Endoscopic papillectomy versus surgical ampullectomy for adenomas and early cancers of the papilla: a retrospective Pancreas2000/European Pancreatic Club analysis.

Objective: Ampullary neoplastic lesions can be resected by endoscopic papillectomy (EP) or transduodenal surgical ampullectomy (TSA) while pancreaticoduodenectomy is reserved for more advanced lesions. We present the largest retrospective comparative study analysing EP and TSA.

Design: Of all patients in the database, lesions with prior interventions, benign histology advanced malignancy (T2 and more), patients with hereditary syndromes and those undergoing pancreatoduodenectomy were excluded. All remaining cases as well as a subgroup of them, after propensity-score matching (nearest-neighbour-method) based on age, gender, anthropometrics, comorbidities, size and histological subtype, were analysed. The median follow-up was 21 months (IQR 10-47) after the primary intervention. Primary outcomes were rates of complete resection (R0) and complications. Groups were compared by Fisher's exact or χ test, Mann-Whitney-U-test and log-rank test for survival.

Results: Of 1673 patients in the database, 1422 underwent EP and 251 TSA. Of them, 23.2% were excluded for missing or inconclusive data and 19.8% of patients for prior interventions or hereditary syndromes. Final histology showed in 24.2% of EP and 14.8% of TSA patients a histology other than adenoma or adenocarcinoma while advanced cancers were recorded in 10.9% of EP and 36.6% of TSA patients. Finally, 569 EP and 63 TSA were included in the overall analysis, with a higher rate of more advanced cases and higher R0 resection rates in the TSA groups (90.5% vs 73.1%; p<0.01), with additional ablation in the EP group in 14.4%. Severe adverse event rates were 3.2% (TSA) vs 1.9% (EP). Recurrence after histological R0 resection was 16% (EP) vs 3.2% (TSA; p=0.01), and additional therapy for R1 resection was applied in 67% of the 159 cases. Propensity-score-based matching identified 62 pairs of EP/TSA patients with comparable baseline patient and lesion characteristics. The initial R0-rate was 72.6% (EP) compared with 90.3% (TSA, p=0.02) with recurrences found in 8% (EP) vs 3.2% (TSA; p=0.07); reinterventions were more frequent in the EP group. Overall survival was comparable.

Conclusions: The rate of patients with poor indications due to non-neoplastic disease or advanced cancer is still high for both EP and TSA; multiple retreatments were necessary for EP. Although EP can be considered an appropriate primary therapy for certain ampullary adenomas, case selection for both therapies (especially with regard to the best step-up approach) should be studied further.