Outcome of surgical resection for hilar cholangiocarcinoma

Outcome of surgical resection for hilar cholangiocarcinoma

MagdyASorour, MD; Mohamed T El-Riwini, MD

CIT Surgery Unit, General Surgery Department, University of Alexandria, Egypt.

Background: Hilar  cholangiocarcinoma  (HC) is an adenocarcinoma  of the extrahepatic biliary tree arising  from the main left or right hepatic ducts or their confluence. HC is still considered to be a disease which is difficult to treat or to cure.  Its prognosis is very poor and the only curative treatment is complete resection with a negative surgical margin. Preoperative evaluation of the tumor is important in order to evaluate resectability and the extent of surgery. Surgical resection involves  either local resection, local excision  with caudate lobectomy or major hepatic resection with caudate lobectomy.The aim of this work was to study the feasibility and outcome of surgical resection  for hilar cholangiocarcinoma.

Methods: This study was carried out on 72 patients with radiological evidence suggesting resectable hilar cholangiocarcinoma presented to the Gastrointestinal Surgery Unit, Main Alexandria University  Hospital  during  the period  from March 2006 till January  2013.   All patients'presentations, laboratory and radiological data, surgical procedures, complications, follow-up and survival data were collected

Results:  Between March 2006 and January 2013; 72 patients  (56 males and 16 females) were admitted for resection of a radiologically suspected resectable hilar cholangiocarcinoma. All patients presented with obstructive jaundice of varying degrees. The tumors were Bismuth­ Corlette typelin36patients, typellin 24patients, type!Ilain 5patientsandtype!Ilbin 7patients. All patients underwent surgical resection. Forty-six patients underwent bile duct resection with hepaticoj([/unostomy and regionallymph node dissection. Fourteen patients underwent bile duct resection and lymph node dissection  with caudate lobectomy.  Three patients underwent right hepatectomy, two patients underwent extended right hepatectomy,  four patients underwent left hepatectomy and three patients underwent extended left hepatectomy;  all with hilar bile duct resection, caudate lobectomy and regional lymph node dissection. Complete resection (RO) was achieved in 65 patients {90.3%), while 7 patients  (9.7%) had incomplete resection. All over 9 patients  (12.5) developed recurrence. The 3 and 5-years overall survival rate  for all patients, using the Kaplan-Meier actuarial curve, were 51.4% and 34.7% respectively.

Conclusion: Cholangiocarcinoma remains a devastating disease. Most patients have unresectable tumors at the time of diagnosis and have a dismal prognosis. Complete resection is the only treatment that offers any hope of long-term  survival but is possible in  few patients. Furthermore,  even after resection,  disease recurrence is common. Acfjuvant therapy has not been shown to have a role in this disease.

Key words:  Bile duct cancer, hilar cholangiocarcinoma,  Klatskin tumor, staging, surgical resection.

Introduction:

Hilar cholangiocarcinoma (HC) is an adenocarcinoma  of  the  extrahepatic  biliary tree arising from the main left or right hepatic ducts or their confluence. This tumor has been referred to as Klatskin tumor after Dr. Gerald

Klatskin published his paper in 1965.1 About two-thirds  of  extrahepatic  bile  duct  (EBD) cancers arise at the  hepatic hilum (Klatskin tumor) with one third arising from the distal common  bile  duct.2,3 The  diagnosis  of  HC should be suspected in patients with painless

jaundice     whose    CT    scan    demonstrates dilated intrahepatic  bile ducts with a normal gallbladder and extrahepatic biliary tree.4

HC  is  still  considered  to  be  a  disease which   is   difficult   to   treat   or   to   cures. Its  prognosis   is  very  poor  and  the  only curative   treatment    is   complete   resection with  a negative  surgical  margin.2,3,6,7  Most patients   with   cholangiocarcinoma  present with  advanced  disease that is not amenable to  surgical  treatment.8  Palliative  resection, surgical bypass procedures, and various types of  intubation  and  drainage  procedures  are associated with 3-year survival rates from 0% to 4%.4 Adjuvant therapy (chemotherapy and radiation therapy) has not been shown clearly

to reduce recurrence risk8.

