Background: s/Aims: Pancreaticoduodenectomy (PD) is the standard treatment for periampullary tumors, but it is technically challenging. Evidence directly comparing open, laparoscopic, robotic, and hybrid approaches is limited. This study conducts a network meta-analysis (NMA) to compare the perioperative and oncologic outcomes of open PD (OPD), laparoscopic PD (LPD), robotic PD (RPD), and hybrid PD. Methods: We searched PubMed, EMBASE, and the Cochrane Library for studies published between January 1994 and August 2024.We included randomized controlled trials and comparative observational studies that evaluated at least two PD approaches. Perioperative outcomes were the primary endpoints, while oncologic safety served as a secondary endpoint. A random-effects NMA was performed, establishing treatment hierarchies through ranking probabilities (PROSPERO ID: CRD420250365864). Results: A total of 78 studies were included (5 randomized and 73 retrospective). RPD was associated with lower blood loss compared to OPD (mean difference [MD], –163.85 mL) and LPD (MD, –84.14 mL). Hospital stays were also shorter for RPD compared to OPD (MD, –2.50 days) and LPD (MD, –1.88 days). In contrast, OPD was the most time-efficient approach compared to LPD (MD, –77.61 minutes) and RPD (MD, –73.30 minutes). Mortality rates, severe complications, clinically relevant postoperative pancreatic fistula rates, and reoperation rates were comparable across all surgical approaches. In terms of oncologic safety, lymph node yield and R0 resection rates were similar for all modalities. Conclusions While OPD is the most time-efficient approach, RPD provides significant advantages in reducing intraoperative blood loss and shortening hospital stays compared to both LPD and OPD.