The following manuscript details a stepwise approach to the robot-assisted pancreaticoduodenectomy performed at the University of Pittsburgh Medical Center.
Since its first report in 2003, robotic pancreaticoduodenectomy (RPD) has gained popularity among pancreatic surgeons. Inherent advantages of the robotic platform, including three-dimensional vision, wristed instruments, and improved ergonomics, allow the surgeon to recapitulate the principles of open pancreatoduodenectomy allowing safe oncologic dissection, hemostasis, and meticulous reconstruction. Over the course of the past decade, significant strides have been achieved in outlining the safety, feasibility, and learning curve of the robotic Whipple. When performed by high volume pancreatic surgeons experienced in RPD, recent comparative effectiveness studies show potential advantages compared to the open technique, including reductions in hospital stay and morbidity. National data also show reductions in conversion rates compared to its laparoscopic counterpart. Although long-term oncologic data are still needed, short-term oncologic surrogates of margin resection and lymph node harvest suggest no compromise in oncologic outcomes. As pancreatic surgeons increasingly integrate robotics into their practice, proficiency-based training and credentialing will be necessary for the safe application and dissemination of RPD. Here, we provide the detailed steps of a robotic pancreaticoduodenectomy performed at the University of Pittsburgh Medical Center.
Pancreaticoduodenectomy (PD) is a complex operation that combines a challenging resection and a meticolous reconstruction. During its early inception, the traditional open approach was frought with high complication rates and a mortality rate approaching 25%. In the last three decades, improvements in the surgical technique and perioperative care led to corresponding improvements in outcomes, with a reduction in mortality to less than 5%, especially at high volume centers1,2,3. Despite this, morbidity remains substantial. With advancements in surgical technology, minimally invasive surgical approaches through laparoscopy or robot-assisted surgery have emerged in an effort to curb this morbidity. Since its first report in 2003, interest in robotic pancreaticoduodenectomy (RPD) has grown by pancreatic surgeons4,5. Inherent advantages of the robotic platform, including three-dimensional (3D) vision, wristed instruments, and improved ergonomics, allow the surgeon to recapitulate principles of open PD (OPD) in a minimally invasive manner, including safe oncologic dissection, hemostasis, and meticulous reconstruction4,6,7,8,9,10. The goal of this manuscript is to provide the detailed steps of an RPD performed at the University of Pittsburgh Medical Center (UPMC)11,12,13.
In the presented case study, a 42-year-old female with a previous history of intraductal papillary mucinous neoplasm (IPMN), initially presented with acute pancreatitis. Computed tomography (CT) of the abdomen revealed a 3.3 cm pancreatic head lesion with associated dilatation of the main pancreatic duct (Figure 1A,B), with a mixed type IPMN. Endoscopic ultrasound (EUS) confirmed the existence of an irregular, heterogenous cyst measuring 3.1 x 2.0 cm in the pancreatic head with mixed solid and cystic components and main PD duct dilation (Figure 1C). EUS cytology revealed the presence of atypical cells with no high-risk molecular mutations14,15. Biochemical workup including serum tumor markers were normal, with CA19-9 12 U/mL. Based on the Fukuoka criteria, this patient was recommended to have a PD and was deemed a suitable candidate for the robotic approach16.
This protocol follows the guidelines of the University of Pittsburg Medical Center human research ethics comittee (Institutional Review Board: PRO15040497)
1. Preoperative workup and selection
2. Anesthesia
3. Patient positioning
4. Placement of ports and liver retractor
5. Resection phase
6. Reconstruction phase
In the representative case, the total operative time was 225 min with an estimated blood loss (EBL) of 50 mL (Table 1). The patient was admitted to the surgical ward. Her postoperative course followed the UPMC institutional ERAS pathway. We routinely assess JP amylase at POD#1 and #3 to assess for pancreatic fistula and practice early drain removal on POD 3-5 when possible. The patient’s JP amylase levels were 403 U/L and 68 U/L, respectively. Therefore, the drain was removed on POD#3. The patient was discharged on POD#6.
Pathologic analysis of the specimen revealed invasive moderately-differentiated adenocarcinoma (0.2 cm) centered in the pancreatic head and arising in a branch-duct IPMN (3.7 cm) with extensive high-grade dysplasia with no positive lymph nodes in any of the 32 resected. There was no evidence of lymphatic, venous, or perineural invasion. Final AJCC 8th edition stage was pT1aN0M0. The patient was recommended to undergo adjuvant chemotherapy with FOLFIRINOX as per the PRODIGE 24 trial17. The patient completed the therapy and remains without any evidence of disease.