In the extrahepatic ducts, three distinct macroscopic subtypes of cholangiocarcinoma are well described: sclerosing, nodular, and papillary.9 Sclerosing tumors are the most common subtype and are more common at the hilus than in the distal bile duct. Sclerosing tumors are very firm and cause an annular thickening ofthe bile duct, often with diffuse infiltration   and  fibrosis   of  the   periductal tissues and can form an associated mass. Nodular tumors are characterized  by a firm, irregular  nodule  of tumor that  projects  into the lumen ofthe duct. Features ofboth types are often seen, hence the frequently used description nodular sclerosing. The papillary variant   accounts   for   approximately    10% of all cholangiocarcinomas and, although occasionally seen at the hilus, is more common in the distal bile duct.9 These tumors are soft and friable  and may be associated with little transmural  invasion.  Tumor fragments  may be present  in the  distal  bile  duct,  resulting in intermittent jaundice. Recognition of this variant is important because it is more often resectable and may have a more favorable prognosis   than   the   other  types,   although this  has  not  been  proved  definitively.9  It can be divided further into papilloma type, intraductal growing type, mucin-producing type, and cystic type.lO,ll

Ninety percent of patients with HC have painless   jaundice,    10%  have   cholangitis, and   56%  have   systemic   symptoms   such

as malaise, abdominal discomfort, nausea, anorexia, and weight loss_l2,13 Preoperative evaluation   of   the  tumor   is  important   in order   to   evaluate    resectability    and   the

extent  of surgery.2,3,6  In  patients  with  hilar

cholangiocarcinoma, evaluation must address four critical components of resectability: level and extent  of tumor  within the  biliary tree, vascular invasion, hepatic lobar atrophy, and distant  metastatic  disease.  Several  methods are  proposed  to  evaluate  tumor  extension. Bismuth-Corlette classification has been used to define the longitudinal tumor extension.l4

Resectability of the tumor can be evaluated by the Blumgart T-staging system combined with the AJCC cancer staging  system (American

Joint  Committee  on  Cancer). 15 Assessment

of the lateral spread and soft tissue extension can be evaluated  based on the TNM staging system.l6,17 Lymph node metastasis  is a significant factor helping to determine patient outcome after surgery for HC.l8,19

MRI, CT, endoscopic retrograde cholangiography (ERC), and, perhaps, endoscopic ultrasound (EUS) are used most frequently to diagnose and stage HC. The Bismuth-Corlette   type   I  tumor   is  defined by the presence  of a lesion  confined  below the confluence of the right and left hepatic ducts. This type oftumor can be treated with segmental resection ofthe EBD and regional lymph node dissection. Hepatectomy is unnecessary if the resection margin is microscopically  confirmed to be negative by frozen pathology during surgery.20 Bismuth type II tumors extend to the confluence of the right and left hepatic ducts. Bismuth type  II tumor  can be treated  by bile duct resection with  hepaticojejunostomy   and  regional lymph  node  dissection.  Caudate  lobectomy is mandatory  when the tumor  infiltrates caudate  bile duct  branches_20,21 A type  Ilia tumor  extends to the bifurcation of the right hepatic duct, and a type IIIb tumor extends to the bifurcation of the left hepatic duct. Hilar bile duct resection with hemihepatectomy including the caudate lobectomy and regional lymph  node  dissection  is  the  standard surgical method for type III tumor.20-22 Type IV tumors extend to the bifurcation  of both

the right and left hepatic ducts and have been generally  regarded as inoperable except for liver transplantation; multicentric tumors  are also included  in this category.l7

A preoperative T staging  system,  defined by   biliary   tumor    extent,   the   presence   or absence  of  portal vein  involvement, and the presence  or absence  ofhepatic lobar atrophy, may   be  useful   for   predicting  resectability and   the   likelihood  of   finding   metastatic

disease.23,24

The aim of this work was to study the feasibility and outcome  of surgical  resection for hilar cholangiocarcinoma.

