Abstract
Introduction: ROSE entails immediate pathological evaluation of diagnostic specimens in the procedure room, facilitating real-time feedback to operator on specimen adequacy and diagnosis. There is ongoing debate about its role in the field of Endobronchial Ultrasound Transbronchial Needle Aspiration (EBUS-TBNA). Methods: Retrospective cohort analysis of prospectively maintained Greater Manchester EBUS database encompassing all linear EBUS procedures between 01/01/2017 and 31/12/2018 at the ROSE-centre versus all linear EBUS procedures in the same period at a non-ROSE centre. Diagnostic performance plus procedural factors such as lymph node (LN) stations sampled and sedation doses were examined then stratified according to EBUS indication. Results: In total 1650 consecutive EBUS procedures were examined across two centres. Using ROSE resulted in statistically significant reduction in number of nodes sampled, time to pathology and sedation doses for most indications. In 697 staging EBUS, sensitivity at the ROSE-centre was 95% (95% CI 91-97%), negative predictive value (NPV) 93% (95% CI 88-96%) with prevalence of mediastinal nodal metastases of 27% (103/376) versus non-ROSE sensitivity 85% (95% CI 79-90%), negative predictive value 86% (95% CI 80-90%) with prevalence of mediastinal nodal metastases of 32% (103/321). In 329 diagnostic EBUS, using ROSE resulted in a statistically significant reduction in number of LN stations sampled per procedure (median 1 station [1-1] with ROSE vs 2 [1-2], p <0.001). Diagnostic performance was higher at the ROSE centre including increases in sensitivity of 7% for diagnostic EBUS in advanced lung cancer, 20% for Isolated Mediastinal Hilar Lymphadenopathy (IMHL) and 17% for diagnosis of nodal metastases from extra-thoracic malignancy. Conclusion: This study suggests ROSE may provide additional value in diagnostic performance in EBUS and warrants further discussion in an evolving lung cancer and bronchoscopic landscape.