Background: Screening for cervical cancer precursors by Papanicolaou cytology is a public health success story; however, its low sensitivity entails unnecessary referrals to colposcopy of healthy women with equivocal (ASCUS) or mild dysplasia (LSIL) cytology.
Objective: We assessed the accuracy of p16/Ki-67 immuno-testing for triage of low grade cervical cytology.
Search strategy: We systematically searched Medline, Embase, CRD and Cochrane databases, and handsearched key references.
Selection criteria: Eligible studies included women with ASCUS or LSIL cervical cytology who had undergone p16/Ki-67 testing and subsequent verification by colposcopy-directed biopsies and histologic analysis.
Data collection and analysis: We extracted data on patient characteristics and test conduct, diagnostic accuracy measures and assessed the methodological quality of the studies. R software was used to perform a bivariate analysis of test performance data.
Main results: Five eligible studies were identified. Four of the studies had high risk of bias. In the LSIL subgroup, the sensitivity of p16/Ki-67 testing ranged from 0.86 to 0.98, compared with 0.92-0.96 of high-risk HPV testing (hrHPV); specificity ranged from 0.43 to 0.68 versus 0.19 to 0.37, respectively. In the ASCUS subgroup, sensitivity ranged from 0.64 to 0.92 (p16/Ki67 test) versus 0.91 to 0.97 (hrHPV); specificity ranged from 0.53 to 0.81 versus 0.26 to 0.44, respectively.
Authors' conclusions: p16/Ki-67 testing cannot be recommended for triage women with ASCUS or LSIL cytology due to insufficient high-quality evidence. Further studies on test performance and the impact of p16/Ki-67-based triage on health outcomes are needed for a definitive evaluation of its clinical utility.
Keywords: Cervical cancer screening; Ki-67; human papillomavirus; molecular diagnosis and prognosis; p16INK4A.
© 2014 Royal College of Obstetricians and Gynaecologists.