European Journal of Cardiovascular Medicine

Since 1935 [1] Pancreaticoduodenectomy (PD) is performed for various malignant and benign conditions .Traditionally considered a demanding task, the duodenum-pancreas deep location in the retroperitoneum and connections with major vessels increase technical requirements. While depth may be easily predicted from experience and, in spite of intensifying the technical difficulty, is rarely related to specific complications, anatomical variations in duodenopancreatic arterial pattern (celiac axis and the superiormesenteric artery - SMA) may carry a risk of potentially life-threatening vascular injury when not recognized before the pro-cedure by appropriate preoperative imaging.

Indeed, the anatomical variation of the common hepatic artery (CHA) originating as a branch of the SMA is rare but not in significant, reaching 1.5e4% [2,3]. In this conformation ,the CHA is located behind the head of pancreas, relatively far from the right aspect of the portal vein whereas the gastroduodenal artery (GDA) arises rom CHA. During PD, if this situation is not recognized, CHA injury Or interruption can easily occur, leading to massive hepatic necrosis or biliary ischemic cholangitis [4,5].

This technical report aims to share our experience in two consecutive cases of patients undergoing PD and concomitantly presenting with an anatomical variation of CHA, originating as a branch of the SMA.

In March 2012 and August 2012, we recorded two consecutive patients undergoing PD and presenting with a CHA originating as a branch of the SMA (Fig. 1).

Patients were a 51-year-old man and a 71-year-old woman without personal history for gastrointestinal affection and major comorbidity. Both were diagnosed with adenocarcinoma of the head of pancreas. Preoperative investigations did not reveal  distant metastasis, peritoneal carcinomatosis or any other surgical contraindication. Contrast enhanced CT-scan showed the rare anatomical arterial pattern (Fig. 2). After presentations during multidisciplinary meeting, PD was proposed.

Procedures were realized using the same modified technique, through a midline incision. The great omentum detachment enabled to reach the retrogastric cavity. In order to complete the stomach mobilization, the lesser sac was opened. Starting with an hepatic flexure mobilization, a Kocher manoeuvre was performed in order to fully mobilize the duodenopancreatic bloc. Then, the mobilization of the third and fourth duodenum and of the first jejunal loop was achieved. An anterograde cholecystectomy was completed, enabling to identify the common bile duct, the intra-pedicular CHA and the portal vein in the hepatic pedicle. At the lower border of the pancreas, the mesenterico-portal axis was identified and dissected. Using mechanical stapler, the stomach was split in order to achieve the antrectomy. The anterior aspect of the portal vein was dissected off the posterior aspect of the pancreatic neck.

Differently from the traditional PD technique , the pancreas was cut before individualizing CHA origin and GDA. The SMA and CHA origin were then dissected from the uncinate process; the GDA was dissected at the upper border of the pancreas (Fig. 3), and divided after being clamped in order to check the CHA pulse. The common bile duct was dissected and cut, the jejunum was divided. The uncinate process was freed together with the retroportal lamina, thus enabling to remove the specimen. Then, the reconstruction was performed: end-to-side pancreatojejunostomy using Haka-mada technique [6] on the first jejunal loop, end-to-side hep-aticojejunostomy and a distal Roux-en-Y end-to-side gastrojejunostomy. Blood loss was estimated to 400 and 500 mL ,respectively. Histopathology showed a pT2N0M0 and pT3N0M0 pancreatic adenocarcinoma, respectively; margins were free in both cases.

Outcomes were unremarkable; there was no bleeding, no pancreatic leakage, and no infections. No biological sign of liver ischemia [6] was reported.

Fig.1.Michels'Hepatic Arterial type 5: Anatomy.

Fig.2.ContrastenhancedCT-scan,coronalslide:arterialpattern.

The most common pattern of arterial blood supply to the liver (CHA arising from the celiac axis, dividing in GDA and proper hepatic artery, this latter dividing distally into right and left hepatic branches) reaches 51.7-75.7% [2,3] of patients. During PD, unrecognized anatomical variations of liver and pancreas arteries may lead to vascular injuries, causing intra operative bleeding and/or postoperative liver or bowel ischemia. While there is no worry about a damaged GDA (which is normally secured and divided during PD), an injury to the CHA or to the SMA may have dramatic consequences. Indeed, injury to the CHA, the properhepatic arteryor one of its branches may lead to a chronic biliary ischemia in the related hepatic territory [5].Such an ischemia may concern just the supplied territory [7] or cause, in the most severe cases (mostly

Fig. 3. Picture after pancreas section, vessels  skeletonization. A:Pancreatic stump (head of pancreas). B: Hepatic Artery: Common Hepatic Artery (central yellow loop) and Gastroduodenal Artery (right-sided blue loop).C:Superior Mesenteric Artey (left-sided blue loop).

