Cervical cytology is a subjective assessment of cellular changes that could be biased by knowledge of an HPV status. With the advent of primary HPV screening in the Netherlands, the place of cytology will shift from a primary screening role to a secondary triage role. This may not only increase the percentage of abnormal cytology but this can be further enhanced by the knowledge of positive HPV status. We have set up this study to measure the possible effect of knowledge of the HPV status on cervical cytology as can be observed in the new primary HPV screening in the Netherlands.
Prior to the start of the renewed Dutch cervical screening program based on primary HPV testing, we have asked two pairs, each consisting of two experienced Dutch cytologists to examine and classify a set of 100 slides in two rounds, with a wash-out period of at least two months. About 50% of the slides was hrHPV positive. Ideally, this experiment would be conducted in a blinded manner in the five screening laboratories, in order to quantify any possible shift in classification of abnormality (expected to increase by 10-15%). However, this was considered too burdensome for the laboratories. Therefore, the experiment was done with two pairs of volunteers, providing them with limited information, i.e. a measurement in cytology reproducibility. Briefly, a set of glass slides was derived from a blinded pilot of hrHPV cotesting in population based screening in the Netherlands (DuSC study), adjusted for age and expected percentage of abnormalities and cleared of any marker on the slides. During the first round, slides were offered to the experts without any information on hrHPV status, while, after a two-month wash-out period, the reordered slides were offered with information on hrHPV status per slide. Each expert wrote down classification scores during each round. The interrater reliability and the test-retest reliability of the cytologists were calculated by means of Cohen’s Kappa and evaluated.
Results will be shown.
Based on preliminary results, the HPV-bias seems to be limited i.e. below the anticipated 15%. Abnormality rate of cytology should be measured carefully during the implementation of the program with adjustment of cytology criteria if necessary to avoid undue burdening of colposcopy clinics.