Diagnostic performance of rapid on-site evaluation during bronchoscopy for lung cancer: A comprehensive meta-analysis

Background: Lung cancer is the leading cause of cancer-related mortality worldwide. Screening high-risk populations for lung cancer with low-dose computed tomography (LDCT) reduces lung cancer mortality. Bronchoscopy is a diagnostic procedure used to monitor patients suspected of having lung cancer after LDCT. Rapid on-site evaluation (ROSE) can improve the diagnostic accuracy of endobronchial ultrasound–guided transbronchial needle aspiration (EBUS-TBNA), although its diagnostic value remains unclear. In this meta-analysis, the authors evaluated the diagnostic accuracy of ROSE during bronchoscopy. Methods: The PubMed, Embase, and Cochrane Library databases were searched for studies evaluating the diagnostic accuracy of ROSE for lung cancer during bronchoscopy. Studies evaluating the performance of ROSE and articles providing sufficient data for constructing a 2 × 2 table on a per-lesion basis were included. A meta-analysis was conducted using a bivariate random-effects model. Results: In total, 32 studies involving 8243 lung lesions were included with a pooled sensitivity of 91.8% and a pooled specificity of 94.9%. Subgroup analysis of 12 studies involving 2929 specimens from patients who underwent computed tomography revealed a pooled sensitivity of 93.8% and a pooled specificity of 96%. Further subgroup analysis of seven studies on the diagnostic outcomes of ROSE for intrathoracic or mediastinal lymph nodes through EBUS-TBNA for lung cancer staging revealed a pooled sensitivity of 90.1% and a pooled specificity of 96.9%. Conclusions: ROSE exhibited high sensitivity and specificity for diagnosing lung cancer during bronchoscopy. It also exhibited high sensitivity in detecting lung cancer in patients undergoing LDCT and higher specificity for nodal staging with EBUS-TBNA.