The treatment of hilar cholangiocarcinoma continues to evolve, and improved survival time and a better life quality remains the primary aim. Hepatobiliary resection has become common in treating such patients [1-5], but radical high resection increases the difficulty in traditional biliary-enterostomy [3-5]. We performed extended high hilar resection and enlarged hepatic portal-enterostomy on 51 patients with hilar cholangiocarcinoma and analysed their clinical data.

Patients

      From March 2016 to March 2018, 56 hilar cholangiocarcinoma patients were referred for surgical treatment. Extended high hilar resection and enlarged hepatic portal-enterostomy was performed on 51 cases of hilar cholangiocarcinoma patients. One patient with cirrhosis was not included because the pathological diagnosis was hepatic cellular cancer with the invasion of hilar bile duct. Stent enterostomy was performed on 5 patients with metastasis in the abdomen cavity. Duplex ultrasonography (US), Computed Tomography (CT) and/or Magnetic Resonance Cholangiopancreatography (MRCP) were applied to evaluate the range of invasion and distant metastasis of tumours. Percutaneous transhepatic biliary drainage (PTCD) and bile refusion were performed on patients with a bilirubin level higher than 400 mmol/L and Child Grade B or above. Endoscopic retrograde cholangiopancreatography (ERCP) was performed when necessary. The Total Bilirubin (TB) level ranged from 132 to 623 mmol/L. Vitamin K was administered at a dose of 30 mg/ day for 3 to 15 days before operation. Symptoms of anaemia, hypoproteinaemia, and electrolyte disorders were treated and improved. Malnourished patients were given total parenteral nutrition preoperatively.

Surgical procedure

      The abdomen was entered and explored thoroughly, through a right lateral subcostal incision with left extension to the xiphoid. The ligamentum teres was elevated and the inferior surface of the liver was exposed. Precise assessment of tumour extension was evaluated. Careful palpation of the liver was performed to rule out suspected masses within the liver. Some lesser omentum was dissected and the lesser sac was accessed to palpate the caudate lobe. A Kocher manoeuvre was performed in some patients when the retroduodenal lymph nodes were suspected with metastases [3,4].

      The common bile duct was transected above the duodenum generally and turned upward. The lymph node, fat, nerve fibres, and other connective tissue along the common bile duct were cleared to achieve skeletonization of the hepatoduodenal ligament if radical resection was possible. The entire extrahepatic biliary apparatus and adjacent lymphatic tissues were turned upward to expose the portal vein. The dissection was continued upward towards the hilum, skeletonizing the portal vein and hepatic artery. It was easier to dissect the posterior aspect of the bile duct tumour from the portal vein and hepatic arteries to the level of the bifurcation of the portal vein provided that there was no invasion to the vessels. The gallbladder was freed from the liver, and the liver hilum was fully exposed anteriorly by taking down the gallbladder. At this point, the proximal extent of the tumour was assessed by palpation.

      The incision was made on the liver surface with the resection of at least 1 cm above liver tissue at the base of segment IV. The hepatic parenchyma was dissected using an electric scalpel combined with cavitron ultrasonic surgical aspirator. Hepatic parenchyma bleeding was controlled by electric coagulation and suturing. The branches of bile duct were transected using scalpel or scissors. Some tissues of caudate lobe were resected, combined right/or left lobectomy with caudate lobectomy was carried out when necessary (Figure 1). Rapid frozen section was performed to observe whether the margin of resection was tumour-free if radical resection was possible. The bile ducts with a diameter of over 5 mm were exposed and the wall was sutured with the surrounding tissues (Figure 2).

      When radical resection was possible, then the branches of portal and/or hepatic artery were invaded, the invaded branch was ligated and resected. Partial resection of the portal branch or portal vein were possible. If partial hepatectomy was required, the ipsilateral branch of portal vein and hepatic artery were ligated and en bloc caudate lobectomy was performed. When palliative biliary drainage was conducted, tumour tissue was resected as much as possible, however, the resection line should be assessed carefully to avoid heavy haemorrhage.

      A Roux-en-Y loop of jejunum was prepared and brought up in a retrocolic fashion (approximately 40 cm distal to the ligament of Teres). Anastomosis was carried out in an end-to-side fashion between the hepatic parenchyma incision line and the side of jejunum loop using a single layer of 0 silk. The anastomotic opening included all of the openings of the bile ducts without ligation of any biliary radicles. Suture was performed with a 1/2 radian suture needle, and the knot was tied in situ with moderate tension to avoid laceration of hepatic tissue. Posterior wall anastomosis was a difficult part of the surgery because there were few suturable parts of liver tissue anterior to inferior vena cava. Suture of the liver caudate tissue was not performed too deep as a result of the risk of injury to the inferior vena cava. Two to three stents were inserted into the bile duct branches with a diameter greater than 5 mm and were fixed using one single interrupted suture of 4-0 Vicryl suture with the bile duct wall, and one stent tube was pulled out through the wall of the jejunum loop (Figure 3). The intestinal wall and hepatic parenchyma were sutured together for anterior wall anastomosis of the jejunum loop with interrupted sutures, and all biliary radicles were drained into the hepatic portal bile lake (Figure 4). When palliative resection was performed, it was possible to suture the post wall of the jejunum with the tumour tissue. After completion of the anastomosis, air was injected into the jejunum loop from the stent tube to test whether there was leakage of the hepatic parenchyma-enterostomy anastomosis (Figure 5). Drainage was routinely employed near the anastomosis [3,4].

