Introduction: For non-invasive intraductal papillary and mucinous neoplasm (IPMN) with limited extent, pancreaticoduodenectomy (PD) or distal pancreatectomy (DP) seem excessive due to the risk of pancreatic insufficiency. Enucleation (EN) or medial pancreatectomy (MP) are not commonly performed for IPMN. The aim of this study was to evaluate the feasibility and results of EN and MP for non-invasive IPMN.
Patients and methods: Of 249 patients with IPMN, we attempted a limited resection in 50 (20%) EN (n=31) or MP (n=20) with routine intra-operative frozen section pathology. One attempted EN was converted to MP. Indications for surgery were pain/pancreatitis (44%), suspicion of main duct involvement (28%), mural nodules in branch duct (14%), branch duct>30 mm (8%) or suspicion of mucinous cystadenoma (6%). Follow-up clinical assessment and MRI were performed on a yearly basis.
Results: Of the 31 attempted enucleations, 5 (13%) were immediately converted (4 PD, 1 MP) due to technical reasons (n=3) or due to findings on frozen section (n=2). At definitive pathological examination (accuracy of frozen sectioning=98%), branch ducts were involved by mild (n=21), moderate (n=7) or high grade dysplasia (n=2). One patient underwent a double EN. Of 20 attempted medial pancreatectomies, 8 (40%) required additional segmental resection due to significant IPMN lesions at pancreatic margins; 3 of the additional resection margins were tumor-free, and 5 were involved by IPMN (4 conversions to PD or DP, one contra-indication to PD). Overall, 49 pancreatic margins were analyzed by frozen sectioning with 98% accuracy. Resected specimens of 16 MP showed involvement by mild (n=7), moderate (n=7) or high grade dysplasia (n=2). There was no postoperative mortality. Median length of stay was 21 and 30 days respectively after EN and MP. Pancreatic fistula rate was 54% and 81% respectively after EN and MP. Three patients underwent early re-operation for hemorrhage. Overall median follow-up was 24 months (3-121). All patients are alive, 2 patients (5%) have presented with recurrent pain and 4 have developed tumor recurrence on imaging follow-up (4/33=12%). Two patients (5%) developed de novo diabetes (one after EN combined with DP) and a third patient developed worsening of pre-existing diabetes plus exocrine insufficiency. No patient had surgery for recurrence.
Conclusions: EN and MP are feasible for non-invasive IPMN. Their significant early morbidity is counterbalanced by low rates of both long-term functional disorders and tumor recurrence.