Cervical cancer screening with cervicography and cytology

Abstract

The results of colposcopy and directed biopsies were compared with the cytological findings and cervicography in 1539 asymptomatic outpatients undergoing routine cervical cancer screening in order to compare the accuracy of cervicography and cytology and to assess the potential benefits of the combination of both screening tests. With the new reporting criteria, cervicography looked less sensitive (53% vs. 56%; P = 0.718) and significantly less specific vs. 98%; P = 0.040) than cytology. In comparison with cytology and at the expense of a recall rate of 9%, the combination of both screening tests led to an increase of the sensitivity (77% vs. 56%; P = 0.013) but also to a decrease of the specificity (95% vs. 98%; P < 0.0001) and of the positive predictive value (44% vs. 60%; P = 0.051). When the original reporting criteria were applied to cervicography, the combination of both screening tests led to optimal sensitivity and negative predictive value (100%, respectively) at the expense of an unrealistic recall rate of 20%. Consequently, cervicography should not be considered as an alternative to cytology for cervical cancer screening since its accuracy is not significantly better and its rate of technically defective tests is significantly higher. The combination of both methods increases the sensitivity of cervical cancer screening at the expense of a high recall rate, of which cost effectiveness has to be assessed.