Pancreatic leak (PL) following pancreaticoduodenectomy (PD) remains one of the most challenging and potentially life-threatening complications in pancreatic surgery. When retrograde endoscopic access to the pancreaticojejuna (PJ) anastomosis is not feasible, treatment options are limited, and surgical revision carries high morbidity and mortality.

Currently, no clear or standardized guidelines exist regarding the optimal endoscopic management of these leaks. In our experience, retrograde approach becomes necessary when the pancreatic fistula coexists with biliary anastomotic dehiscence and/or necrosis of the jejunal limb, as these situations require a complex, multi-target intervention. Conversely, in isolated PLs, retrograde access is often technically demanding and ultimately redundant.

We aimed to evaluate the feasibility, safety, and clinical outcomes of a novel endoscopic technique for the management of PLs after PD: EUS-guided antegrade transanastomotic intrapancreatic stenting (EUS-ATIS) using a fully covered self-expandable metal stent (FCSEMS).

This was a single-center, retrospective case series (July 2022–August 2025) of consecutive PD patients with PLs secondary to PJ dehiscence who underwent EUS-ATIS. Primary endpoints were technical and clinical success. Secondary endpoints included time to leak closure, adverse events (AEs), need for rescue therapy, recurrence, and follow-up outcomes.

Fifteen patients were evaluated; EUS-ATIS could not be performed in 2 due to lack of a safe and adequate endosonographic window for puncture, resulting in a feasibility rate of 86.7%. Thirteen patients underwent successful EUS-ATIS (61.5% male; median age 68 years [IQR 59–73.5]). All had International Study Group on Pancreatic Surgery (ISGPS) grade B postoperative pancreatic fistula; 10/13 presented with high-output PLs, and all were septic at the time of the EUS intervention. In two cases, shorter FCSEMSs (40 mm in length) were initially placed, resulting in persistent leakage. Leak resolution was achieved after placement of a second, coaxial FCSEMS to extend the intrajejunal segment and divert pancreatic flow further downstream within the jejunal limb. Median endoscopic sessions per patients were 1 (IQR 1-2). Median therapeutic delay was 10 days (IQR 6.5–13.5). Technical success and clinical overall success rates were both 100%. Median time to leak closure was 7 days (IQR 5.5–11.5). Early AEs occurred in 3/13 (23%), all non-severe and successfully managed either endoscopically or conservatively. No late AEs or leak recurrences were observed during a median follow-up of 189 days (IQR 95–410). No patient required radiologic or surgical rescue. One death (7.7%) occurred, unrelated to the procedure or PL persistence. Pancreaticogastrostomy stents were removed in 7/13 patients after a median of 3 months (IQR 2–3). FCSEMS were removed in only 2 patients, as retrograde endoscopic access was often unfeasible.

To our knowledge, this represents the first reported use of EUS-ATIS with FCSEMS placement for the treatment of PLs caused by PJ dehiscence after PD. This innovative approach provides a safe, effective, and minimally invasive alternative to surgery by restoring anastomotic continuity, diverting pancreatic flow from the dehiscence site, and promoting definitive healing. Early outcomes are favorable. Continued patients’ enrollment and longer follow-up will be essential to confirm long-term safety and reproducibility.