Intraductal papillary mucinous neoplasms (IPMN) are important premalignant neoplasms of the pancreas. Extent of resection can be difficult to determine as pancreatic ductal involvement is often diffuse but the morbidity of total pancreatectomy is rarely warranted. We reviewed our experience with resected IPMN to characterize the importance of frozen section in determining extent of resection and long-term outcome. Seventy patients (33 women, 37 men; mean age 64.4) operated on for IPMN from 1992 through 2008 were reviewed. There were 39 patients with main duct disease (MD), 28 with side-branch disease (SB), and 3 with mixed disease. Of patients with MD and mixed disease, 11 (26%) had pancreatic duct dilation (diameter > 5 mm) localized to the head, 4 (10%) localized to the tail, 16 (38%) had diffuse duct involvement, and 12 (29%) had a mass. Based on preoperative imaging and operative findings, 9 (13%) underwent total pancreatectomy, 38 (54%) pancreaticoduodenectomy, 22 (31%) distal pancreatectomy, and 1 (1.4%) was managed with central pancreatectomy. Final pathology of the resected specimen was characterized as adenoma in 17 (24%), borderline in 26 (37%), carcinoma in situ in 13 (19%), and invasive carcinoma in 14 (20%). Frozen section examination of the pancreatic margin was performed in 49 of the 61 patients that underwent partial pancreatectomy. Frozen section was characterized as low grade dysplasia (LGD) in 9 patients (18%), IPMN/adenoma in 8 (16%), and “negative for malignancy” in 32 (65%). There were no cases of carcinoma in situ or invasive cancer on frozen section. Additional resection occurred in two patients as a result of the frozen section findings; one for a focus of IPMN and one for foci of dysplasia at the margin. At a mean follow-up of 17.4 months (range 1 to 118 months), three patients (4.3%) had invasive carcinoma detected in the pancreatic remnant with the final pathology of the original specimen being: main duct IPMN/invasive cancer, main duct IPMN/carcinoma in situ, and side-branch IPMN/borderline. Pathology of the pancreatic duct margin in these patients was benign IPMN in the first patient and “negative for neoplasm” in the other two. In conclusion, the preoperative assessment of extent of ductal disease is the most valuable tool in predicting the extent of resection. The principal utility of frozen section on the resection margin is in determining the presence of invasive cancer. Long term follow-up of the remnant is important following resection of all patients with IPMN regardless of original pathology or margin status.