Figure 1: Preoperative diagnostic imaging. (A) and (B) IPMN in the head of the pancreas with associated main pancreatic ductal dilatation. (C) EUS demonstrating heterogenous pancreatic head mass with mixed solid and cystic components. Please click here to view a larger version of this figure.
Figure 2: Patient positioning and anesthesia setup. Patient is positioned supine in a split leg table with all pressure points padded. Patient table is positioned to accommodate for both the surgical robot and the anesthesia devices. This figure was reproduced with permission from Intuitive Surgical, Inc. Please click here to view a larger version of this figure.
Figure 3: Port placement. Purple 8 mm ports (robotic arms [A] 1–3), green 12 mm umbilical port (camera port), green 12 mm left lower quadrant port (assistant), red 5 mm right lower quadrant port (assistant), left lateral 5 mm port (liver retractor). This figure was adapted with permission from Springer, Journal of Gastrointestinal Surgery, Performing the Difficult Cholecystectomy Using Combined Endoscopic and Robotic Techniques: How I Do It. Magge, D. et al25. Please click here to view a larger version of this figure.
Clinicopathologic Treatment and Outcome Data, adopted from Zureikat, AH et al. Ann Surg. 2016. | |||||
Variable | All Patients | RPD | OPD | P-value | Represented Case |
Age | 65 | 67 | 65 | 0.07 | 44 |
Male sex, % | 52.90% | 55.45 | 52.26 | 0.41 | Female |
BMI, kg/m2 | 26.3 | 27.5 | 26.1 | <0.001 | 24.41 |
Prior abdominal surgery, % | 43.8 | 51.18 | 41.86 | <0.001 | None |
Pancreatic cancer, % | 50.8 | 33.18 | 55.32 | <0.001 | Yes |
Pancreatic duct diameter (>8mm), % | 6.3 | 15.74 | 3.55 | <0.001 | 1 mm |
Pancreatic texture (Soft), % | 49.2 | 69.43 | 43.35 | <0.001 | Soft |
Operative time, min | 325 | 402 | 300 | <0.001 | 225 |
Estimated blood loss | 300 | 200 | 300 | <0.001 | 50 |
Transfusion, % | 16.4 | 16.11 | 16.52 | 0.89 | None |
Major complications, % | 23.8 | 23.7 | 23.87 | 0.96 | None |
Severe wound infection, % | 13 | 11.37 | 13.41 | 0.43 | None |
Pancreatic Fistula (Grade B/C), % | 23.8 | 13.7 | 9.1 | 0.04 | None |
Length of stay, days | 8 | 8 | 8 | 0.98 | 6 |
Table 1: Comparison of the represented case with national data9.
With advances in the surgical technology, laparoscopic and robot-assisted surgeries are being increasingly used in gastrointestinal and hepatobiliary procedures. Conventional laparoscopy is associated with benefits over open surgery for many procedures. However, inherent limitations such as decreased surgical dexterity, suboptimal ergonomics, lack of wristed instruments, and 2-D visualization, have limited its dissemination to complex gastrointestinal operations such as PD.
Contrary to laparoscopy, the robotic platform allows for the minimally invasive operations to be performed under 3D vision, with enhanced dexterity and the use of articulating (wristed) instruments. The Si is an older system and is the basis for which the authors have performed the vast majority of RPDs. The main inherent advantage of the older model (e.g., Si) is the use of a larger (12 mm) robotic camera with improved definition over the 8 mm camera (e.g., Xi). However, in this case both the newer and the older versions are used interchangeably for RPD. Regardless of the model, RAS allows for the open PD principles to be adhered to when performing the minimally invasive surgery. Despite concerns over oncologic outcomes, morbidity, cost, and training, several single, multi-institutional, and national series have demonstrated the safety and feasibility of RPD5,7,8,15. More recent data demonstrate that RPD can be associated with improvements in morbidity and length of stay compared to the open approach and reductions in conversion compared to the laparoscopic approach9,18,19,20,21.
Based on our experience at UPMC, several factors are needed for successful implementation of RPD. These include an institutional commitment to program success with necessary training and mentorship, prior surgeon experience in open pancreatic surgery, use of two staff surgeons to navigate through the initial learning, availability of a large case volume (2–4 cases/month), prospective assessment of perioperative outcomes, and dedicated operating room staff.
