Abstract
Based on the best estimates of the prevalence of human Papillomavirus infection in the United States, the overall HPV prevalence in the target population is approximately 20%. The prevalence varies greatly with age, being as high as ∼50% in the third decade to less than 5% in the sixth. These data have implications for a discussion about the utility of human Papillomavirus testing as a screening procedure. (
I
The question, therefore, is whether screening for HPV using some sort of molecular diagnostic would be superior for selecting the population at risk for cancer development? The answer to this superficially simple question is unfortunately complex. Part of the problem is technical. Which HPV test should be used? HPV testing, as with all molecular diagnostics, is continually evolving. Until recently there was only one commercially available FDA-approved test for HPVs, the Hybrid Capture tube test (HCT) marketed by Digene Diagnostics. The sensitivity, specificity, and predictive values of that test with a 10 pg/ml cutoff value for a group of 11–13 high-risk viruses has been fairly well characterized. Compared to PCR analysis using L1 consensus primers, the HCT has a lower sensitivity. However, the HCT is more specific for the presence of clinically detectable cervical abnormalities compared with PCR which, because of its higher sensitivity, picks up a significantly higher proportion of patients without clinically detectable disease.
As noted above, molecular technologies continue to evolve rapidly. The next-generation Hybrid Capture test (HC2) is relatively semi-automated, uses a microtiter format, and has up to 50 times the analytic sensitivity of the present test. Whether or not the improved sensitivity is of clinical benefit greatly depends on whether one is using the test for screening vs diagnosis and the population characteristics. The interplay among sensitivity, specificity, and disease prevalence must be constantly kept in mind when the utility of any test is evaluated. Likewise, PCR/amplification technologies are rapidly evolving. In addition, the expanding sequence database of all relevant human papillomaviruses make it likely that the new, powerful DNA-chip technologies may possibly replace or augment current HPV testing methods (Stoler 2000a).
Might HPV testing be a better screening method? This question has been most thoroughly examined by workers in the Netherlands, who have proposed using an extremely sensitive PCR-based method as the first step in a cervical cancer screening program (van Ballegooijen et al. 1997; Stoler 2000a). If one were designing a cervical cancer screening program from scratch, this approach makes a tremendous amount of sense. Almost 100% of the pathology of interest is HPV-positive. Conversely, if an individual were not HPV-positive, the incidence of disease would be so low as to make screening almost worthless. Combining the high prevalence of human papillomaviruses in the pathology of interest with the relatively long time frame from acquisition of infection until the development of the target cervical cancer immediately brings the relative value of initial triage based on HPV status into focus. The lower the prevalence of HPV in the population to be screened, the better the performance profile of an extremely sensitive HPV screening test. For example, the incidence of cervical cancer in women under 25–30 years of age is extremely low, and the prevalence of HPV in the United States drops from approximately 40% at age 20 to 10–20% at age 30–40 (or as low as 4–5% at age 30, as it is in the Netherlands). Under these conditions, it may not make sense to spend resources on screening young women, most of whom develop only transient low-grade lesions. The Dutch proposal seeks to screen the entire population at age 30 with the most sensitive available HPV test combined with a single cytological screening. Patients who are positive on either test would be entered into a program of more intense routine screening, whereas the “double-negative” patients would be returned to the general population pool that would be then screened on a long-interval basis of 5 or even 10 years. Again, if the prevalence of detectable virus is low and the disease prevalence is also low, such a system makes for extremely rational triage and resource utilization. The arguments become even stronger if the cost and reliability of the HPV test become comparable to those of cytological methods. Prospective studies addressing a rational basis for HPV primary screening are needed and planned in the Netherlands and at other sites. Whether or not such a program could be tested in the United States is debatable, given the relative mobility of the United States population and the strongly ingrained emphasis on annual Pap smear screening. However, recent data derived from studies of ASCUS patients support the concept that current generation HPV tests will effectively identify patients with existent HSIL with a sensitivity almost identical to that of colposcopy while sparing almost half the population colposcopic examination (Manos et al. 1999; ALTS Group 2000; Schiffman and Adrianza 2000; Schiffman et al. 2000; Solomon et al. 2001). Such data clearly point towards the development of a primary screening program utilizing an HPV test as the first step.