 and 18, which cause cancer, and two other subtypes involved in genital warts. Later that month, the Centers for Disease Control and Prevention’s (CDC) Advisory Committee on Immunization Practices recommended that girls ages 11 and 12 receive the vaccine, and added it to a program that provides immunizations at no cost to Medicaid recipients and certain other children. GlaxoSmithKline is also developing an HPV vaccine, which should spur competition.
But the HPV vaccine in no way negates the need for screening, because it’s only 70% effective, noted both Stoler and Elizabeth Unger, Team Leader at the HPV Laboratory in the Division of Chromic Viral Diseases at the CDC. “While the high-risk subtypes cause 70% of all cancers, 30% are left behind,” she explained, while Stoler predicted that if the vaccine is widely used over the next 10 years, Pap abnormalities will result from HPV subtypes other than 16 and 18. That situation will diminish the power of the current vaccines’ predictive value, and require more sensitive and specific HPV tests, Unger said.
Pap’s Staying Power
Even with more HPV tests on the market, conventional, direct-smear Pap and liquid-based formats from Cytyc and TriPath Imaging are unlikely to disappear any time soon, Unger maintained. “Pap tests in all formats have been shown to be more specific but less sensitive than HPV DNA tests. While the HPV test has excellent negative predictive value, it is less specific.” More importantly, physicians who are experienced and comfortable with the Pap will be slow to give it up, predicted Unger, adding that some clinicians are concerned about the “quirks” of Digene’s test, including cross-hybridization of the HPV subtypes. “There are still people out there looking for a Pap test with increased sensitivity. Screening programs have been organized around the Pap, and they have brought significant reductions in invasive cervical cancer wherever they have been introduced. The test is not perfect, but on a population basis it is very effective.”
Nolte has seen reluctance on the part of gynecologists to give up the Pap. Guidelines that recommend HPV tests for screening women age 30 and older haven’t changed screening practices at his institution, judging by what tests he sees ordered. “Not all guidelines are accepted by all clinicians,” he observed. “More of these HPV tests aren’t happening in my environment.”.