BACKGROUND - Endobronchial ultrasound-guided transbronchial needle aspiration (EBUS-TBNA) is increasingly used to sample mediastinal and hilar lymph nodes and has excellent diagnostic test characteristics. The determinants of sampling adequacy, however, have not been extensively examined. We set out to determine which procedural variables were associated with acquisition of tissue sufficient for pathologic analysis during EBUS-TBNA.
METHODS - A retrospective analysis of all EBUS-TBNA cases performed over 32 months by 10 proceduralists at our institution was completed. Variables potentially associated with sampling adequacy were analyzed.
RESULTS - A total of 1304 procedures performed by 10 proceduralists while the patient received conscious sedation were included for analysis. Sampling adequacy was 94.2% overall and varied with the primary proceduralist (87% to 99.2%; P<0.001). Diagnostic yield per procedure for malignancy or a specific benign diagnosis was 43.2% overall. Proceduralists with a higher average number of lymph node stations sampled per procedure had improved sampling adequacy (parameter estimate=1.32; P=0.007). Sampling adequacy was lower with lymph nodes smaller than 10 mm (parameter estimate=-0.7; P=0.002) but was not associated with procedural environment (hospital procedural suite vs. clinic-based procedural suite) (P=0.08), lymph node station (P=0.69), propofol use (P=0.90), or average annual proceduralist cases performed (P=0.21). Only 6/216 (2.8%) patients had subsequent procedures (EBUS-TBNA or surgery) that indicated the initial EBUS-TBNA had inadequate sampling potentially leading to a missed cancer diagnosis.
CONCLUSIONS - Excellent EBUS-TBNA sampling adequacy can be achieved by pulmonologists in a large group setting, who are not exclusively dedicated to interventional pulmonary medicine, using only moderate conscious sedation.