Defining the criteria for local resection of ampullary neoplasms

To delineate factors determined preoperatively, which predict successful local resection of ampullary neoplasms.

Retrospective review of case series of the author's experience from 1988 through 1995. The median follow-up of patients with malignancies was 29 months.

Tertiary care university teaching hospital.

Twenty-seven patients underwent surgery. The decision to perform either an ampullectomy or pancreaticoduodenectomy (PD) was based on the size of the lesion, the presence of a "field defect" (ie, familial polyposis), depth of invasion determined by preoperative endoscopic ultrasound, and extent of pancreatic and bile duct involvement seen on endoscopic retrograde cholangiopancreatography.

Fourteen patients underwent ampullectomy, 12 patients underwent PD, and one patient had a retroperitoneal node biopsy performed without resection of the primary tumor.

Resectability, morbidity, and mortality.

Depth of invasion was accurately determined in nine of 12 patients studied by preoperative endoscopic ultrasound. Preoperative endoscopic biopsy specimens were obtained in 21 patients and were inaccurate in seven of 21 cases. The length of stay following local resection was 10.5 +/- 3.7 days vs 15.4 +/- 5.8 days following PD (P=.02). One patient died following PD, and there were no deaths following ampullectomy. Six of 12 patients undergoing PD had postoperative complications vs two of 14 patients undergoing local resection.

Ampullectomy is the procedure of choice for resecting benign lesions smaller than 3 cm, small neuroendocrine tumors, and T1 carcinomas of the ampulla. While endoscopic ultrasonography is helpful in identifying stage T1 lesions suitable for local resection, no preoperative test proved accurate enough to substitute for clinical judgment and intraoperative pathological confirmation.