Interpreting microbiopsies in cervical smears - A cytohistologic approach

OBJECTIVE: To assess interobserver variation in the diagnosis of thick tiss ue specimens (microbiopsies) in cytology smears and histologic sections tak en from them, to evaluate the applicability of MIB-1 in histologic sections from microbiopsies and to evaluate whether processing microbiopsies in inc onclusive smears has additional diagnostic value. STUDY DESIGN: Cytologic smears were selected in which there were diagnostic disagreements between pathologists and cytologists and microbiopsies were present. Interobserver variation among three pathologists and three cytolog ists in the diagnosis of these microbiopsies was investigated. The smears w ere processed for histologic sections, and interobserver variation between pathologist diagnoses were analyzed. An additional histologic slide stained for MIB-1 was used far consensus diagnosis. The consensus diagnosis was co mpared with available follow-up and its sensitivity and specificity determi ned. The value of applying the microbiopsy technique in slides diagnosed as inadequate or atypical squamous cells of undetermined significance (ASCUS) was analysed. RESULTS: From a series of 62,334 cervical smears, 49 with microbiopsies wer e selected. If was possible to derive histologic slides from 38 cases. Inte robserver variability in the diagnosis of microbiopsies and histologic sect ions from them was moderate- kappa =.44 (SE =.06) and kappa=.44 (SE =.09), respectively. In the consensus meeting for all cases, a conclusive diagnosi s was reached. The Pearson correlation coefficient between the consensus di agnosis and MIB-1 staining was r =.62. The sensitivity of the consensus dia gnosis for the follow-up diagnosis was 71% and the specificity 60%. Diagnos es on approximately 50% of slides diagnosed as inadequate or ASCUS could be made. CONCLUSION: The histotechnical workup of microbiopsies is not difficult; ho wever, their diagnosis can be a problem. Adequate diagnostic criteria are n ot available. Aided by MIB-1 staining, histologic sections from microbiopsi es can be diagnosed, and the diagnoses correlated with follow-up in most ca ses. Processing of microbiopsies in smears with an inconclusive cytologic d iagnosis or a diagnosis of ASCUS allowed correct diagnosis in 50% of cases in this study.