The triage of human papillomavirus (HPV)-positive women is needed to avoid overreferral to colposcopy. p16(INK4a) immunostaining is an efficient triage method. p16(INK4a) /Ki-67 dual staining was introduced mainly to increase reproducibility and specificity compared with stand-alone p16(INK4a) staining.
Within a pilot project, HPV-positive women were referred to colposcopy if cytology was abnormal or unsatisfactory or HPV testing was still positive after 1 year. For 500 consecutive women, a slide obtained during colposcopy was immunostained for p16(INK4a) /Ki-67 and independently interpreted by 7 readers without previous experience with dual staining. Four of these readers were experts in cervical pathology and 3 were not. Kappa values for multiple raters, sensitivity, and specificity for cervical intraepithelial neoplasia type 2-positive histology were computed. Because women with normal cytology were underrepresented, estimates for all HPV-positive women were obtained as weighted means of cytology-specific estimates.
The overall kappa for HPV-positive women was 0.70 (95% confidence interval [95% CI], 0.60-0.77). Kappa values were not found to be significantly different between expert and nonexpert readers with regard to cervical cytology but were significantly increased (P =. 0066) after consensus discussion. The overall specificity estimate for HPV-positive women was 64.0% (95% CI, 57.4%-70.2%): 66.7% (95% CI, 59.8%-73.0%) for experts and 60.5% (95% CI, 59.8%-73.0%) for nonexperts. Among women with abnormal cytology, the sensitivity was 85.5% (95% CI, 77.9%-90.8%): 85.8% (95% CI, 77.9%-91.2%) for experts and 85.1% (95% CI, 76.6%-90.9%) for nonexperts.
p16(INK4a) /Ki-67 immunostaining demonstrated good reproducibility and specificity when triaging HPV-positive women. Dual-staining interpretation can be performed, after short training, even by staff who are not experts in cervical cytology. This allows HPV-based screening with triage to be performed in settings in which such expert staff is not available.