Background and study aims: Histology is the gold standard for the diagnosis of pancreatic adenocarcinoma. However, the negative predictive value of endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) for the diagnosis remains low. The aims of this prospective multicenter study were: (1) to compare the performance of contrast-harmonic EUS (CH-EUS) with that of EUS-FNA for the diagnosis of pancreatic adenocarcinoma; (2) to assess the intra- and interobserver concordances of CH-EUS.
Patients and methods: A total of 100 consecutive patients with a solid pancreatic mass of unknown origin were prospectively included at three centers (July 2009 - April 2010). All patients were examined by CH-EUS followed by EUS-FNA. Absence of vascular enhancement at CH-EUS was regarded as a sign for pancreatic adenocarcinoma. The final diagnosis (gold standard) was based on pathological examination (EUS-FNA, surgery) or 12-month follow-up.
Results: The final diagnoses were: 69 adenocarcinoma, 10 neuroendocrine tumors, 13 chronic pancreatitis, and 8 other lesions. In diagnosing adenocarcinoma, CH-EUS and EUS-FNA had respective accuracy, sensitivity, specificity, positive predictive value, and negative predictive value of 95 %, 96 %, 94 %, 97 %, and 91 %, and of 95 %, 93 %, 100 %, 100 %, and 86 % without significant difference. Five false-negative cases with EUS-FNA were correctly classified by CH-EUS. Interobserver agreement (seven endosonographers) was good (kappa 0.66). Intraobserver agreement was good to excellent (kappa 0.76 for junior, 0.90 for senior).
Conclusions: The performance of CH-EUS for the diagnosis of pancreatic adenocarcinoma was excellent. The good intra- and interobserver concordances suggest an excellent reproducibility. CH-EUS could help to guide the choice between surgery and follow-up when EUS-FNA is inconclusive.
© Georg Thieme Verlag KG Stuttgart · New York.