by Shelia Hoar Zahm, Sc.D.
Human papillomavirus (HPV) is understood to cause virtually all cases of cervical cancer. In recent years, scientists and clinicians have used this knowledge to develop two successful and complementary preventive methods: (1) vaccination against HPV infection and (2) HPV testing. HPV testing detects treatable cervical cancer precursors with greater sensitivity than can be achieved by the Pap test (cytology) alone.
DCEG staff members have contributed substantially to a series of guideline meetings focused on introducing HPV testing into cervical screening programs in the United States. The first set of meetings addressed general screening, or the testing of presumed normal women. These meetings led to the publication of new screening guidelines in 2012 (described in the July 2012 issue of Linkage). Multiple organizations, including the American Cancer Society, the American Congress of Obstetricians and Gynecologists, and the U.S. Preventive Services Task Force, endorsed the use of concurrent HPV testing with Pap testing (referred to as co-testing) at five-year intervals for women who test negative on both tests. Compared with current practice, the new guidelines recommend longer intervals between routine screenings. Mark Schiffman, M.D., M.P.H., a senior investigator in the Clinical Genetics Branch (CGB), noted, “The high sensitivity and negative predictive value of HPV DNA testing make the test amenable to fewer screens in a lifetime than Pap testing.”
With new screening guidelines in place, the focus turned to reaching agreement on the clinical management of women who test positive when screened for HPV. To support this goal, DCEG scientists, in collaboration with Kaiser Permanente Northern California (KPNC), used data from 1.4 million KPNC members to provide key evidence for the formulation of new risk-based guidelines for the clinical management of women with abnormal cervical cancer screening tests. The resulting 2012 American Society for Colposcopy and Cervical Pathology Consensus Guidelines and the eight reports with the supporting data from the KPNC study were published in a 2013 supplement of the Journal of Lower Genital Tract Disease.
The screening and clinical management strategies included in the guidelines are based on essential features of the natural history of cervical cancer: infection with HPV; persistence, rather than clearance, of the infection; development of a precancerous precursor lesion (effectively, carcinoma in situ); and invasion. Because all cervical cancer cases go through these steps, scientists can develop algorithms based on precancer and cancer rates among women with similar test results who have been followed over time. These calculations can then be used for managing women with various results from HPV testing, Pap testing, and colposcopy (a diagnostic exam during which biopsies are taken).
Risk prediction models developed by DCEG were pivotal to the development of the new screening guidelines that recommend HPV-Pap co-testing as the primary screening method, with longer intervals between routine screens. This approach was applied to the management guidelines. Risk calculations reduce a complex battery of tests over time into a single number that forms the basis for action. Hormuzd A. Katki, Ph.D., a tenure-track investigator in the Biostatistics Branch and lead author on the reports supporting the new management guidelines, explained, “If two different combinations of test results imply the same risk of disease, then we manage them similarly. This is the principle of ‘equal management of equal risks.’”
Risk calculations for HPV-Pap co-testing were compared, or benchmarked, with implicit risk thresholds underlying previously accepted Pap-only screening. For example, at KPNC, women with low-grade Pap results, for which immediate colposcopy is currently prescribed, have a five-year risk of 5.2 percent of developing precancer or cancer. Therefore, under the new guidelines, women with any combination of HPV and Pap results that carries a five-year risk of 5.2 percent or more of developing precancer or cancer should be referred for colposcopy. Women with a five-year risk of around 2.6 percent should be advised to return for testing in one year, because that has been the accepted practice for women with an equivocal Pap result. Women with a five-year risk below 0.26 percent need not return for screening for three years because they have risk equal to a woman with a negative Pap test.
Through long-term follow-up of the women in the KPNC study, which began HPV-Pap co-testing in 2003, DCEG investigators and their KPNC collaborators determined the risks of disease for various combinations of HPV and Pap test results. Dr. Katki and his colleagues conducted separate analyses for each combination of HPV and Pap results to calculate risk and then compared their analyses with the benchmark Pap-only thresholds to develop clinical management recommendations. The analyses considered sequential HPV-Pap test results, biopsy results, and excisional treatments for precancers. Because of the large number of women in the study, the excellent follow-up rate, and the detailed medical information, the analyses were powerful enough to detect clinically meaningful distinctions even for the lowest levels of risk.
The addition of HPV testing to cytology is creating novel screening results. One of the most challenging subgroups is comprised of women with positive HPV tests but negative Pap tests (see Figure 1). Although the longer-term risk of precancer and cancer is elevated among women in this subgroup, cytology is still normal. The positive HPV test could reflect a transient infection that will clear on its own or a very early precancer that might not yet be detectable on biopsy. Thus, without evidence of abnormal Pap results, immediate referral for colposcopy is not recommended for this subgroup. Instead, these women are counseled to return for testing in one year. If infection persists, the risk of precancer and cancer is much higher and referral for colposcopy is then appropriate.
However, as Nicolas Wentzensen, M.D., Ph.D., M.S., a senior investigator in the Hormonal and Reproductive Epidemiology Branch, explained, “It is far from ideal to wait for one year to see if HPV clears naturally. We need new biomarkers to identify infections that are destined to persist and cause precancer or cancer.” Promising biomarkers include p16/Ki67 dual staining, HPV genotyping, HPV genome methylation, and HPV RNA testing. Dr. Wentzensen and colleagues are evaluating several of these biomarkers within Kaiser and another DCEG project called the Study to Understand Cervical Cancer Early Endpoints and Determinants as well as the Costa Rica Vaccine Trial, which provides the opportunity to evaluate biomarkers in a vaccinated population.
In the interest of applying HPV testing globally, Dr. Schiffman and colleagues have investigated whether alternative screening strategies are needed in low-resource countries or in countries in which the age patterns for HPV infection differ from those of the United States. In the United States, the prevalence of HPV infection peaks among women in their 20s and declines with age.
Julia C. Gage, Ph.D., M.P.H., a research fellow in CGB, described a project she conducted in Nigeria, where the prevalence of carcinogenic HPV infection remains high at all ages. “In each region, we need to adjust the screening program to fit the age-specific HPV prevalence pattern and available treatments,” Dr. Gage said. “In sub-Saharan Africa, although HPV prevalence does not decline with age, HPV-based screening programs might still be best targeted to mid-adult women because treatment of precancers found is most feasible before menopause.”
Although the new guidelines are a substantial step forward in incorporating the principles of benchmarking and “equal management of equal risk” into the full screening and management program, the resulting complexity is a challenge for clinicians. The complexity will only increase in the future as guidelines are updated to incorporate new biomarkers, such as HPV genotyping and p16/Ki67, and to address the screening of vaccinated women. One solution is to create a hand-held computer application for clinicians that would present HPV risk estimates based on the best data and evidence from sources all over the United States and the world. Perhaps, the management guidelines could be included as well.
“If guidelines are too complex, clinicians’ and patients’ compliance may be reduced,” Dr. Katki explained. “Calculating risk of cervical precancer or cancer for each combination of clinical tests may be the best way to cut through the complexity. Although these new guidelines are a big step forward, much work remains to bring the benefits of HPV testing to all women worldwide.”