Feature Story

The HPV test as a primary screen

Back to Feature Story

October 2000
William Check, PhD

In the United States, HPV testing is approved as an adjunct to Pap cervical cytology. But in "low-resource" countries, where widespread Pap cytology may not be feasible, it is being evaluated as a possible primary screen.

Dr. Jerome Belinson, immediate past chair of gynecology and obstetrics at the Cleveland Clinic, headed a study in China in which 1,997 women, ages 35 to 45, had six examinations: self-collected swabs for HPV screening (under observation of a nurse); four-quadrant fluorescent spectroscopy readings; a ThinPrep Pap smear, read manually and by an AutoPap instrument; direct HPV detection on the ThinPrep specimen; diagnosis by "visual inspection" using acetic acid; and colposcopy and biopsy, which allowed calculation of true specificity. "We were able to biopsy all patients, even those with normal cytology, because we used a special instrument that takes smaller biopsies and is virtually painless," Dr. Belinson says. Although the majority of the study workers were Chinese, including gynecologic oncologists, gynecologic nurses, and technicians, it was a collaborative project between China and the United States.

This study, Dr. Belinson says, was performed in an area of China that is high-risk for cervical cancer, and it enrolled only women who hadn’t previously been screened and were between the ages of 35 and 45. Accordingly, prevalence of high-grade lesions (=CIN2) was quite high, 4.3 percent. Sensitivity of HPV self-test for high-grade lesions was 83 percent, and specificity was 86 percent. For HPV direct testing, sensitivity was 95 percent and specificity 85 percent. Accuracy of manual cytology depended on the cutoff: For =ASCUS, sensitivity was 94 percent and specificity 78 percent; for =HSIL, the figures were 77 percent and 98 percent (for =CIN2 on biopsy).

Dr. Belinson and colleagues concluded that both HPV direct test and ThinPrep Pap =ASCUS "are highly sensitive" but that "neither is highly specific."

"Both are relatively expensive, and if histologic confirmation is obtained, they become even more expensive," they noted, hardly a prescription for a low-resource setting.

Dr. Belinson is embarking now on a 10,000-patient study in China under more realistic conditions. For example, he says, "Our field person who recruits patients will hand the self-test to the woman and she will do it herself there in the village."

An analogous study in Guanacaste, Costa Rica, was coordinated by Dr. Mark Schiffman, of the Division of Cancer Epidemiology and Genetics, National Cancer Institute. To eliminate referral bias, recruiters went door-to-door to enroll 8,554 subjects, all of whom were examined by HPV testing, conventional Pap cytology, PapNet, manual ThinPrep cytology, cervigram, and visualization by a clinician. If any test was positive, the woman was referred for colposcopy, and a biopsy was taken of any lesion seen on acetostaining. All cancers and most high-grade precursors (=CIN2) were identified, a total of about 140 (1.6 percent).

"Our goal was to determine the ROC [receiver operating characteristic] curve for the available Hybrid Capture II assay," Dr. Schiffman says. The question: What cutpoint would yield the best performance for detecting cervical cancer and its immediate precursors, CIN2 and CIN3? At the optimal cutoff (1.0 pg/mL), the HPV assay was more sensitive than conventional Pap testing (88.4 percent versus 77.7 percent) but less specific (89 percent versus 94.2 percent). Referral rates were 6.9 percent for conventional Pap and 12.3 percent for HPV testing. Thin-layer cytology "matched" the sensitivity and specificity of HPV testing.

In an accompanying editorial (JAMA. 2000;283:108-109), Jack Cuzick, PhD, of the Imperial Cancer Research Fund, wrote that the Guanacaste study and a similar one conducted in Cape Town, South Africa, demonstrated that HPV testing "potentially can enhance the sensitivity of cervical screening and also could be used to reach women who otherwise might not undergo such screening. However," he added, "whether widespread HPV testing is feasible or affordable and whether it will eventually lead to fewer cases of invasive cervical cancer or reduce morbidity and mortality of the disease will require further study. . . . "