ENDOSCOPIC SNARE PAPILLECTOMY OF AMPULLARY ADENOMAS: PREDICTORS OF INCOMPLETE RESECTION AND POST-PAPILLECTOMY PANCREATITIS

BACKGROUND: Endoscopic Snare Papillectomy (ESP) is a minimally invasive option for the management of noninvasive ampullary adenomas. The primary aim of this study is to determine risk factors for incomplete resection of ampullary adenomas following ESP, as well as the impact of rectal indomethacin and pancreatic duct (PD) stenting on post-papillectomy pancreatitis (PPP).

METHODS: This retrospective multicenter study included consecutive patients undergoing index ESP for ampullary adenomas from 1/2012-9/2020 at two referral centers. Patient demographics, procedural indications, details, adverse events, pathology, and subsequent interventions were collected. Wilcoxon rank-sum test and Fisher’s exact test were used to compare continuous and categorical variables. Stepwise logistic regression was used to identify predictors for residual disease and PPP.

RESULTS: Of 136 included patients, 24 (18%) carried a diagnosis of Familial Adenomatous Polyposis. Overall median adenoma size was 20mm (IQR, 15mm, 25mm) and 47 (35%) adenomas were laterally spreading. Common bile duct extension was seen on pre-procedure EUS or index ERCP in 25 (19%) patients. En bloc resection was used in 77 (57%) patients. Advanced neoplasia was seen in 60% of patients (Table 1). Median duration of surveillance was 899 days (IQR 382, 2041 days). Residual neoplasia was seen in 66 (49%) patients on first surveillance endoscopy, of whom 6 (9%) had biopsy-confirmed residual malignancy. Factors not meeting inclusion threshold during stepwise regression (P<0.20) were adenoma size, location, and presence of periampullary diverticulum. Factors associated with increased risk of residual adenoma were biliary extension (OR 6.87, 95% CI 2.10-22.48) and piecemeal resection (OR 2.32, 95% CI 1.09-4.92) (Table 2). Median number of endoscopic sessions required for successful ampullary adenoma eradication was 1 (IQR 1 session, 2 sessions). Surgical pancreaticoduodenectomy was ultimately required in 13 (10%) patients.

PPP occurred in 20 (15%) patients. Prophylactic PD stent was placed in 122 (90%) patients, of whom 62 (51%) underwent PD stenting alone and 60 (49%) received combination PD stenting and rectal indomethacin. Decreased rates of PPP were seen with PD stenting (OR 0.22, 95% CI 0.06-0.78, P=0.020), but not with rectal indomethacin alone (OR 0.82, 95% CI 0.32-2.12) or in combination with PD stenting (OR 0.49, 95% CI 0.18-1.37).

CONCLUSIONS: Residual adenoma was seen in nearly half of patients with ampullary adenomas after ESP. CBD extension and piecemeal resection were predictive of residual disease. ESP was associated with a 15% rate of PPP, which was lowered by PD stenting. Reduction of PPP was not demonstrated by the addition of rectal indomethacin or indomethacin alone.