Robotic whipple procedure

Introduction

The first minimally invasive pancreaticoduodenectomy (PD) was reported in 1994, followed by the first published description of a robot-assisted PD in 2003. ^, The robot-assisted approach to PD offers surgeons many advantages compared with standard laparoscopy, and its use has led to increased rates of minimally invasive PD. Multi-institutional studies have since confirmed the safety and efficacy of robot-assisted PD. Robot-assisted PD has been consistently associated with longer operative times but is associated with less blood loss and fewer major postoperative complications. Importantly, in pooled data from surgeons past the initial learning curve, there is no difference in margin status or suboptimal lymphadenectomy.

Indications

The indications for robot-assisted PD should mirror those for open PD. Despite this similarity, there appears to remain a selection bias from published reports, with only 33% of robot-assisted PD being performed for pancreatic adenocarcinoma, compared to 55% in one large, multi-institutional study for open PD. Early concerns regarding the lack of oncologic equivalence of the minimally invasive approach have been unfounded, as multi-institutional studies and meta-analysis showed similar outcomes. This oncologic equivalence was strengthened when studies evaluated outcomes from surgeons who were past the initial learning curve. Although satisfactory oncologic outcomes can be achieved, there are some technical limitations for complex resections that should be acknowledged by most surgeons. Common exclusions to performing the procedure should include patients with very large or bulky tumors and those with locally advanced disease. While portal vein involvement has traditionally been an indication for an open procedure, reports on venous resection and reconstruction have emerged from high-volume centers. However, this should be reserved for specialty centers with significant experience in robot-assisted PD.

The learning curve for robot-assisted PD has been reported to range between 20 and 80 cases, depending on the metrics used. In examining console time as the primary measure, as few as 20 cases were needed to plateau for surgeons already experienced with open PD and robot-assisted distal pancreatectomy. A series from the University of Pittsburgh showed an improvement in blood loss and conversions at 20 cases, a decrease in pancreatic fistula after 40 cases, and a reduction of operative time after 80 cases. Other measures such as complication rates, length of stay, and readmission rates continued to show improvement over the 120 cases reported in the series. Importantly, in considering the learning curve, one must recognize that this reflects not only the increasing technical competency of the operating surgeon but also the progression of the assistants and the entire surgical and anesthetic team.

Patient selection is critical for those surgeons who are early in their learning curve of robot-assisted PD. Early cases should focus on distal cholangiocarcinomas, ampullary carcinomas, and smaller pancreatic malignancies. These cases should present a straightforward resection as well as a dilated pancreatic duct, making reconstruction easier. Likewise, small pancreatic endocrine tumors and cystic neoplasms can provide for relatively easy resections but often at the expense of a reconstruction made more difficult by smaller pancreatic and/or bile ducts. With experience, more complex cases will become achievable in a minimally invasive fashion with the robot.

Patient preparation

An important part of patient preparation is education. Our patients attend procedure-specific, preoperative educational classes not only detailing the operative experience but also providing a framework for postoperative recovery. These preoperative classes help to establish patient expectations and are suggested to reduce anxiety and facilitate timely discharge.

Operating room setup and patient positioning

Patients receive indocyanine green in the preoperative holding area to assist with identification of the biliary tree, utilizing the near-field florescence technology available on the da Vinci robotic system (Intuitive Surgical, Inc., Sunnyvale, CA). Preoperative antibiotics and deep vein thrombosis prophylaxis are administered and re-dosed as appropriate, per typical surgical guidelines. Sequential compression devices are applied prior to the induction of anesthesia.

Following the induction of anesthesia, a Foley catheter, arterial line, and nasogastric tube are placed. These items are usually left overnight following surgery and are removed on postoperative day 1. Additional intravenous access is obtained at the discretion of the anesthesia team. Central venous catheters are used as dictated by comorbidities or difficulty with intravenous peripheral access. Fluid management is guided throughout the procedure by monitoring stroke volume variation (SVV) using the Vigileo device (Edwards Lifesciences, Irvine, CA). A previous report showed that restrictive fluid management guided by SVV during the resection portion of the procedure was associated with lower rates of postoperative pancreatic leak and delayed gastric emptying.

Patient positioning

We presently perform all robot-assisted PD procedures on the da Vinci Xi system, allowing for more freedom in patient position and operative room arrangement than previous models (standard and Si). Patients are placed in a supine position on the operative table with arms extended. The operative table is oriented at a 30–45-degree angle from the anesthesia personnel. The side-docking ability of the da Vinci Xi and this operative room layout provide ample clearance for the robot and ergonomic working space for the bedside assistant, while simultaneously giving the anesthesia personnel continued access to the patient’s airway and extended arms.

After ensuring that the patient has appropriate padding and is securely positioned on the operating room table, the patient is placed in a slight reverse Trendelenburg position, usually ranging from 8–12 degrees, depending on the body habitus of the patient. Patient positioning in a steep reverse Trendelenburg position is typically unnecessary. The patient is prepped and draped in the standard surgical fashion. We have found that the use of trauma drapes with attached pouches is beneficial in organizing the necessary cables and tubing.

Port placement

Entry method can be achieved using a Veress needle, Hasson, or optical entry at the discretion of the operating surgeon and taking into consideration the patient’s body habitus and prior procedures. We typically initiate insufflation and entry into the abdomen at the infra-umbilical position using a Veress needle followed by a 12-mm assistant port. The robotic ports are then placed in symmetrical spacing along a horizontal line above the umbilicus. The ideal port placement is shown in Fig. 31.1 . The general locations for the ports include the right anterior axillary, right midclavicular, left midclavicular, and left anterior axillary lines. The four robotic ports may need to be adjusted cranially or caudally based on the patient’s body habitus. Additionally, large, protuberant, or small abdomens may require adjusting the spacing between the ports. Ensuring adequate spacing between the robotic arms on patients with small abdomens can prove difficult, and often the ports need to be placed below the level of the umbilicus for extremely small patients.

Additional minor modifications to spacing may be required, depending on the surgeon’s preference for having two arms to the right or left of the camera. Our standard technique places the camera in the right midclavicular line with two robotic instruments controlled by the right hand of the surgeon. Many other reports describe a preference for two left-sided instruments. This will be a personal preference acquired with time and experience. Likewise, the placement of assistant ports will be dictated by the final location of the robotic ports and by body habitus. In our common layout, the 12-mm infra-umbilical port placed at the time of entry is utilized as the assistant port as well as the specimen extraction site. If this site is not ergonomically satisfactory for the assistant or is impeded by body habitus, an additional 5-mm assistant port can be placed away from the remaining ports. Some surgeons report the regular use of a dedicated liver retractor inserted through an additional left upper quadrant port or a Nathanson retractor placed in a subxiphoid position; however, this is not routinely part of our practice.

The robot is usually docked from the patient’s right side, but it could alternatively be docked from the left side or overhead. The camera is placed in Arm #2 (right midclavicular line) and targeting is performed with the umbilical fissure of the liver as the primary target. This allows for ample visualization and utility of the arms throughout the resection phase as well as the range required for reconstruction.

All robot-assisted PD procedures performed at our institution are single-surgeon cases with a dedicated, highly skilled surgical technologist. The availability of dedicated surgical technologists and circulating nurses has streamlined our robotic pancreatectomy program from a logistical and financial standpoint.

Key operative steps