Diagnostic accuracy and adequacy of rapid on-site evaluation performed by a pulmonologist on transbronchial needle aspiration specimens (DREPA): a prospective study.

Abstract

Rapid on-site evaluation (ROSE) during transbronchial needle aspiration (TBNA) is conventionally performed by pathologists. However, availability of a pathologist in the bronchoscopy suite is often an issue. We aimed to study if a pulmonologist, after receiving a short period of training in cytopathology, is able to assess the adequacy of onsite samples during TBNA. A pulmonologist was initially trained by a pathologist in examining cytology slides and assessing sample adequacy on TBNA smears. During TBNA, one slide from each needle pass was stained on-site using rapid Giemsa stain and was labelled as ROSE slide. The remaining slides were sent to the pathology laboratory for definitive cytological analysis. The ROSE slides were examined by a pulmonologist and a pathologist blinded to each other's interpretation. Level of agreement between the pulmonologist and pathologist was assessed by estimating Cohen's kappa. A total of 172 slides from 35 patients were prepared for ROSE and evaluated independently by pulmonologist and pathologist. For adequacy, the pulmonologist and pathologist agreed in 143 out of the 172 slides (83% agreement), κ 0.649 (p < 0.001). For diagnostic categories, the pulmonologist and the pathologist agreed in 143 out of the 172 slides (83% agreement); κ 0.696 (p < 0.001). The sensitivity, specificity and accuracy of ROSE performed by the pulmonologist with respect to that performed by the pathologist was 66.2%, 96.8% and 83.1% respectively. After a short period of training in cytopathology, a pulmonologist can assess for adequacy of TBNA ROSE slides in the bronchoscopy suite.