The Effect of the Operator Learning Curve on Diagnostic Accuracy for Lung Cancer with Endobronchial Ultrasound Guided Transbronchial Needle Aspiration (EBUS-TBNA): A Single-Center Observational Study

Abstract

We aimed to investigated the relationship between diagnostic accuracy of ultrasound-guided trans-bronchial needle aspiration (EBUS-TBNA) and the number of procedure performed by operator. We included all consecutively attended patients undergoing EBUS-TBNA for diagnosis of lung cancer. The pathology prearations were evaluated by the same single pathologist. Rapid on-site examination (ROSE) was performed for all patients. To determine the impact of operator experience on the accuracy of EBUS-TBNA-based diagnosis, patients were divided into three equally sized groups based on the chronological order of their EBUS-TBNA assessment. A total of 270 patients were included in the study. The most common indication for undergoing the EBUS-TBNA procedure was tumor staging (44.1%). ROC analysis revealed that the study bronchoscopist's diagnosis of malignancy based on EBUS-TBNA images were consistent with the final pathologist's diagnosis in 73% of the first 90 cases (p= 0.02), 90% of the second 90 cases (p= 0.01), and 98% of the third 90 cases (p= 0.01). In the first 90 cases, the sensitivity of the bronchoscopist's diagnosis was 74%, the specificity was 78%, the positive predictive value was 76%, and the negative predictive value was 85%. In the second 90 cases, the sensitivity was 94%, the specificity was 92%, the positive predictive value was 95%, and the negative predictive value was 95%. In the third 90 cases, the sensitivity was 98%, the specificity was 98%, the positive predictive value was 98%, and the negative predictive value was 100%. We observed that as the number of EBUS-TBNA procedures that the operator performed increased, his diagnostic accuracy approached perfection.