Factors predictive of malignancy and endoscopic resectability in ampullary neoplasia

OBJECTIVE: Endoscopic treatment of ampullary lesions has been well described, though it remains uncertain if specific features predict malignancy, and whether identifiable factors are associated with successful endoscopic resection of benign lesions. METHODS: Fifty-six consecutive patients undergoing endoscopic evaluation of ampullary neoplasia between March 2000 and May 2004 were included in the study. Clinical presentation, underlying medical conditions, endoscopic treatment, endoscopic ultrasound (EUS) to define extent of local involvement, pathology results, and outcome were documented. Data elements for analysis included EUS findings, lesion lifting with submucosal injection, age, gender, tumor size, and endoscopic intervention. Analyses were performed to determine the ability to predict malignancy and the ability to extirpate benign lesions. RESULTS: Thirty-one males and 25 females were included; mean age was 62 yr. Final diagnoses included 29 adenomas, 20 adenocarcinomas, 4 adenomyomas, 2 paragangliomas, and 1 neuroendocrine tumor. Thirty of 35 patients with benign lesions had extirpation with a mean of two endoscopic procedures. Complications of endoscopic resection included cholangitis (1), bleeding (2), and pancreatitis (4). The presence of malignancy was associated by multivariate analysis with the inability to obtain a cleavage plane with saline injection. Univariate analysis also identified EUS T stage as a predictor of malignancy. In benign lesions, none of the analyzed variables predicted successful endoscopic resection. CONCLUSION: In ampullary lesions, failure to achieve a cleavage plane with submucosal injection is the strongest predictor of malignancy followed by EUS T stage. Endoscopic treatment of benign ampullary neoplasia is effective; no factor was predictive of successful extirpation.