Post-operative management

      Nutritional support was given postoperatively for three days to patients with the balance of calories, protein, fat, trace elements and vitamins. It was administered by the peripheral intravenous route. If a nasoenteric tube was placed before operation, early enteral nutrition was utilized. Drainage was removed postoperatively. Stent tube was pulled out, after a cholangiography was obtained with injection of contrast medium across it.

      The study included 28 male and 23 female patients with ages ranging from 35 to 78 years old. All of the fifty-one patients underwent high hilar resection and hepatic hilar parenchymaenterostomy anastomosis. Five to twelve biliary radicles were exposed. Partial tumour resection and palliative biliary drainage was conducted on 17 patients, and one patient had a bile fistula for 2 months. Combined right/left hepatolobectomy with caudate lobectomy was performed on 16 patients. Eleven patients underwent right hepatic artery resection and partial resection of portal vein was performed on 4 patients. The operation time ranged from 190 minutes to 430 minutes, and the amount of bleeding was from 150 to 1200 ml. Two weeks post-surgery, the levels of serum aspartate aminotransferase and alanine aminotransferase decreased significantly. Icterus was significantly improved and the serum TB level decreased postoperatively. Postoperative complications occurred in 11 patients (one case of anastomosis leakage, five cases of wound infection, 3 patients with hepatic abscess, one case of haemobilia, one case of bile leakage from the hepatic transection surface). The postoperative hospitalization of patients were from 12 to 30 days. The patient with bile leakage died of pneumonia two months post operation. The performance of the patients were evaluated according to Karnofsky Performance Scale. The average Karnofsky score was 83 (ranging from 69-100), indicating that some patients had symptoms or signs but could engage in normal activities. The score increased significantly 2 months post operation.

      Cholangiocarcinoma accounts for 3% of all gastrointestinal cancers, and about 50-60% of them are hilar cholangiocarcinoma which occurs at the bifurcation of the left and right hepatic ducts. The tumours grows slowly, and many of them are diagnosed as advanced lesions with low radical resection rates and long-term survival rates. Local recurrence and re-obstructive jaundice are the main causes of morbidity and mortality post operation. The purposes of the management for these patients are to improve the quality of life and prolong the survival period. Surgical treatment is one of the major options, and expanding the scope of resection and postponing the reobstruction after biliary-enteric restoration, are the main measures. The main objectives in the therapy of hilar cholangiocarcinoma are complete tumour excision with negative margins if possible, and restoration of biliary-enteric continuity [6-11]. If resection is not clearly unfeasible, patients should be fully evaluated for possible curative resection before any intervention is performed since stentassociated infection and inflammation renders assessment and exploration difficult [1,2,6,11,12]. If resection is clearly not feasible, transtumoral drainage or paratumoral surgical bypass can be performed traditionally and offer palliation [1,7,11,13].

      Traditionally, the liver hilum is fully exposed anteriorly by taking down the gallbladder and lowering the hilar plate by incising Glisson's capsule along the base of segment IV [13,17]. It is not infrequent that removal of the tumour results in discontinuity of the right anterior and posterior sectoral ducts and/or separation of one or more caudate ducts from the left hepatic duct. In previous experience for hilar cholangiocarcinoma, after resection of tumours, the large bile ducts were refitted and end-to-side anastomosed with the jejunum and some small bile ducts (less than 1 to 2 mm) were drained not satisfactory or ligated. Expansion of the proximal bile ducts would compress the portal vein in the same Glisson’s sheath. This procedure often led to portal hypertension and cholestasis, which increased the incidence of cholangitis and bacterial infections[13-15]. In this study, the incision was made at, at least 1 cm above the Glisson's capsule along the base of segment IV to resect more hepatic tissues and achieve tumour-free margins[3, 4]. Portal parenchyma– enterostomy was performed without any bile duct integration or ligation. All bile ducts converged into the hepatic portal bile lake, and the bilirubin level of patients dropped close to normal levels after surgery. This procedure also provided effective biliary drainage in palliative resection patients.

      Hepatoportoenterostomy (Kasai portoenterostomy) was first proposed by Morio Kasai in 1974 for the treatment of children with congenital biliary atresia [18], and since then, it has been used for the treatment of high bile duct injury and cholangiocarcinoma [19]. Some modifications have been made on the basis of previous experiences. The main differences lie in that: 1) The hilar plate was conserved and after exposure of the hilar bile duct in the surgery procedure described by Kasai, fibrous tissue was anastomosed with the jejunum, whereas in this procedure, the hilar tissues were skeletonized, and the hilar plate, part of the caudate lobe, and the surrounding portal parenchyma were resected. 2) There were few or no hilar bile ducts in congenital biliary atresia patients, whereas in hilar cholangiocarcinoma patients, there were more than 5 inflatable hilar bile duct radicals. 3) Higher hilar resection should be performed to achieve tumour-free margins in hilar cholangiocarcinoma patients. 4) The posterior wall was anastomosed with the resection surface of the caudate lobe. 5) Some hilar cholangiocarcinoma patients were with inflammation around the hepatic hilar.

      This study was supported by The Natural Science Foundation of Shandong Province of China (2016GSF201180).