Clinical value of fully automated p16/Ki-67 dual staining in the triage of HPV-positive women in the Norwegian Cervical Cancer Screening Program

Abstract

BACKGROUND: More accurate biomarkers in cervical cytology screening could reduce the number of women unnecessarily referred for biopsy. This study investigated the ability of p16/Ki-67 dual staining to predict high-grade cervical intraepithelial neoplasia (CIN) in human papillomavirus (HPV)–positive women from the Norwegian Cervical Cancer Screening Program. METHODS: Automated p16/Ki-67 dual staining was performed on liquid-based cytology samples from 266 women who were HPV-positive at their secondary screening. At a mean of 184 days after p16/Ki-67 staining, 201 women had a valid staining result and a conclusive follow-up diagnosis (histological diagnosis or HPV-negative diagnosis with normal cytology findings). The sensitivity and specificity for predicting the follow-up diagnosis were compared for cytology, p16/Ki-67 dual staining, and their combination. RESULTS: Sixty-seven percent of the study sample was p16/Ki-67–positive. The sensitivity of p16/Ki-67 staining for predicting CIN-2/3 was statistically significantly higher than the sensitivity of cytology (0.88 vs 0.79; P =.008), but this was not true for the prediction of CIN-3 (0.94 vs 0.88; P =.23). The specificity of cytology for predicting CIN-3 was significantly higher than the specificity of p16/Ki-67 staining (0.35 vs 0.28; P =.002), but this was not true for CIN-2/3 (0.35 vs 0.31; P =.063). For predicting CIN-2/3 and CIN-3, combination testing gave potentially better sensitivity (0.95 and 0.96, respectively) and better specificity (0.49 and 0.50, respectively). CONCLUSIONS: In a population of HPV-positive women, p16/Ki-67 dual staining was more sensitive but less specific than cytology for predicting high-grade CIN. The advantage of using both tests in different combinations is the potential for increasing the specificity or sensitivity in comparison with both methods performed individually. Cancer Cytopathol 2017;125:283–291. © 2016 American Cancer Society.