INTRALABORATORY QUALITY ASSURANCE IN CERVICAL VAGINAL CYTOLOGY - EVALUATION OF INTERCYTOLOGIST DIAGNOSTIC REPRODUCIBILITY/

Diagnostic reproducibility and accuracy in evaluating cervical/vaginal smears were the focus of this study concerning intralaboratory qualit y control. A set of 120 cytological samples was evaluated by 15 cytopa thologists whose experience ranged from 3-29 yr The study report form was based on the 1988 Bethesda System. Intercytologist reproducibility (with respect to sample adequacy, epithelial cell abnormalities, and presence of cellular changes associated with HPV, Human Papilloma viru s) was evaluated using the Kappa statistic. Poor reproducibility in de fining sample adequacy was observed (K = 0.24). The agreement on epith elial cell abnormality definition was good (K = 0.64); the lowest repr oducibility was observed for High-grade squamous intraepithelial lesio n/cervical intraepithelial neoplasia HSIL/CIN II (specific K = 0.37) a nd Atypical cells of undetermined significance ACUS (specific K = 0.44 ). The agreement in detecting cellular changes associated with HPV was fair (K = 0.49). Diagnostic accuracy Sol epithelial cell abnormalitie s, assessed by comparison with a consensus reference diagnosis was goo d (K = 0.74). (C) 1997 Wiley-Liss, Inc.