Feature Story

Opening the door to HPV testing

In search of better methods
The HPV test as a primary screen

October 2000
William Check, PhD

In the genre of mystery stories known as the police procedural, there are no angst-ridden, wisecracking heroes, no sudden revelations, and only an occasional dramatic resolution. Rather, the detectives use methodical investigation and repeated interrogation to ferret out their quarry.

Much the same can be said of chasing cervical cancer and its high-grade precursors with routine Pap cervical cytology screening. Like a real-world detective, the Pap smear is not glamorous or infallible-it lacks the sensitivity and specificity desired in an ideal laboratory test, and interpretation is not highly reproducible, even among expert cytopathologists. In 1988, more than 30 years after the procedure was introduced, experts were still seeking a rational classification of Pap cytology readings.

Yet, like dogged detectives, periodic examinations with Pap cervical cytology are ultimately effective. Persistence is the key, as Jerome Belinson, MD, immediate past chair of gynecology and obstetrics at the Cleveland Clinic, points out. "In the U.S. everything works better because you can do it multiple times," says Dr. Belinson, who directed a study of alternative screening methods for cervical cancer and its precursors in China. "That is the reason the Pap smear works as well as it does-it gets repeated."

Elizabeth Unger, MD, PhD, acting chief of the Human Papillomavirus Section at the Centers for Disease Control and Prevention, points out that the sensitivity and specificity of traditional cytology screening can be low. "The reason why we have decreased cervical cancer mortality is repeat testing," she agrees. In fact, Dr. Unger suggests, public health messages may have created falsely elevated expectations about preventing cervical cancer with the Pap smear. "It kind of boggles the mind that this whole process works at all," she says.

Preventing cervical cancer with repeat Pap screening depends on women visiting a physician’s office regularly, which requires vigilance and has attendant costs. To improve the quality of the Pap smear and increase its sensitivity, several approaches have been tried, with the goals of detecting a higher percentage of high-grade precursor lesions while reducing the number of unnecessary colposcopies and eliminating the need for repeat visits.

One method that has been under intense investigation in the last few years is testing cervical samples for human papillomavirus. HPV is associated with virtually all cervical cancers and high-grade precursor lesions [high-grade squamous intraepithelial lesions (HSIL), which comprise cervical intraepithelial neoplasia 2 and 3 (CIN2 and CIN3) and carcinoma in situ]. Most clinical studies of HPV testing have used Digene’s Hybrid Capture assay, the only FDA-approved kit for this purpose. Originally the kit included only the five types of HPV that present the highest risk for causing cervical cancer, Dr. Unger notes. As more types were discovered and the assay was found to produce negative results in women with high-grade cervical cytology, the number of types detected was increased to 13 in the Hybrid Capture II version. "Digene has worked really hard to make their test reproducible," Dr. Unger says. "Interlaboratory comparisons are very good."

In some studies, HPV testing has been combined with collection of cervical samples in a liquid medium, the ThinPrep collection system from Cytyc Corp., Boxboro, Mass. When a cervical specimen is collected in liquid, part is used to make a slide for cytology and the remainder is stored for HPV testing if indicated.

Cytopathologists’ assessments of the results of trials of HPV testing are positive. "I think that the data are very promising to encourage the use of HPV testing for triage in women with atypical cytology," says Diane Davey, MD, director of cytopathology at the University of Kentucky Medical Center, Lexington. "From what I have seen, it seems to be extremely sensitive and is an excellent triage for women with ASCUS. It looks like nearly all women who are later found to have biopsy-proven high-grade lesions or cancer, five percent to 15 percent of those with ASCUS, will have a positive HPV result."

Speaking of collection in a liquid-based medium, Dr. Davey says, "The biggest advantage of ThinPrep, and the fact that I would like to stress, is that you can avoid having the patient come back."

R. Marshall Austin, MD, PhD, medical director and director of cytopathology and gynecologic pathology services at Coastal Pathology Laboratories, Charleston, SC, also sees HPV triage as "a very valuable test." In studies in China and Costa Rica (see "The HPV test as a primary screen," page 68), combining ThinPrep with HC II had sensitivity for clinically significant high-grade and malignant lesions of nearly 100 percent," he says.

In practice, the Cleveland Clinic’s Dr. Belinson says, "We use HPV as a secondary screen for ASCUS Pap smears. If the HPV test is negative, we do not do a colposcopy." Drs. Davey and Austin also offer HPV testing for ASCUS triage; Dr. Davey sends it to a reference laboratory because volume does not yet justify bringing it in-house.

