Interobserver agreement in contrast harmonic endoscopic ultrasound.

ABSTRACT

BACKGROUND AND AIM: Contrast harmonic endoscopic ultrasound (CH-EUS) was recently introduced to clinical practice; its reproducibility among endosonographers is unknown. Our aim was to assess the interobserver agreement (IA) in CH-EUS. METHODS: Fifteen endosonographers (eight experienced and seven non-experienced) from 11 Italian EUS centers evaluated 80 video-cases (40 solid pancreatic lesions, 20 pancreatic cystic lesions and 20 submucosal lesions) of CH-EUS, according to the degree of enhancement, the pattern of distribution and the washout of the contrast agent. IA within each group and between the two groups of observers was assessed with the Fleiss kappa statistic. RESULTS: Overall IA was moderate for the uptake and fair for the pattern of distribution and the washout. In solid pancreatic lesions, IA was moderate for the uptake and fair for the pattern and the washout. In cystic pancreatic lesions, IA was uniformly moderate for the assessment of uptake, slight for the pattern and fair for the washout. In submucosal tumors, IA was substantial for the uptake, slight for the pattern and fair for the washout. Non-experienced endosonographers demonstrated, in most cases, comparable IA with the experienced ones. CONCLUSIONS: Interobserver agreement among endosonographers for CH EUS was satisfactory. In particular, overall IA varied from slight to substantial, being fair in the majority of cases. Inherent structural features of the lesions, as well as technical differences between the variables assessed, could have accounted for the fluctuation of the results. Outcomes of IA were reproducible between experienced and non-experienced endosonographers.