Key points are not available for this paper at this time.
OBJECTIVE: To calculate pooled risk estimates for combinations of cytology result, human papillomavirus (HPV) 16/18 genotype and colposcopy impression to provide a basis for risk-stratified colposcopy and biopsy practice. DATA SOURCE: A PubMed search was conducted on June 1, 2016, and a ClinicalTrials.gov search was conducted on June 9, 2018, using key words such as "uterine cervical neoplasms," "cervical cancer," "mass screening," "early detection of cancer," and "colposcopy." METHODS OF STUDY SELECTION: Eligible studies must have included colposcopic impression and either cytology results or HPV 16/18 partial genotype results as well as a histologic biopsy diagnosis from adult women. Manuscripts were reviewed for the following: cytology, HPV status, and colposcopy impression as well as age, number of women, and number of cervical intraepithelial neoplasia (CIN) 2, CIN 3, and cancer cases. Strata were defined by the various combinations of cytology, genotype, and colposcopic impression. TABULATION, INTEGRATION, AND RESULTS: Of 340 abstracts identified, nine were eligible for inclusion. Data were also obtained from three unpublished studies, two of which have since been published. We calculated the risk of CIN 2 or worse and CIN 3 or worse based on cytology, colposcopy, and HPV 16/18 test results. We found similar risk patterns across studies in the lowest risk groups such that risk estimates were similar despite different referral populations and study designs. Women with a normal colposcopy impression (no acetowhitening), less than high-grade squamous intraepithelial lesion cytology, and HPV 16/18-negative were at low risk of prevalent precancer. Women with at least two of the following: high-grade squamous intraepithelial lesion cytology, HPV16- or HPV18-positive, and high-grade colposcopic impression were at highest risk of prevalent precancer. CONCLUSION: Our results support a risk-based approach to colposcopy and biopsy with modifications of practice at the lowest and highest risk levels.
Silver et al. (Tue,) studied this question.