Involving CHA or proper hepaticartery),the acute necrosis of whole liver [4,7]. An injury to the SMA leads to small bowels and right colon ischemia/necrosis, and is often fatal.

In presented cases, a pivotal role in avoid in garterial injuries was played by preoperative CT scan and anatomical variations research. Therefore, a careful analysis of CT scan should be considered mandatory, not only to define the operability (with radical intent) of candidate to PD, but also to find out a typical arterial patterns and plan the correct surgical strategy.

Considering the surgical technique, reported cases of PD in patients presenting with anatomical variations suggest some commentaries.

First, splitting the pancreas before completing CHA and GDA dissection makes the procedure safer and easier. Indeed, since the CHA does not come across the portal vein at the upper border of pancreas (Figs.1-3), the pancreas can be divided without taking any risk of arterial injury. Then, the head of pancreas is retracted on the patient's right side in order to achieve an appropriate surgical exposure, thus allowing CHA and SMA dissection: this simple manipulatin enables to visualize and perform the skeletonization. Moreover, the pancreatic head retraction towards the right side ofthe patient also makes easier the uncinate process/retroportallamina dissection and the ligation of pancreatico-jejunal branches from the SMA.

Second, as it may be deduced by Figs.1e3, the sharp dissection and skeletonization of the CHA (running behind the head of pancreas and the tumor) and the GDA (that could have been Confused with CHA) along their atypical paths is the best way to avoid misunderstandings. It definitely represents the main difficulty of the procedure. Although obviously two cases do not set the rule, there was no surgical accident, no intraoperative major blood loss (400 and 500 mL, respectively) and outcomes were unremarkable.

Considering this new approach ,it does not allow to complete an“arteryfirst” approach [8,9] (the pancreas is divided prior skeletonization of the SMA). The “artery first” approach is supposed to increase oncological benefits by reducing positive marginre-sections rate. Unfortunately, no randomized controlled trial has demonstrated the “artery first” approach as reducing R1 resection rate (one trial is currently performed in Germany). Thus, replaced or accessory right hepatic artery in pancreaticoduodenectomy has been studied [10] :no difference was demonstrated in margin status or survival and encased accessory right hepatic artery should be considered as an encased SMA and be treated with neoadjuvant chemotherapy.

Since unusual patterns in celiac axis, SMA and related branches are relatively common, surgeons planning a PD must analyze in depth imaging (mainly CT-scan) and look for those patterns. Furthermore, some rare variations may be met and change typical surgical plan; keeping on mind vessels to save and those to scarify may allow to accomplish unusual but effective procedures.

1. Whipple, W.B. Parsons, C.R. Mullins, Treatment of carcinoma of the ampullaof Vater, Ann. Surg. 102 (1935) 763e779.
2. R. Hiatt, J. Gabbay, R.W. Busuttil, Surgical anatomy of the hepatic arteries in1000 cases, Ann. Surg. 220 (1994) 50e52.
3. H. Rong, W.F. Sindelar, Aberrant peripancreatic arterial anatomy.Consid-erations in performing pancreatectomy for malignant neoplasms, Am. Surg. 53(1987) 726e729.
4. Rosenbaum, H.L. Egbert, Liver necrosis and death following hepatic arteryligation, JAMA 149 (1952) 1210e1212.
5. Deltenre, D.C. Valla, Ischemic cholangiopathy, Semin. Liver Dis. 28(2008)235e246.
6. Hakamada,S.Narumi,Y.Toyoki,M.Nara,K.Ishido,T.Miura,N.Kubo,
7. Sasaki, An easier method for performing a pancreatico-jejunostomy for thesoftpancreasusingafast-absorbablesuture,WorldJ.Gastroenterol.14(2008)1091e1096.
8. K.Cho,S.S.Kim,Y.S.Do,K.B.Park,S.W.Shin,H.S.Park,S.W.Choo,I.W.Choo,
9. Ischemic liver injuries after hepatic artery embolization in patientswithdelayed postoperative hemorrhage following hepatobiliary pancreatic sur-gery, Acta Radiol. 52 (2011) 393e400.
10. Weitz, N. Rahbari, M. Koch, M.W. Büchler, The“arteryfirst”approach forresectionofpancreaticheadcancer,J.Am.Coll.Surg.210(2010)e1e4.
11. Pessaux, E. Marzano, E. Rosso, A plea for the artery-first dissectionduringpancreaticoduodenectomy, J. Am. Coll. Surg. 211 (2010) 142e143.
12. Turrini, E.A. Wiebke, J.R. Delpero, F. Viret, K.D. Lillemoe, C.M. Schmidt,Preservation of replaced or accessory right hepatic artery duringpan-creaticoduodenectomy for adenocarcinoma: impact on margin status andsurvival, J. Gastrointest. Surg. 14 (2010) 1813e1819.