Learning curve of electromagnetic navigation bronchoscopy for diagnosing peripheral pulmonary nodules in a single institution

Background: Electromagnetic navigation bronchoscopy (ENB) is a novel technology that is designed to diagnose peripheral pulmonary lesions (PPLs). The purpose of the study was to explore the learning curve and evaluate the diagnostic yield and safety of ENB in diagnosing peripheral pulmonary nodules. Methods: A total of 40 patients discovered with peripheral pulmonary nodules suspicious for malignancy were chronologically enrolled into the study. Biopsy, brushing and flushing specimens were obtained through the extended working channel (EWC). Radial endobronchial ultrasound (R-EBUS) and fluoroscopy were used as means for localization and auxiliary implements in the procedure of ENB. Immunohistochemistry (IHC) and driver genes testing were performed on biopsy samples when it was necessary. All the patients performed chest radiographs to exclude pneumothorax after the procedure. Data of all patients were recorded prospectively and analyzed retrospectively. The learning curve was evaluated using cumulative sum (CUSUM) method. Results: The mean diameter of the 40 nodules was 21.1±5.3 mm. The average navigation time and total operation time were 8.1±3.2 and 24.6±4.1 min, respectively. The overall diagnostic yield was 82.5% (33/40). CUSUM analysis of learning curve based on the navigation time and total operation time both could be best modeled as polynomials and divided the learning curve into two phases at the point of case 14. The learning curve based on diagnostic yield did not demonstrate an obvious turning point. No complications occurred in the 40 procedures. Conclusions: ENB is a safe technology with a high diagnostic yield in diagnosing peripheral pulmonary nodules. The learning curve based on the procedure time could be stable after 14 procedures and there was no obvious learning curve based on the diagnostic yield for an experienced pulmonologist.