Evaluation of Cervicography Screening for Cervical Cancer in a High Risk Population

Statement of the problem CervicographyTM was first described in 1981 as a visual screening system for early detection of cervical neoplasia and cancer. Early studies to assess the validity of cervicography showed the method to have an acceptable sensitivity but an unacceptably low specificity for mass screening. Following revision of the cervicography classification scheme, specificity improved, but at the expense of lowered sensitivity. Most previously published studies have had some methodologic inadequacies which may have affected the outcome. Methods Cervigrams were taken for 8460 women who enrolled into a population-based, natural history study of cervical neoplasia in Guanacaste Province, Costa Rica. Cervicography and three cytologic screening tests were the basis for referral for colposcopic examination and directed biopsy. Initial cervicography classification was compared with a referent diagnosis determined by histology and three cytologic tests, cytology, and presence of cancer-associated human papillomavirus types. Cervicography was submitted to additional review and arbitration to achieve an optimal classification. Interobserver agreement was assessed, and the performance of the optimal cervicography result was compared with the referent diagnosis. Sensitivity, specificity, and predictive values were estimated, and results were stratified by characteristics of the woman and visual characteristics of the cervigram image. Digital colposcopic images were interpreted to evaluate the perceived appropriateness of the decision to biopsy and biopsy placement, and the impact of these on sensitivity and specificity.