Data from our experience suggests the learning curve of RPD to be approximately 80 cases22. Notably, this is quite similar to the learning curve of OPD as demonstrated by three other reports.1,23,24 Reductions in EBL and operative conversions occur early (20 cases), while a decrease in clinically relevant pancreatic fistula rate occurs after 40 cases. Operative time, a surrogate of procedural efficiency, is optimized after 80 cases. Following identification of our learning curve, we established a training program with the objective of disseminating safe robotic pancreatectomy. This stepwise mastery-based curriculum includes five main components: 1) mastery of the console, 2) virtual reality, 3) innanimate and bio-tissue drills, 4) live operative proctoring, and 5) continuous quality improvement and assurance11,13,25.
There are a few technical considerations for RPD that warrant emphasis. During the operation, communication between the bedside and console surgeons is paramount. Both surgeons must adhere to the same operative plan and anticipate each other’s maneuvers. In the resection phase, A3 plays a key role in retraction of the specimen to allow for optimal exposure. There are three critical parts in the operation that can result in significant intraoperative hemorrhage: 1) dissection of LOT and linearization of the duodenum after the proximal jejunal transection, 2) dissection of the inferior pancreatic border to begin the retropancreatic tunnel, and 3) dissection of the uncinate process. These phases demand extreme caution and warrant a thorough knowledge of operative anatomy. Control of hemorrhage may be challenging and requires a fascile ability to suture with 4–0 and 5–0 monoflament sutures, an experienced bedside assistant to control suction, and the ability to perform a quick and safe open conversion if bleeding is not controlled. In the reconstruction phase, A3 similarly plays a major role, since it is often utilized to grasp and retract previously placed sutures in a cranial direction to allow for countertension when placing sutures.
In conclusion, we provide a stepwise description of our RPD technique. Our technique follows principles of open PD, while allowing safe and oncologically sound application of a minimally invasive approach to this complex operation.
The authors have nothing to disclose.
Nothing to acknowledge.
3-0 V-Loc sutures | Medtronic (Minneapolis, MN) | VLOCMo614 | Barbed Absorable Suture |
4-5 Fr Freeman Pancreatic Flexi-Stent | Hobbs Medical (Stafford Springs, CT) | 6542, 6552 | Pancreatic Duct Stent |
5-0 PDS (polydiosxanone) | Ethicon (Somerville, NJ) | D10063 | Synthetic Absorbable Suture |
Cadíere forceps | Intuitive (Sunnyvale, CA) | 470049 | Surgical Robot Instrument |
Da Vinci Si | Intuitive (Sunnyvale, CA) | Surgical Robot | |
Da Vinci Xi | Intuitive (Sunnyvale, CA) | Surgical Robot | |
Endo Clip 10 mm Applier | Covidien (Dublin, Ireland) | 176619 | Laparoscopic Titanium Clip Applier |
Endo GIA 45 mm Curved Tip Articulating Vascular Stapler with Tri-Stapler Technology | Covidien (Dublin, Ireland) | EGIA45CTAVM | Laparoscopic Surgical Stapler |
Endo GIA 60 mm Articulating Stapler with Tri-Stapler Technology | Covidien (Dublin, Ireland) | EGIA60AMT | Laparoscopic Surgical Stapler |
Endo GIA 60 mm Curved Tip Articulating Vascular Stapler with Tri-Stapler Technology | Covidien (Dublin, Ireland) | EGIA60CTAVM | Laparoscopic Surgical Stapler |
EndoCatch Gold 10 mm Specimen Pouch | Medtronic (Minneapolis, MN) | 173050G | Specimen Extraction Bag |
EndoCatch II 15 mm Specimen Pouch | Medtronic (Minneapolis, MN) | 173049 | Specimen Extraction Bag |
Fenestrated bipolar forceps | Intuitive (Sunnyvale, CA) | 470205 | Surgical Robot Instrument |
GelPOINT Mini Advanced Access Platform | Applied Medical (Rancho Santa Margarita, CA) | CNGL3 | Laparoscopic Abdominal Access Platform |
Large needle driver | Intuitive (Sunnyvale, CA) | 470006 | Surgical Robot Instrument |
Large SutureCut needle driver | Intuitive (Sunnyvale, CA) | 470296 | Surgical Robot Instrument |
LigaSure Blunt Tip Laparoscopic Sealer/Divider | Medtronic (Minneapolis, MN) | LF1844 | Laparoscopic Bioplar Device |
Mediflex liver retractor | Mediflex (Islandia NY) | Laparoscopic Liver Retractor | |
Monopolar curved scissors | Intuitive (Sunnyvale, CA) | 470179 | Surgical Robot Instrument |
Permanent cautery hook | Intuitive (Sunnyvale, CA) | 470183 | Surgical Robot Instrument |
ProGrasp forceps | Intuitive (Sunnyvale, CA) | 470093 | Surgical Robot Instrument |