In this study, robot-assisted surgery demonstrated safety equivalent to laparoscopic surgery, with no significant difference in early complication rates. Despite a shorter follow-up period, robot-assisted surgery was associated with accelerated postoperative recovery, particularly in pediatric patients, and showed the potential to reduce long-term postoperative complications.

Complications after CBD surgery can occur on both the hepatic and pancreatic sides. On the hepatic side, IHBD stenosis is a known risk factor for hepatolithiasis, cholangitis, and cholangiocarcinoma. Most IHBD stenoses in the CBD are considered congenital, resulting from septal or membranous strictures [7]. Hepatolithiasis can lead to recurrent cholangitis, a major risk factor for bile duct cancer [1, 2].

To mitigate these risks, we routinely performed bile ductoplasty for hilar IHBD stenosis in both open and laparoscopic surgery, and favorable outcomes of hilar bile ductoplasty have been previously reported [7]. Careful examination and removal of the membrane or septum that causes IHBD stenosis have been shown to effectively prevent hepatolithiasis after surgery. In this study, we performed bile ductoplasty in more than 50% of robot-assisted surgeries. These procedures increased the operative time, but did not increase complications and demonstrated feasibility comparable to that of laparoscopic procedures. The robotic platform offers advantages, such as tremor filtration, motion scaling, and three-dimensional visualization, which enable more precise and thorough dissection. In this study, the robot’s 3D magnification and multi-joint functionality enabled relatively easy and safe bile ductoplasty for incision and suturing of a stricture area that would be difficult to access laparoscopically (located deep in the hepatic hilum, with the direction of the operation on the stricture area not matching the axis direction of the forceps) (Supplementary Fig. 4). Several studies reported that robot-assisted surgery was more effective than laparoscopic surgery, particularly in terms of anastomosis and suturing [10, 11]. Although bile ductoplasty in robot-assisted surgery has not been widely evaluated, our findings suggest that the system facilitates accurate handling of stenotic lesions. In the Rob group, no anastomotic stenosis and only one case of hepatolithiasis/cholangitis were observed, despite the shorter follow-up period.

However, certain peripheral IHBD stenoses, particularly those located deep within the liver, cannot be surgically addressed. Stenosis in the hilar region or near the secondary branches can be removed using intraoperative bile ductoplasty. In one patient in the Rob group, cholangitis developed postoperatively due to pre-existing stenosis in the right posterior section, which could not be surgically corrected. These deep IHBD anomalies are rare but warrant long-term follow-up. Intraoperative bile ductoplasty remains important in reducing the risk of complications.

In the pancreas, residual IPBD can cause postoperative pancreatitis, pancreatolithiasis, and carcinoma [12]. Therefore, complete IPBD excision should be performed. As far as we know, no reports have evaluated robot-assisted surgery for the CBD in detail with respect to intrapancreatic ductal treatment. In this study, no significant difference in the incidence of residual IPBD or related complications was observed between groups. In CBD surgery, we perform intraoperative cholangiography to evaluate the anatomy of the bile duct and pancreatic duct, and then resect the intrapancreatic bile duct at the very edge of the pancreatic duct junction. The detachment of the IPBD requires extremely delicate manipulation to avoid damaging the blood vessels and pancreatic tissue surrounding the bile duct. Robot-assisted surgery provides a high-definition 3D view and magnified field of view, enabling clearer visualization of fine blood vessels and the boundaries between pancreatic tissue and bile ducts. This allows safe and accurate dissection. In addition, multi-jointed forceps and image stabilization enable manipulation in deep, narrow spaces that are difficult to access with laparoscopic surgery, allowing for more precise and stable incisions and dissections.

In the Rob group, hospital stay was significantly shorter, enteral feeding was started earlier, and drain duration was reduced. These benefits were even more pronounced in pediatric patients. International multicenter studies and meta-analyses have also demonstrated the shorter postoperative hospital stay or fewer postoperative complications associated with robot-assisted surgery, demonstrating its safety and feasibility [6, 13, 14]. These findings suggest that robot-assisted surgery may reduce surgical trauma, minimize bowel manipulation, and expedite the recovery of intestinal function.

This study included a wide patient population ranging from neonates weighing approximately 3 kg to adult patients weighing up to 90 kg aged 60 years. This diversity not only adds to the generalizability of the study but also raises important questions regarding patient selection, surgical adaptation, and perioperative management across age groups. It is especially remarkable that robot-assisted surgery demonstrated comparable or even superior short-term outcomes, such as shorter hospital stay and faster initiation of enteral feeding, despite the inclusion of adult cases. Furthermore, the subgroup analysis suggested that robot-assisted surgery accelerates postoperative recovery, especially in pediatric patients. Minimally invasive surgery is particularly important in children, and these results support the widespread use of robot-assisted surgery in pediatric surgery.

