Page 65 - Screening for Cervical Cancer: Systematic Evidence Review

Chapter III. Results

Herrington and colleagues studied 167 patients with 99 with persistent “borderline”

69
smear, 39 with “wart virus changes”, and 29 with “mild dyskaryosis.” Like women in the

United States, German women with moderate or severe dysplasia are referred immediately for

colposcopy and possible treatment. The group evaluated in this study is, therefore, much like the

group for whom HPV triage would be of interest in the United States. In this setting, consensus

PCR had sensitivity of 87.5%, specificity of 62.2%, positive predictive value of 42.2% and

negative predictive value of 94.0% for detection of histology proven high-grade lesions.

Bollen and colleagues studied British women including only those with mild or moderate

66
dysplasia or cytology. Of the test systems they evaluated in their patient population, SHARP

PCR (for high-risk HPV types) was relevant to this review. The sensitivity for detecting HSIL

was 94.6%; specificity, 40.3%; positive predictive value, 39.8%; negative predictive value,

94.7%. In these 3 cytology-defined studies as well as the colposcopy-referral studies, the

negative predictive value of testing for high-risk HPV types is greater than 90%, suggesting the

feasibility of using negative HPV results to lengthen the time until repeat Pap test and to reduce

the number of women who have colposcopy.

Benefits

The benefits of HPV testing as a screening test, an adjunct to screening, or a triage tool

are not documented in prospective cohort studies or trials evaluating outcomes among women

who received conventional cytology only compared to an alternate strategy that includes HPV

testing.

65