Introduction

Portal annular pancreas (PAP) is an anatomical anomaly in which the uncinate process of the pancreas extends extensively behind the superior mesenteric vein (SMV) and/or portal vein (PV), and then fuses with the dorsal surface of the pancreatic body [1]. Pancreatic tissue annularly surrounds the SMV–PV and requires careful consideration during surgical planning for pancreatic resection. A case of PAP is herein presented and the technical issues are discussed.

Case report

A 61-year-old female with PAP was referred to the hospital for pylorus-preserving pancreaticoduodenectomy for papilla Vater adenocarcinoma. Preoperative computed tomography (CT) showed that the PV was annularly surrounded by pancreatic parenchyma and PAP was diagnosed (Fig. 1a). Endoscopic retrograde pancreatography (ERP) demonstrated the inferior head branch of the pancreatic duct (IHBPD) in the uncinate process (Fig. 1b). The connection between IHBPD and the main pancreatic duct (MPD) could not be identified in the pancreatic body. Magnetic resonance cholangiopancreatography (MRCP) revealed the same findings (Fig. 1c). The intraoperative findings were compatible with CT. The uncinate process extended extensively behind the PV, and then fused with the dorsal surface of the pancreatic body (Fig. 2). The pancreas was divided above the PV in order to liberate the PV from pancreatic annulation, and then the body of the pancreas was divided again (Fig. 3). Reconstruction was performed by a modification of the method described by Child with pancreatojejunal anastomosis performed by duct-to-jejunum, end-to-side pancreaticojejunostomy. [2] Pancreatic duct-to-jejunal anastomosis was performed with 9 interrupted sutures using monofilament slowly absorbable material (5-0 Maxon, Covidien Co.). The pancreatic stump and jejunal seromuscular layer were closely approximated with 5 interrupted sutures using monofilament non-absorbable material (4-0 Nespylene, Alfresa Pharma Co.) as described by Kakita et al. [3]. A 7 French (Fr) polyethylene tube was placed in the MPD as a lost stent. Postoperative pancreatography of the specimen demonstrated extension of IHBPD in the uncinate process without connection to the MPD in the pancreatic body (Fig. 4). The fusion between the uncinate process and the dorsal surface of the pancreatic body was histologically examined. There was connective tissue between the extended uncinate process and pancreatic body, and no ductal continuity was identified between the IHBPD in the uncinate process and the MPD in the pancreatic body (Fig. 5). The postoperative course was uneventful.

Fig. 1
figure 1

Preoperative findings of a portal annular pancreas by computed tomography and endoscopic retrograde pancreatography. CT shows the portal vein (arrow) is surrounded annularly by pancreatic parenchyma (arrow heads; Fig. 1a). Endoscopic retrograde pancreatography and MRCP shows that the inferior head branched pancreatic duct (arrow) was visualized as well as the main pancreatic duct (arrow head; Fig. 1b, c). The common hepatic duct (asterisk) and duct of Santorini (filled diamond) were also visualized

Full size image
Fig. 2
figure 2

Intraoperative findings of the portal annular pancreas. The pancreatic head (H) was divided from the pancreatic body (B) with a stapling device above the portal vein (asterisk). The uncinate process (arrow) extended behind the PV and fused with the dorsal surface of the pancreatic body. The fusion was superior to the splenic vein (dagger symbol) The pancreatic body was divided again along the dotted line

Full size image
Fig. 3
figure 3

Schematic illustration of the planes of pancreatic division. The pancreas was first divided above the portal vein (arrow head) and then subsequently divided again to the left of the fused area between the uncinate process and pancreatic body (arrow)

Full size image
Fig. 4
figure 4

Postoperative pancreatography of the specimen. The pancreas was first divided above the SMV–PV with a stapling device (inverted triangles). The pancreas was divided again in the body (diamond). The inferior head branched pancreatic duct (arrow) extended in the uncinate process stretching to the dorsal surface of the pancreatic body. The duct of Santorini (asterisk) was also visualized

Full size image
Fig. 5
figure 5

Histological examination of the fusion between the uncinate process and the dorsal surface of the pancreatic body (magnified with a loupe). There were connective tissues (C) between the uncinate process (U) and the pancreatic body (B). No ductal continuity was identified between the main pancreatic duct in the body (asterisk) and the branched duct in the uncinate process

Full size image

Discussion

PAP is a rare anatomical anomaly with a reported incidence of only 1.14 % [1]. PAP is classified as suprasplenic, infrasplenic and the mixed type depending upon where the uncinate process fuses in relation to the splenic vein [4]. PAP requires division of the pancreatic parenchyma at least twice to liberate the SMV–PV. The present case was the suprasplenic type and required pancreatic division twice (Fig. 4), whereas it is necessary to divide the pancreas three times in patients with the mixed type. The uncinate process is derived from the ventral pancreatic bud and contains its own pancreatic duct [5]. Therefore, a retroportal pancreatic duct is invariably present in PAP to drain the uncinate process with or without ductal continuity between its branch and the MPD in the pancreatic body. A retroportal MPD develops if the IHBPD connects to MPD in the body [4, 6, 7]. The current case clearly demonstrated the presence of IHBPD in the extended uncinate process on both preoperative ERP and postoperative pancreatography, but there was no ductal continuity with the MPD in the pancreatic body either radiologically or histologically. Dividing the pancreas anywhere in the uncinate process would have produced two independent pancreatic dissecting planes, one with the MPD and the other with an isolated IHBPD, and such a procedure would have increased the risk of pancreatic fistula. In fact, previous reports of pancreaticoduodenectomy for PAP described a prolonged postoperative course due to the development of a pancreatic fistula [1, 6, 8–10]. One solution to prevent pancreatic fistula might have been to ligate the retroportal pancreatic duct [4, 7], but two independent pancreatic dissecting planes would have still remained. Therefore, it was safer to dissect the whole uncinate process to produce only one dissecting plane with the MPD in the pancreatic body. This procedure produces only one dissecting plane with the MPD even in the case of a retroportal MPD. The current case did not histologically demonstrate the fusion and ductal continuity between the uncinate process and the dorsal surface of pancreatic body. It might have been possible that there were only physiological adhesions between the uncinate process and pancreatic body. Only one pancreatic division above the SMV–PV might have been sufficient in that case, with separation of the physiological adhesions in order to produce only one dissecting plane with the MPD in the pancreatic body. However, lysis of tight adhesions between the pancreatic body and uncinate process carries some risk of pancreatic fistula from the raw pancreatic surface once it is injured during separation. Therefore, the pancreas was first divided above PV in the present case, and the final division was performed to the left of the fusion between the uncinate process and pancreatic body (Fig. 3). In conclusion, PAP should be treated by dividing the pancreatic body after liberation of the SMV–PV on the left of the fusion between the uncinate process and pancreatic body to reduce the risk of a pancreatic fistula after pancreaticoduodenectomy.