The first published report of a successful pancreaticoduodenectomy was published by Allen O. Whipple in 1935 [1]. Whipple reported 3 patients who underwent a 2-stage procedure with pancreatic duct ligation: one patient died in the perioperative period; another died 8 months later from cholangitis, and the last from metastases after 28 months. This initial report was followed by a series describing a single-stage procedure [2], the fundamentals of which we recognize today as the Whipple procedure. These fundamentals included (1) resection and reconstruction in one stage; (2) avoidance of cholecystoenterostomy by implantation of the bile duct into the jejunum, and (3) implantation of the pancreatic duct into the jejunum. Following the modification of pylorus preservation by Traverso and Longmire [3], the technical aspects of pancreaticoduodenectomy have remained essentially unchanged since Whipple described the procedure in 1935.

Postoperative mortality and morbidity remained significant hurdles to the widespread implementation of pancreaticoduodenectomy for many years after the initial description. Mortality rates approaching 30% were common for the next several decades. In 1968 John Howard reported a series of 41 consecutive pancreaticoduodenectomies without a mortality [4]. This series was followed by improvements by John Cameron and colleagues [5] at Johns Hopkins, who standardized the technical aspects of this procedure and set the modern-day gold standard for outcomes. Attention to surgical detail combined with advances in critical care and anesthesia led to steady and dramatic improvements in postoperative outcomes following pancreaticoduodenectomy. Recent reports reproducibly demonstrate a morbidity rate of 30% to 40% with 1% to 3% mortality [6].

Recent refinements of the pancreaticoduodenectomy have focused on the implementation of minimally invasive approaches. Gagner and Pomp [7] described the first laparoscopic pancreaticoduodenectomy in 1994, a procedure that lasted nearly 24 hours. Since then, Kendrick and Cusati [8] and Palenivelu and colleagues [9] have reported large series of minimally invasive pancreaticoduodenectomies with outcomes comparable with those of large open series. Most recently, robotic-assisted minimally invasive approaches have been described by the groups led by Gulianotti, Melvin, Zeh, and Moser [10–12].

Major pancreatic resection remains the final frontier of minimally invasive surgery, because of the twin technical challenges of controlling hemorrhage from major vessels and reconstructing the biliary and pancreatic ducts with acceptable morbidity (Box 1). The minimally invasive approach offers potential advantages compared with open surgery: (1) decreased incisional pain may lead to improved recovery time and decreased hospital stay; (2) improved postoperative recuperation and performance status may permit earlier initiation of adjuvant therapy in a higher percentage of patients with pancreatic cancer [13]. It is important that initial concerns regarding the oncologic equivalency of minimally invasive resection for cancer have proved to be unfounded in other malignancies such as colon and gastric cancer [12,14–16]. The third potential advantage of minimally invasive pancreatic resection applies to the group of patients with radiographically identifiable precursor lesions such as mucinous cystic neoplasms who may require prophylactic pancreatectomies to prevent the progression to pancreatic cancer. The availability of a minimally invasive approach with equivalent or superior recovery times might alter the risk/benefit ratio of pancreatectomy in favor of earlier intervention and improve patient acceptance of prophylactic surgery. Lastly, the technological progression in all of surgery is toward smaller more minimally invasive procedures. Reluctance or refusal on the part of hepatic/pancreatic/biliary tract (HPB) surgeons to explore innovations risks obsolescence.

Laparoscopic surgery has evolved significantly since its introduction in the early 1970s. Although advanced laparoscopic procedures are being performed at many centers, advanced procedures that require complicated resection and reconstruction such as pancreaticoduodenectomy (PD) remain limited to a few specialized centers. A total of only 146 laparoscopic PDs were reported in the world’s literature in the first 14 years following Gagner’s description in 1994 [17]. Palanivelu and colleagues [9] presented 75 cases, and Kendrick and Cusati [8] reported 62 cases of totally laparoscopic PDs. These two series demonstrate that laparoscopic pancreaticoduodenectomy can be performed safely with acceptable morbidity, although their results may be difficult to generalize to other centers [18]. The slow implementation of laparoscopic techniques for pancreaticoduodenectomy is likely the result of the limitations inherent to current technology, namely, 2-dimensional imaging, limited range of instrument motion, and poor surgeon ergonomics [17]. In this situation the surgical principles are altered to meet the limitations of the technology, leading to reluctance on the part of many HPB surgeons. A minimally invasive approach to pancreaticoduodenectomy that recreates well-established surgical principles would be a significant advance.

