Evidence-Based Medicine Versus Liquid-Based Cytology

I this issue, Arbyn et al (see p. 167) update the evidence about the accuracy of liquid-based cytology compared with conventional cytology, and the results are not encouraging.1 In summary, best evidence suggests that liquid-based cytology does not lead to more disease detection and appears to increase false-positive testing, sobering news given that over 80% of U.S. obstetrician–gynecologists surveyed2,3 report using this screening method. The review reveals a surprising lack of high-quality studies performed in screening settings. The recent landmark Italian randomized trial, however, is an exception and deserves focused attention.4 Over 45,000 women aged 25–60 years were randomized to either liquid-based or conventional cytology. Colposcopy was performed in all women with atypical squamous cells of undetermined significance or worse. In comparing these techniques head-to-head, two important clinical questions were answered. First, liquid-based tests were more likely to be interpreted as abnormal and equivocal. Second, despite more positive tests, liquidbased cytology did not lead to detection of more high-grade cervical intraepithelial neoplasia (CIN). In other words, all extra-positive tests appeared to be falsely positive. Arbyn’s meta-analysis of summary data showed similar results. How could a technology be so widely implemented before the appropriate studies have been performed to assess benefits and harms? Reasons are myriad. Many clinicians likely converted to liquid-based cytology to facilitate human papillomavirus (HPV) testing for management of atypical squamous cells of undetermined significance (ASC-US) tests, a strategy designed to decrease unnecessary colposcopy by identifying women at low risk of high-grade CIN (ie, those with negative HPV tests). Does liquid-based cytology combined with HPV testing for ASC-US management decrease uncertainty and overall colposcopy rates compared with conventional cytology with either repeat cytology or HPV testing for ASC-US management? On the basis of results from the Italian trial,5,6 the answer is not obviously “yes.” In that trial, 3.84% of women randomized to liquid-based cytology would have undergone colposcopy for atypical cytology/HPV-positivity or low-grade squamous intraepithelial lesion or worse. In the conventional cytology arm, colposcopy in everyone with ASC-US or worse triaged a near-identical proportion of women to colposcopy (3.81%). Note, however, that colposcopy rates in traditional practice would be lower since the majority of clinicians would not perform colposcopy for a single ASC-US result.2 The reasoning that HPV triage of ASC-US unequivocally justifies use of liquid-based cytology, therefore, is contestable. Some clinicians may have been enticed by claims of fewer unsatisfactory tests with liquid-based cytology. Evidence from randomized trials See related article on page 167.