The operative time was significantly longer in the Rob group than in the Lap group, likely due to the greater proportion of adult cases. Adult patients tend to have more severe inflammation and adhesions owing to delayed diagnosis. In addition, in adult cases, the Roux-en-Y limb was constructed intraperitoneally and routed via the retrocolic pathway, with intraperitoneal fixation of the limb and closure of the mesenteric defect and Petersen’s space. These technical steps, which are not typically required in pediatric cases, may have contributed to longer operative times. Nonetheless, subgroup analyses showed comparable operative durations between the Rob and Lap groups when stratified by age, suggesting that these differences were due to patient background and procedural complexity rather than the robotic approach itself. Moreover, previous studies indicated that although the total operation time was longer, robot-assisted surgery resulted in significantly shorter cyst resection and hepaticojejunostomy times [15]. There are reports that robot-assisted surgery is useful in reducing the time required for hepaticojejunostomy compared to laparoscopic surgery, not only in children but also in adults [16, 17]. The long duration of hepaticojejunostomy in robot-assisted surgery was thought to be due to the time required for docking and the time needed to change the equipment. It was suggested that the longer surgery times in the Rob group were due to the higher number of adult cases, and once the surgeon overcomes the learning curve, surgery time would be further reduced.

Laparoscopic surgery was performed from 2013, while robot-assisted surgery began in 2021, reflecting changes in surgical techniques over time. Analysis limited to the period when both procedures were performed (2021–2023) showed that the significant difference in enteral feeding initiation period and drain placement period disappeared; however, a significant difference in length of hospital stay remained. These results suggest that advances in perioperative care, imaging, and surgeon experience may have improved overall outcomes in the Rob group, demonstrating superiority in comparisons across all periods. Logistic regression analysis demonstrated that robot-assisted surgery was an independent predictor significantly reducing the risk of complications. Furthermore, Cox regression analysis demonstrated that robot-assisted surgery was an independent predictor for shorter hospital stays, earlier initiation of feeding, and shorter drain retention periods. After adjusting for era effects and patient background, robot-assisted surgery was suggested to potentially yield improvements in postoperative outcomes. Therefore, this study suggested that robot-assisted surgery may have an advantage over laparoscopic surgery in accelerating postoperative recovery.

In view of the fact that cases of non-dilatation type were included in the cohort, which could act as a confounding factor, a sensitivity analysis was conducted. Sensitivity analysis showed that excluding these cases did not significantly alter our main findings. These results indicate that the presence or absence of bile duct dilatation did not significantly affect the conclusions of this study, reinforcing the robustness of our main findings. However, for a more detailed analysis, large-scale studies involving a large number of cases of non-dilatation type will be necessary in the future.

We describe some limitations. First, this was retrospective study. However, our center has performed the largest number of robot-assisted CBD surgeries in Japan; selection bias may have been introduced, especially as more adult cases were included in the Rob group due to the recent expansion of robot-assisted surgery. The significantly high proportion of adult patients in the robot group may still cause bias in the results. We did not intentionally choose laparoscopic or robot-assisted surgery based on the difficulty of the cases. Laparoscopic and robot-assisted surgeries were performed in patients requiring emergency surgery due to bile duct perforation. Second, the shorter follow-up period in the Rob group (median 1 vs. 6 years in the Lap group) may underestimate the incidence of late complications. The results of this study are valid for comparing short-term postoperative outcomes, but conclusions regarding long-term late complications should be interpreted with caution. Our findings should be validated in long-term prospective multicenter studies with larger cohorts. Third, based on the status of robot introduction in 2021, the results may reflect initial experiences. Previous studies have reported that approximately 25–30 cases are required for the learning curve of robot-assisted CBD surgery to stabilize [10, 18]. During the study period, we were in the middle of the learning curve. At this stage, there may be variability in surgical time and some postoperative outcomes. While it cannot be ruled out that this may have influenced the results of this study, it is considered that it may provide important information for evaluating outcomes after CBD surgery. Fourth, this study did not perform multiple comparison corrections. The possibility of statistical significance being obtained by chance could not be ruled out. Therefore, the findings of this study should be verified by large-scale prospective studies and multicenter collaborative studies. Our results should be regarded as merely indicative of the direction for future research. Fifth, differences in experience among surgeons may have affected outcomes. In particular, robotic surgery was performed by a limited number of surgeons, so the results may have depended on the skills of the surgeons. However, we have adjusted for major confounding factors using multivariate analysis, demonstrating that the differences in outcomes between surgical procedures are independent of these factors. Future studies should involve more surgeons in a multicenter collaborative study to examine the effects of surgeons in greater detail.

Robot-assisted CBD surgery appears safe and is associated with faster short-term recovery, particularly in children. Given shorter follow-up and potential confounding, definitive long-term comparative effectiveness requires adjusted, time-to-event analyses and longer observation.