Abstract

BACKGROUND

Several factors have been found to affect the accuracy of EBUS-TBNA, however, the effect of the pathologist experience has not been explored.

AIM

We aimed at assessing the effect of pathologist experience in EBUS accuracy in a historical cohort recruited at a tertiary care university hospital.

METHODS

Patients were sequentially recruited for 5 years. All EBUS involved at least one experienced bronchoscopist (experience>100 EBUS), but pathologists had various levels of experience. Definitive diagnosis was made by histopathology (94%) or serial imaging (6%). Factors potentially affecting diagnostic accuracy, including pathologist, bronchoscopist, sex, age, indication, nodal station (NSt), nodal size, were explored using multiple logistic regression adjusting for confounders.

RESULTS

A total of 663 linear EBUS were included. Mean age was 63.7 years + 12.9, 72% were male, most frequent indications were mediastinal adenopathies in patients under suspicion of thoracic (69%) or extra-thoracic cancer (10%) and inflammatory/infectious adenopathies (8%). Median adenopathy size (IQR) was 14mm (10mm-20mm). Most common NSt were 7 (29%), 4R (25%) and 10R (15%). Overall accuracy was 81%. Experience of pathologists (P=0.02) and NSt 4R (P=0.046) were associated with a higher accuracy while adenopathy size<1cm (P=0.007), NSt 11R (P=0.002), 10R (P=0.028), and certain pathologists independently of experience (P=0.003) were related with lower diagnostic accuracy. The bronchoscopist did not affect accuracy when EBUS was performed by at least one experienced bronchoscopist (P=0.13).

CONCLUSION

Experience and other individual factors of pathologists, as well as NSt and nodal size, affected EBUS-TBNA accuracy.