Background. Although the clinical benefit of rapid on-site cytologic evaluation (ROSE) during flexible bronchoscopy has been reported, its quality management has not been fully investigated. Here, we report a one-year experience of quality control with ROSE in our institution. Objective and Methods. Between October 24, 2013 and October 31, 2014, ROSE was done in 183 of 269 patients who underwent bronchoscopy in our division. ROSE was compared with pathological diagnosis in terms of its sensitivity, specificity, positive predictive value, negative predictive value, and accuracy. During the study period, diagnostic results were reviewed among pulmonologists, cytological screeners, and pathologists. Results. Malignant tumor was the most common diagnosis (110 patients). Discordant cases were reviewed every four months and diagnostic assessments were partially modified. As a result, positive predictive value increased while accuracy was maintained. Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of ROSE compared with pathological diagnosis were 83%, 85%, 89%, 77% and 84%, respectively. Conclusion. Introducing ROSE in our department, quality control was properly done with periodical review.
View full abstract