Introduction 1 2 3 4 5 6 3 7 8 11 12 Patient Selection and Technique Contrast-enhanced computed tomography (CT) is used to evaluate a pancreatic head mass and its relation to vascular structures. CT accurately diagnoses mesenteric vein involvement, aiding in operative planning. The CT is also used to assess the length and caliber of the left renal vein, the status of the kidneys bilaterally, and the presence of the left gonadal and adrenal veins, which serve as collateral venous drainage. Additional imaging of the pancreas with endoscopic ultrasound (EUS) is frequently used to further evaluate the location and extent of any venous involvement. 1 1 Figure 1 a b  The superior border of the pancreas is approached by incising the gastrohepatic ligament. The right gastric and gastroduodenal arteries are routinely ligated and divided. Retraction of the common hepatic artery cephalad allows dissection of the portal vein (PV), thereby, allowing completion of the plane between the pancreas and the PV-SMV. The gallbladder is dissected from the liver, and the hepatic duct is encircled near the cystic duct junction. If a pylorus-preserving PD is planned, the superior and inferior aspects of the duodenum are skeletonized, individually ligating the right gastroepiploic vessels. The ligament of Treitz is mobilized, and the proximal jejunum is divided. Sequential ligation and division of the bowel mesentery exposes the uncinate process. The mobilized duodenum and jejunum are passed beneath the superior mesenteric vessels into the right upper quadrant. 2 2 Figure 2 a b  3 3 Figure 3 a b  Mobilization of the portal vein superior to the pancreas and the peritoneum along the root of the small bowel mesentery may provide length for the SMV or PV segment. This is accomplished by ligating and dividing small branches to the SMV, PV, and splenic vein (SV). Although primary end-to-end anastomosis is preferred, if interposition grafting is necessary, autologous vein and specifically the left renal vein is utilized for two reasons. First, the vein may be exposed within the same operative field, thereby, eliminating a second operative field and dissection. Second, the caliber and wall thickness of the vein is similar to the portal vein in most instances, providing good handling and suturing properties. Harvest of the left renal vein is undertaken after the retroperitoneal dissection when the specimen remains attached to the portal vein segment only (before venous resection). This allows the best assessment of the need for interposition grafting and minimizes the amount of clamp time by harvesting the graft before SMA and venous occlusion. 4 4 4 Figure 4 a b c  5 6 7 Figure 5 Vascular clamps are used to control the superior mesenteric vein, splenic vein, and portal vein before sharp dissection and resection of the involved venous segment. Inflow occlusion of the superior mesenteric artery during reconstruction reduces bowel engorgement.  Figure 6 posterior view  Figure 7 In the example shown, reconstruction of the PV-SMV confluence with an interposition graft utilizing left renal vein was performed initially, followed by reimplantation of the splenic vein into the graft (end-to-side) Gastrointestinal reconstruction is performed in a standard fashion after completion of venous reconstruction.  Results Nine patients have undergone reconstruction of the SMV-PV during PD with an autologous left renal vein graft. There were seven men and two women with a mean age of 57 years (range, 31–77). Preoperative abdominal CT had suggested mesenteric vein involvement in seven of the nine patients. EUS was completed in three patients. In one patient, EUS suggested there was no vein involvement, while in the remaining two patients, EUS did suggest involvement. Preoperative serum creatinine levels ranged from 0.8 to 1.1 mg/dl in these patients (normal values 0.9–1.4 mg/dl ). Mean follow-up was 6.8 months. The procedure consisted of three standard PD and six pylorus-preserving PD. Venous reconstruction consisted of eight interposition grafts and one patch graft. The patch graft was located on the lateral edge of the SMV and PV. Five of the interposition grafts were placed in the SMV, inferior to the confluence. One interposition graft was placed between the SMV and PV with reimplantation of the splenic vein; an additional was placed between the SMV and PV without reimplantation of the splenic vein, and the final graft was in the portal vein. The mean operating time was 7.8 hours (range, 6.5–9.5). The mean tumor size was 3.4 cm (range, 2.2–5). The mean estimated blood loss was 1,300 ml (range, 350–2,500). Eight patients were found to have node-positive disease with six of these patients noted to have histological involvement of the venous segment, while one additional patient had pathologically negative lymph nodes and no evidence of malignant invasion of the vein. In two patients, the uncinate margin was microscopically positive. One patient was monitored overnight in the intensive care unit. There were no operative mortalities, and reoperation was not required in any of the patients. The mean length of hospitalization was 14 days (range, 9–29). Immediate perioperative morbidity included a superficial wound infection in one patient, delayed gastric emptying in one patient, and postoperative gastrointestinal bleeding in one patient. None of the patients experienced a pancreatic leak. No hematuria was noted. One patient was diagnosed with ascites and stenosis of the left renal vein interposition graft anastomosis 1 month after the operation. This patient had a congenitally cystic (nonfunctioning) left kidney. The left renal vein was reduced in caliber, but felt to be adequate for grafting at the time of the original operation. The patient underwent stenting of the graft by interventional radiology with resolution of her symptoms. Eight patients underwent adjuvant treatment, which included radiation therapy in six patients. None of the six patients receiving radiation therapy experienced a decrement in renal function after radiation therapy. Two patients had died 8.3 and 18.2 months after the operation of recurrent disease. Median survival has not been reached. 1 Table 1 A Comparison of Creatinine Levels Serum Creatinine Concentrations (mg/dL) Patient Preoperative Peak Time of Discharge 1 1.1 1.1 1.1 2 1.1 1.4 1.1 3 1 1.5 1 4 0.9 1.3 1.1 5 1 1.3 1 6 0.9 1.3 1.3 7 0.8 0.9 0.9 8 0.8 1.1 0.9 9 0.8 0.8 0.7 Discussion 3 4 13 3 14 9 12 Conclusions Resection offers the only chance at cure for patients with pancreatic cancer, and potentially curative resection may require venous resection. When reconstruction of the venous system necessitates the use of interposition grafting, autologous vein interposition grafts are preferred. The left renal vein provides an additional choice for an autologous graft, and its use is distinguished by ease of harvest and maintenance of renal function. The use of the left renal vein for interposition grafting and patch repair should be considered by surgeons experienced in SMV-PV reconstruction during PD.