Growing acceptance of HPV testing in this setting rests primarily on cross-sectional data from two large U.S. trials. At Kaiser Permanente of Northern California, 973 women who had an ASCUS Pap smear had HPV testing, repeat Pap smear, and cervical histology. Prevalence of histologic HSIL or cancer was 6.7 percent. Sensitivity of HPV testing for HSIL was at least equivalent to that of repeat Pap testing. Among women with an ASCUS Pap smear, the authors estimated that immediate colposcopy of HPV-positive women and repeat Pap testing of all others would provide overall sensitivity of 96.9 percent (JAMA. 1999;281:1605-1610). At Kaiser Permanente, this algorithm is now recommended for ASCUS Paps.

Data from ALTS (ASCUS/LSIL Triage Study) were reported in March at the meeting of the American Society for Colposcopy and Cervical Pathology. Diane Solomon, MD, of the National Cancer Institute, project director of ALTS, explains that the study enrolled about 5,000 women, most with ASCUS but a minority with LSIL. Women were assigned to either immediate colposcopy, HPV testing followed by colposcopy if they were HPV-positive or repeat cytology if they were HPV-negative, or conservative management. Liquid-based collection was used.

It had already been reported that 83 percent of women with LSIL in ALTS were HPV-positive. Thus, HPV testing cannot be used for triage in this group because the test has no discriminatory power.

At the American Society for Colposcopy and Cervical Pathology meeting, Dr. Solomon reported on women who entered with ASCUS. In this group, HPV triage was very sensitive, detecting 96 percent of women who were ultimately found to have high-grade disease. However, HPV testing continues to have suboptimal specificity: Based on HPV positivity, slightly more than half of women with ASCUS would be referred to colposcopy, even though only about 11 percent had high-grade disease.

The sensitivity of repeat cytology using a cytologic diagnosis of ASCUS for referral to colposcopy approached that of HPV testing. "Repeat cytology with a low threshold for referral to colposcopy is certainly also a safe option for managing women with ASCUS provided women return for followup evaluation," Dr. Solomon concludes.

But now that HPV testing has been shown to be very sensitive in identifying women who have high-grade disease-a single HPV test was significantly more sensitive than a single repeat cytology-Dr. Solomon says it is "certainly an appealing option for clinicians and women to consider in the context of an ASCUS cytologic diagnosis. With HPV you get sensitivity with a single test," Dr. Solomon says, while sensitivity with cytology is based on multiple repeats. Patient retention in ALTS was high, but with "an effort well beyond what any office clinician would be able to do," Dr. Solomon notes. Used with liquid-based specimen collection, HPV triage does not require a followup visit.

Eventually every woman in ALTS will go to colposcopy, providing unique longitudinal data. Concluding patient visits are scheduled for January 2001.

But even with strong data emerging, some cytopathologists who favor HPV triage are having trouble getting clinicians to use it. "Clinicians in Kentucky are relatively conservative," Dr. Davey says. "Until it gets to be widespread, probably our doctors won’t order it a lot. I have started to recommend its use in many of my cytology reports," she continues. "But it is the minority of patients in whom it would be useful who are getting HPV testing."

Dr. Davey favors HPV testing since it would save many patients from returning unnecessarily and probably improve sensitivity for finding the women who are at risk for serious disease. She thinks many places will not adopt HPV triage, however, until final data from ALTS appear.

Dr. Davey highlights a group not usually talked about-women over age 40. Many of these women have cytology mimicking preneoplastic changes but actually due to perimenopausal hormonal changes. "HPV triage will be valuable in that group," Dr. Davey says, "for telling hormonal changes from neoplastic or preneoplastic changes. That is where I have been pushing it most, calling clinicians or putting a long note in the report." Colposcopy can cause a lot of anxiety, particularly in these women. "Personally, I would feel very comfortable if I had an ASCUS result and a negative HPV test being followed conservatively," Dr. Davey says.

Regarding reimbursement, Dr. Davey says, "A lot of payers are covering HPV testing. But if you send it to a reference laboratory, as we do, the charges are often higher than what third-party payers or Medicare or Medicaid would pay. So you take a loss or the patient pays some out of pocket." If HPV testing became common, she would bring it in-house and cost would decrease.

Dr. Austin is trying to educate clinicians about the usefulness of adjunctive HPV testing to triage women who have ASCUS Pap smears. "Many women have to come back for repeat visits, repeat smears, and colposcopy," he says. "If you can safely return them to annual screening based on a test from residual vial fluid, that seems to me a good opportunity." He finds that HPV testing has increased hand-in-hand with greater use of ThinPrep.

