Abstract:Objective To discuss the value of rapid on-site evaluation (ROSE) in diagnosis of mediastinal lesions by endobronchial ultrasound-guided transbronchial needle aspirations (EBUS-TBNA).Methods Clinical data of 211 patients underwent EBUS-TBNA with or without C-ROSE on chest computed tomography (CT) showed mediastinal/hilar lesions (including enlarged lymph nodes/masses) were retrospectively analyzed.Results The sensitivity of cytological ROSE (C-ROSE) group and No C-ROSE group were 93.88% and 92.45%, the specificity were 97.73% and 93.85%, respectively (P > 0.05). The sensitivity of the ROSE results interpreted by pulmonologist was 87.76% (43/49) and the specificity was 88.64% (39/44), which were highly consistent with the results of professional cytopathologist (κ = 0.806). Under the guidance of ROSE, the completion rate of the first detection of lung cancer 3 genes (EGFR, ALK and ROS1) was 65.52% (19/29), the subsequent immunohistochemical completion rate was 31.03% (9/29).Conclusion ROSE was applied in the EBUS-TBNA process, clinicians and pathologists had a high degree of consistency in diagnosis, and the pathological types could be immediately known, and provide guidance for the subsequent diagnosis and treatment plans.