Methods:

Between  March  2006  and  January  2013;

72 patients  (56 males and 16 females) with radiological evidence suggesting resectable hilar    cholangiocarcinoma   were    admitted in the Gastrointestinal Surgery Unit, Main Alexandria University Hospital  for surgical resection.  If  unresectability  was   identified at laparotomy, palliative  options were performed and these  patients  were  excluded from the study.

After  approval  of local  ethics  committees

of   both   the   General   Surgery   Department and the Alexandria Faculty  of  Medicine, all patients  included in the study were  informed well about the operative  procedure and the possible  complications and  an  informed written  consent  was obtained  from every patient  before carrying the procedure.

All patients were subjected to the following

Preoperative assessment:

Clinical 1) Complete history taking.  2) Thorough clinical examination

Laboratory work-up.  1) Routine laboratory studies  2) Bilirubin, alkaline  phosphatase, CA19.9, CEA.

Imaging. 1) Abdominal duplex ultrsonography 2) Multi-slice CT (MSCT) abdomen.  CT  angiography whenever indicated. 3) MRI with magnetic resonance cholangiopancreatography (MRCP).  4) Endoscopic ultrasound  (EUS):  whenever indicated.  5. Endoscopic retrograde cholangiography    (ERC):     whenever indicted.      6)     Percutaneous    transhepatic

cholangiography  (PTC):   as  an   alternative to  ERC  when  endoscopy is unsuccessful or technically unfeasible. 7) Laparoscopy.

Cytology.  1) Bile  duct  brushing  cytology

at the time of ERC and PTC. 2) Peritoneal washing at the  time  of laparoscopy. 3)  Fine needle  aspiration cytology at the time of endoscopic ultrasound

Criteria of unresectability23,25

Patient factors

- Medical unfitness

- Advanced  hepatic  cirrhosis

Local tumor-related factors

-Invasion of the secondary biliary radicles bilaterally

- Encasement or occlusion of the main portal vein proximal  to its bifurcation

-  Any  two  combinations (one  unilateral

and one contralateral) of hepatic lobe atrophy, portal  vein  branch  encasement or occlusion and secondary biliary radicles  involvement.

Metastatic Disease

- Histologically proven  metastases to N2 lymph nodes (peripancreatic, periduodenal, celiac, superior  mesenteric, or posterior pancreaticoduodenal lymph nodes)

- Lung, liver, or peritoneal metastases

Metastatic disease to cystic duct, pericholedochal, hilar or portal lymph  nodes (i.e., within the hepatoduodenalligament) did not necessarily constitute unresectability.

Complications related  to biliary  tract obstruction or  previous   biliary  intervention (ie,   cholangitis,   pancreatitis),   if   present, were     treated      before     surgery.     Routine biliary  drainage  of jaundiced patients, not previously stented   and  without   cholangitis, was  not  performed  if an operation could  be performed in a timely fashion (within  one week).  Preoperative portal vein embolization was not used as there was no lobar atrophy  in our senes.

Biopsy  material  from  the  referring  center

was re-examined. In patients unfit for surgery and those with advanced disease, biopsy confirmation was performed, if not done previously.  However,  when  the  imaging studies   suggested  a   potentially  resectable HC,  histologic confirmation of  malignancy was not performed.

Surgical resection:  its extent was determined by preoperative and intraoperative data. Because cholangiocarcinoma is known to spread along the wall of the bile ducts and because the caudate lobe is a frequent site of tumor recurrence following extrahepatic duct resection, a resection that includes the caudate lobe  was  performed.  Tumor  extension  into only the right or the left lobe was resected by an en bloc extended left or right hepatectomy.