Robotic-assisted minimally invasive surgery overcomes many of the shortcomings of laparoscopy, with improved 3-dimensional imaging, 540° movement of surgical instruments, and improved surgeon comfort and precision [19] (Box 2). These technological innovations allow complex resections and anastomotic reconstructions to be performed with techniques identical to open surgery. The authors present here their technical description and outcomes with robotic-assisted major pancreatic resections. This approach maintains maximal adherence to the traditional open surgical techniques with a minimally invasive approach.

To maintain safety and transparency of surgical outcomes, all potential candidates for robotic pancreatic resection are reviewed by the Surgical Oncology Robotic Selection Committee. All robotic procedures are performed by two expert pancreatic surgeons familiar with open pancreaticoduodenectomy and capable of carrying out venous resection and reconstruction whenever indicated. Patients with periampullary malignancies who are candidates for robotic pancreaticoduodenectomy undergo individualized treatment planning based on a validated predictive model to select candidates with the highest likelihood of achieving an R0 surgical resection. The prediction rule was developed and validated in independent cohorts of patients with potentially resectable pancreatic cancer [20]. The model stratifies patients into low risk and high risk for non-R0 surgical outcomes based on findings during preoperative computed tomography (CT) and endoscopic ultrasonography (EUS). High-risk patients are not offered the robotic approach and instead undergo traditional open pancreaticoduodenectomy, given the potential for robotic surgery to compromise oncologic principles in these high-risk patients. Low-risk patients are offered robotic surgery after a detailed consent process and enrollment in a prospective registry of robotic pancreatic surgery.

The predictive factors are: (1) any evidence of arterial or venous vascular involvement on CT; (2) the combination of EUS T-stage and N-stage data to assign a preoperative stage according to the criteria of the American Joint Committee on Cancer (sixth edition), and largest EUS tumor dimension greater than 2.6 cm. Evidence for vascular involvement by CT scan includes minimal abutment of the superior mesenteric or hepatic arteries without extension to the celiac axis, as well as any preoperative suspicion that tumor involves the superior mesenteric vein (SMV)-portal vein (PV) confluence despite the possibility of venous resection and reconstruction. The prediction rule classifies operative findings of metastatic or locally advanced disease as well as positive resection margins as treatment failures. A patient is considered a good candidate for R0 resection (low risk) and should undergo surgery as primary therapy if: (a) the EUS stage is 1A; (b) if there is no vascular involvement, and the EUS stage is greater than 1A and less than 3; or (c) if there is no vascular involvement and EUS stage 2B but the largest tumor dimension is less than 2.6 cm. Otherwise, a patient is a poor candidate for R0 resection (high risk).

In the authors’ published report, the overall resection rate (R0 + R1) among low-risk patients was significantly greater (92%) than that of the high-risk group (53%; P<.0002). Low-risk patients achieved R0 status more frequently than high-risk patients (73% vs 33% R0, P = .0009), despite resection and reconstruction of the PV whenever indicated in both groups. Additional operative findings distinguishing the two risk groups included a greater proportion of unresectable, locally advanced tumors (17% vs 0%, P = .007) as well as unexpected metastatic disease (30% vs 8%, P = .026) in the high-risk group. High predicted risk of surgical failure corresponded to more advanced stages of disease on final surgical pathology, and also correlated with shorter postoperative overall survival. Median survival of low-risk patients was 20.3 months, compared with 12.1 months in those considered at high risk (P = .02).

Robotic-assisted minimally invasive resection of the pancreatic head recreates published methods for open pancreaticoduodenectomy. The technique emphasizes teamwork between two experienced pancreatic surgeons and requires 4-handed cooperation to retract and expose critical structures, the anatomy of which may be distorted by tumor, body habitus, and pancreatitis. The importance of teamwork cannot be overemphasized. Exposure and safe control of bleeding from major vascular structures requires two surgeons familiar with the anatomy to develop a skilled collaboration and the mutual ability to anticipate each other’s movements.