Test positivity cutoff level of a high risk human papillomavirus test could be increased in routine cervical cancer screening

Abstract

We present data on test positivity, relative sensitivity, rates of detection and relative specificity for primary human papillomavirus (HPV) testing with different cutoff levels for test positivity, in comparison to conventional cytology. In 2003-2004, 18,438 women were screened primarily with Hybrid Capture 2 (HC 2) assay, a test for oncogenic HPV DNA, and 21,446 with conventional cytology within the organised screening programme in Finland. A cytological triage test was performed for the HPV positives. Women with cytology equal to low grade squamous intraepithelial lesion (LSIL) or worse were referred for colposcopy. The relative sensitivity measured as relative risk (RR) of any cervical intraepithelial neoplasia (CIN) or cancer was 1.58 for the HPV test at the relative light units (rlu) ratio cutoff 1.00, in comparison to cytology. With the cutoff 3.00, all CIN 2+ lesions were detected. With cutoff 10.00, 2 of the 22 CIN 3+ lesions were missed. Relative specificity for HPV screening for any CIN was 92.6% at cutoff 1.00, 94.6% at cutoff 3.00 and 96.3% at cutoff 10.00. For CIN 3+ specificity estimates for these cutoffs were 92.1%, 94.1% and 95.8%, respectively. Used for routine screening as the sole test, the HPV test cutoff can be increased from the level recommended for clinical use. With HC 2, the detection rate at rlu ratio cutoff 10.00 is still at the level of high-quality conventional screening. At that level, the false positive rate is reduced by about half and the specificity of the HPV test becomes equal to the average specificity of conventional cytology.