To prospectively evaluate the accordance rate between contrast-enhanced Endoscopic Ultrasound (CE-EUS) and Endoscopic Retrograde Cholangiopancreatography (ERCP) in defining intraductal tumor extension and endoscopic resectability of ampullary adenomas (AAs). Secondary aims were: to describe a tertiary referral centre’s experience with endoscopic papillectomy (EP); to assess the accordance rate between endoscopic tumor size and final histopathology; to evaluate the association between CE-EUS enhancement patterns and final histopathology, including stratification by size; to determine CE-EUS uTNM and final pTNM congruity. 

This prospective monocentric interventional study included all consecutive patients referred for EP between January 2024 and June 2025. Inclusion criteria were: age 18–90 years, endoscopic diagnosis of ampullary tumor, absence of lymph node or distant metastases at staging MRI (if performed), ≥3-month follow-up, and informed consent. Exclusion criteria included prior sphincterotomy/stenting, altered anatomy, and metastatic disease. All patients underwent CE-EUS before ERCP. EUS evaluation included lesion size, ductal involvement, ductal diameters, depth of invasion, lymph nodes, presence of CBD/PD stones, and CE-EUS enhancement pattern. ERCP was performed by endoscopists blinded to EUS findings; cholangiography and wirsungography assessed intraductal growth before papillectomy. EP was performed with hot-snare resection; stenting was routinely performed except for cases with concomitant IPMN. Diagnostic accuracy of EUS and ERCP was analyzed using ROC curves and compared with DeLong’s test. Follow-up duodenoscopy with biopsies was scheduled at 3, 6, and 12 months. Residual intraductal disease was treated with endoscopic resection or intraductal radiofrequency ablation (RFA).

To date, a total of 25 patients (52% male; median age 58 years) were included. Median AAs size was 20 mm. CE-EUS suspected intraductal invasion in 36% (n=9), whereas ERCP confirmed it in 32% (n=8), with a median intraductal extension of 5 mm. All lesions were resected endoscopically, with en-bloc resection in 56%. Post-procedural adverse events occurred in 40%, mainly bleeding, all managed endoscopically or conservatively. Final histopathology showed 17 LGD (68%), 5 HGD (20%), 2 no dysplasia (8%), and 1 adenocarcinoma (uT1bN0M0 / pT1bN0M0). R0 resection was achieved in 56%, R1 in 20%, and R2 in 24%, the latter exclusively in cases with intraductal involvement. Excluding the adenocarcinoma and the two lesions not confirmed as adenomas, 15 out of the 22 confirmed AA (68.2%), were uTis/uT1a on EUS, whereas 7 (31.8%) were overstaged as uT1b despite the absence of invasive behavior at final pathology. No significant association emerged between CE-EUS uT staging and dysplasia grade (HGD: 14.3% in uT1b vs 26.7% in uTis/uT1a; p=0.36). EUS and ERCP showed excellent diagnostic accuracy for intraductal invasion (AUC 0.96 vs 1.00; p=0.9) and high accordance rate (88.9%). One minimal 5 mm ductal extension suspected by EUS was not confirmed by ERCP or pathology. Tumor size and CE-EUS enhancement pattern were not associated with malignancy (p>0.1). Kudo and JNET classifications showed limited accuracy (52%). At 3-month follow-up, residual adenoma was found in 37.5% of cases, mostly those with initial ductal involvement. Endoscopic retreatment, mainly RFA, yielded an overall clinical success rate of 81%. No deaths occurred.

According to our limited experience, in patients without nodal or metastatic disease at MRI, EUS and ERCP showed very high accordance rate in assessing intraductal extension of AAs. ERCP was slightly more specific, whereas EUS tended to overestimate minimal intraductal involvement. Furthermore, EUS overstaged uT-staging in over 30% of AAs. CE-EUS enhancement patterns and endoscopic surface classifications were not predictive of malignancy. Routine pre-papillectomy EUS, even with CE, may therefore not be required for all lesions. EP proved to be safe and effective, with high clinical success using a stepwise endoscopic approach including intraductal RFA when required. Lager multicentre studies are needed to refine criteria for EUS selective application.