Background: The accuracy of EUS in the locoregional assessment of ampullary lesions is unclear.

Objectives: To compare EUS with ERCP and surgical pathology for the evaluation of intraductal extension and local staging of ampullary lesions.

Design: Retrospective cohort study.

Setting: Tertiary-care referral center.

Patients: All patients who underwent EUS primarily for the evaluation of an ampullary lesion between 1998 and 2012.

Intervention: EUS.

Main outcome measurements: Comparison of EUS sensitivity/specificity for intraductal and local extension with ERCP and surgical pathology by using the area under the receiver-operating characteristic (AUROC) curves and outcomes of the subgroup referred for endoscopic papillectomy.

Results: We identified 119 patients who underwent EUS for an ampullary lesion, of whom 99 (83%) had an adenoma or adenocarcinoma. Compared with ERCP (n = 90), the sensitivity/specificity of EUS for any intraductal extension was 56%/97% (AUROC = 0.77; 95% confidence interval [CI], 0.64-0.89). However, when using surgical pathology as the reference (n = 102), the sensitivity/specificity of EUS (80%/93%; AUROC = 0.87; 95% CI, 0.76-0.97) and ERCP (83%/93%; AUROC = 0.88; 95% CI, 0.77-0.99) were comparable. The overall accuracy of EUS for local staging was 90%. Of 58 patients referred for endoscopic papillectomy, complete resection was achieved in 53 (91%); in those having intraductal extension by EUS or ERCP, complete resection was achieved in 4 of 5 (80%) and 4 of 7 (57%), respectively.

Limitation: Retrospective design.

Conclusions: EUS and ERCP perform similarly in evaluating intraductal extension of ampullary adenomas. Additionally, EUS is accurate in T-staging ampullary adenocarcinomas. Future prospective studies should evaluate whether EUS can identify characteristics of ampullary lesions that appropriately direct patients to endoscopic or surgical resection.