Physician reluctance is partly financial, Dr. Austin says. "One thing I have seen is that change is often driven by economic factors," he says. "It is really a challenge to put medical data first and to base your policies on that and then fight for reimbursement. There is a strong tendency for people to need to see that things are going to work economically first."

"We had to fight to increase reimbursement for liquid cytology," he adds. "We have been quite successful, but it has been a major effort. And we will have to fight for reimbursement for HPV testing, too."

Illustrating the impact of financial considerations, Dr. Austin says, "When I tell physicians that they can decrease unnecessary colposcopies with this method, some have frankly told me, ’Why would I want to do that? That is how I put my kids through college.’"

Some organizations have been proactive, Dr. Austin says. The American Society for Colposcopy and Cervical Pathology endorsed HPV testing and triage several years ago, but the American College of Obstetricians and Gynecologists has been more conservative.

Adoption of liquid-based specimen collection, too, is driven by dollars. "Reimbursement was not pushing our clinicians in the direction of ThinPrep," Dr. Austin says. In 1999, a "huge" study from Quest with biopsy confirmation provided convincing data on the value of this technology. "Around that time, data on effectiveness of liquid cytology in detecting endocervical glandular lesions had also finally appeared, and we then started heavily converting," Dr. Austin says. "We went in one year from one percent or two percent to one-third. By the end of this year, we will probably be over 70 percent."

But widespread adoption of liquid-based collection depended on obtaining favorable reimbursement in addition to strong data. "Once we became convinced, we went to insurance companies," he says. "We were very aggressive."

In the United States overall, use of HPV testing is increasing. Attila Lorincz, PhD, chief scientific officer at Digene in Gaithersburg, Md., predicts that HPV test sales will double over the next year. "We estimate that approximately 15 percent of ASCUS Pap smears are triaged now with the Hybrid Capture HPV test," Dr. Lorincz says, "and we think that will grow to over 25 percent by next year." He acknowledges that use of HPV testing depends to some extent on adopting liquid-based cytology, which, he estimates, is now used in 40 percent to 50 percent of Pap smears.

Some health centers, however, use "co-collection," taking a second conventional cervical sample and storing it. If cytology is read as ASCUS, this sample is retrieved and sent for HPV testing. In the Maryland Department of Health, three clinics use co-collection and two use ThinPrep. "Some people say co-collection is better because you don’t have to pay for liquid cytology but the patient still doesn’t have to return," Dr. Lorincz says.

Digene’s marketing plan for Hybrid Capture II is "to push to get the entire market on ASCUS in the next three years," Dr. Lorincz says. "That will give us the momentum to move forward on primary HPV screening." Digene has made a submission to the FDA for primary screening based on five studies done in Europe, China, and Latin America. The FDA is deciding to what extent it will be satisfied with those data. "If FDA wants U.S. data, we could use computer modeling based on Caucasian populations similar to the U.S. where HPV as a primary screening tool worked extremely well," Dr. Lorincz says.

Whether Digene will apply for true primary screening is not clear. "Most likely we will be calling for HPV testing in women over 30 years of age in conjunction with a Pap smear every one to three years," Dr. Lorincz says. "We don’t want to replace the Pap smear at this point."

Pathologists are dubious about HPV testing as a primary screen in the United States. Accuracy is "very population dependent," Dr. Unger points out. Disease prevalence strongly affects specificity, which is where HPV tests are weakest. And primary screening trials have mostly been conducted in countries where cervical cancer is much more prevalent. "It is very dangerous to take numbers from a country where HPV is more prevalent and try to say how it would work in the U.S.," Dr. Unger cautions. She favors doing a trial in the United States.

Dr. Davey, too, emphasizes that "The rate of significant disease is so much lower here. So the percentage of positive results that would turn out to be false positive would be so much higher. We would have a major issue trying to figure out how to manage or follow women who have positive results on HPV testing and nothing wrong on cytology and exam." She says, "We would first need to do studies with HPV as a primary screen here in the U.S."

Specificity problems would be particularly acute in younger women, who have a high rate of HPV-positivity but usually nothing abnormal on examination or cytology and most of whom will regress. Yet some mode of screening is still necessary at these younger ages, Dr. Davey says. She sees women age 18 or so with cancer picked up on standard cytology. "We have to continue to push for annual examination and cytology in that younger group," she emphasizes. "We can use HPV testing as triage in young women with an atypical result, but I wouldn’t use it as a primary screen."

