Pancreaticojejunostomy anastomotic stricture (PJAS) is a clinically significant adverse event following pancreaticoduodenectomy. Balloon endoscopy–assisted ERCP (BA-ERCP) is generally considered the first-line approach for PJAS; however, its success is not universal. This study aimed to evaluate the feasibility of endoscopic ultrasonography–guided pancreaticogastrostomy (EUS-PGS) as a drainage method for PJAS.

We retrospectively reviewed 15 patients who underwent endoscopic interventions for PJAS at our institution between February 2015 and July 2025. Technical success was defined as successful drainage via stent placement across the fistula tract or anastomosis, and clinical success as the resolution of pancreas-related symptoms. Adverse events were also assessed.

Fifteen patients were included. The median age was 69 years (range, 18–82 years). Prior surgical procedures consisted of nine pylorus-preserving pancreaticoduodenectomies with Child reconstruction, four subtotal stomach-preserving pancreatoduodenectomies with Child reconstruction, one duodenum-preserving pancreatic head resection, and one middle pancreatectomy with Letton and Wilson reconstruction. Surgical indications included four IPMA cases, three IPMC cases, two chronic pancreatitis cases, two pancreatic ductal adenocarcinoma cases, two bile duct cancer cases, one GIST case, and one traumatic pancreatic injury case. BA-ERCP was attempted in 14 cases, with technical success achieved in 8 (57.1%). EUS-PGS was performed in 6 cases, with technical success in 5 (83.3%). The　initial EUS-PGS failure achieved technical success during the second session. All procedures with technical success also achieved clinical success.　In total, 8 patients were successfully treated with BA-ERCP (BA-ERCP group) and 6 with EUS-PGS (EUS-PGS group), while 1 patient received conservative management after unsuccessful BA-ERCP. The median procedure time was 84 minutes for BA-ERCP and 59 minutes for EUS-PGS (P = 0.181). The main pancreatic duct diameter on CT was significantly larger in the EUS-PGS group than in the BA-ERCP group (5.0 mm vs 3.8 mm, P = 0.02). In the EUS-PGS group, transanastomotic stenting across the PJAS was achieved in 3 cases during the first session and in 2 cases during the second session; 1 case failed both attempts. Adverse events occurred in 2 patients (25%) in the BA-ERCP group, both mild pancreatitis, and in 2 patients (33.3%) in the EUS-PGS group (1 mild pancreatic fistula that resolved conservatively and 1 severe pancreatic fistula requiring endoscopic intervention).

A dilated main pancreatic duct was associated with a lower success rate for BA-ERCP. EUS-PGS appears to be a feasible and effective salvage drainage modality for patients with PJAS.