Dr. Davey believes HPV screening "will be most useful in women possibly over age 35 or clearly 40 and over or in populations where you fear that women will not be screened again for a long time. In those women, you can maximize detection by using HPV and cytology at the same time."

Dr. Austin agrees. "I would think one of the first applications [of HPV primary screening] would be in older women, where it might be possible to demonstrate that you could increase screening intervals in women with negative test results. But," he cautions, "those studies are not easy to do and would take a long time."

Mark Schiffman, MD, MPH, of the Division of Cancer Epidemiology and Genetics at the National Cancer Institute, who coordinated a cervical screening trial in Costa Rica, agrees the HPV test performs much better after women are about age 30. "It is entirely dependent on the society," he says. "In a society in which women didn’t start having sex until age 30, that would be a horrible age to screen." In the United States the peak of CIN2 and 3 is somewhere around age 30, so starting to screen at age 30 would miss a lot of women who have significant high-grade lesions. "Starting to screen [in general] at age 25 makes a lot of sense," Dr. Schiffman says.

Published data have not shown that HPV testing by itself has been 100 percent sensitive, but it does work well with cytology, Dr. Austin says. "That is not really primary screening," he says. "It is still what I would call adjunctive testing." Doing both Pap and HPV screens in all women makes the cost issue especially acute.

Cost-effectiveness is an issue in HPV triage of ASCUS cytology, too. "Ideally we should ask how does this work for cancers," Dr. Unger says. "But cervical cytology in the U.S. has been so successful that there are hardly any cancers anymore." That raises the question of whether a new test will be cost-effective.

A cost-effectiveness analysis is built into ALTS, Dr. Solomon says. And the Kaiser Permanente investigators wrote that their unpublished analysis shows that "savings accrued from decreased visits and procedures were . . . sufficient to offset the cost of implementing ThinPrep Pap testing for all routine screening and subsequent ’reflex’ HPV testing for all ASCUS cases."

HPV triage raises psychological and social health issues as well. "What are the implications of screening every woman in the population for HPV?" Dr. Unger asks. Many will be found positive and will ask, Where did I get this? Can a person be positive with a negative partner in a monogamous relationship? "Are we going to turn cervical cytology into an STD?" Dr. Unger wonders. "Many women don’t want to be tested for STDs," she says. "They say, ’I’m not the kind of woman who would get an STD.’ If they think of HPV as an STD, will they stop going for cervical cytology?

"Before we start putting out HPV test results," Dr. Unger cautions, "clinicians, laboratorians, and patients need to be educated." She finds that some people are desperate to talk to her about HPV infections because of inaccurate information they have received from clinicians. "Some physicians tell patients, ’You are HPV-positive. You are promiscuous.’ If a couple has been married for 10 years and one has positive HPV-like cytology," Dr. Unger finds, "some physicians will say, ’You have an HPV infection and one of you must have been unfaithful.’"

"In any situation in which a woman assumes she is in a monogamous relationship, a positive HPV test-or LSIL or CIN1-makes her question her husband’s behavior," says Dr. Schiffman. Positive findings may imply an STD, but issues of latency and re-emergence complicate the interpretation. "A positive test result does not necessarily prove something about her partner," Dr. Schiffman says. "Handling that revolution in thought is the transition that we need to make."

"As we start to move toward a greater understanding throughout the world that HPV is a sexually transmitted infection, we need to be convincing both men and women that testing is not a threat," he adds. He already see hints of an I- don’t-want-to-know reaction.

Anticipating wider use of HPV triage, Dr. Davey offers practical advice. "Pathologists doing cytology and microbiologists doing HPV testing will be working together in this situation," she notes. "If a liquid-based sample is collected, it will first come into cytology but then be sent for HPV testing to another laboratory or to a reference laboratory. So pathologists need to get this organized, because with ThinPrep you have 21 days for FDA-approved HPV testing from the time of specimen collection." Traditionally, turnaround time for Pap cytology can be a couple of days to a couple of weeks, since it is a screening test. Laboratories may need to change their procedures or priorities and communicate with physicians about how to handle the vial.

Dr. Austin offers a final reminder. "I think the key is to keep the patient as the main focus," he says. "That can be hard to do when so many financially powerful interests, including third-party payers and proprietary manufacturing companies, have entered the conversation."

William Check is a freelance medical writer in Wilmette, IL.