The  resection  was  performed  through  a

bilateral  subcostal  incision.  A self-retaining retractor was used to provide adequate exposure. Following thorough exploration to confirm the absence of peritoneal disease, a complete retroduodenal and retropancreatic lymph node dissection was performed by mobilizing the entire right colon, duodenum, and pancreatic head in a lateral to medial rotation.  The  node-bearing  tissue  was dissected  by  removing  the  anterior  aspect of Gerota fascia of the right kidney, then completely clearing all tissue off the anterior aspects  of  the  right  renal  vein,  right  renal artery, inferior vena cava, and aorto-caval groove.

After completion of the retroduodenal and

retropancreatic  dissection, the  gastrohepatic and  gastroduodenal  ligaments  were  divided along the superior  edge of the stomach  and duodenum. The cystic duct and cystic artery were identified. The cystic artery was ligated and divided. The gallbladder was completely dissected  free from  the gallbladder  bed but left attached to the bile duct by the cystic duct. The gastroduodenal artery, the proper hepatic artery,  and  the  portal  vein  were  dissected circumferentially    from    the     gastrohepatic ligament  and  the  bile  duct.  This  dissection was carried from caudad to cephalad toward the confluence ofthe left and right portal veins and the  liver plate. Portal vein involvement was considered present if the tumor contacted and either distorted or narrowed the vein, or if the  vein  was  encased  or occluded.  After the  bile duct tumor  was dissected  from the anterior surface of the portal vein, the distal common bile duct at the superior aspect of the duodenum  was  divided  and  suture-ligated. The   common   bile   duct,   gallbladder,   and

node-bearing porta hepatis tissue were then reflected in an anterior and cephalad direction to complete the posterior dissection along the right and left bile ducts. A circumferential dissection  of  the  right  and  left  bile  ducts was performed  if tumor  was isolated to the confluence ofthe right and left bile duct (not extending up into either the right or left duct). The right and left bile ducts were then divided sharply at least 1 em proximal to the tumor. The surgical specimen was removed and included the distal right and left bile ducts, the common  hepatic  duct, the common  bile duct with the attached gallbladder, and the regional  node-bearing  tissue.  The specimen was immediately evaluated using frozen­ section  studies  on  the  right  and  left  bile ducts to confirm tumor-negative  margins. At that point, the gastroduodenal artery, proper hepatic artery, left and right hepatic arteries, portal vein, and inferior vena cava had been "laid bare" Figure (1).

The resection was completed by performing

a caudate lobectomy. The capsule of the caudate  lobe was scored with electrocautery and  then  the  caudate   lobe  was  dissected free from the vena cava using an ultrasonic dissecting  instrument.  Small  vessels  within the parenchyma of the caudate lobe were clipped and larger vessels were ligated. The removal of the caudate lobe was completed by dissecting the liver parenchyma free from the anterior aspect of the retrohepatic vena cava. The small venous branches draining directly from the caudate lobe into the vena cava were identified,  suture-ligated,  and  divided  until the caudate lobe had been removed.

Biliary-enteric drainage was reestablished

using a Roux-en- Y jejunal loop. The right and left bile ducts were individually anastomosed end-to-side to the jejunal loop using full­ thickness interrupted 4-0 absorbable vicyl sutures    Figure (2).    The   operation   was completed  by placing  two tube  drains  near the base of the liver. The abdominal wall was closed in layers.

Incases wherethe hilarcholangiocarcinoma extended directly into the right or left bile duct, an extended right or left hepatectomy was performed en bloc with the extrahepatic  bile

duct, gallbladder,  and regional  node-bearing tissue Figures(3,4). Only a single Roux-en-Y hepaticojejunostomy was performed.

A nasogastric  tube  was  placed  during

the operation. We generally remove the nasogastric tube on the 1st postoperative day. On postoperative day 5, we removed the tube drains. All patients received a perioperative dose of a third-generation  cephalosporin antibiotic and then received scheduled intravenous dosages of this antibiotic  during the following 5 days after surgery.

In patients who had a bile leak from a biliary-enteric  anastomosis  that was drained by  the  surgically  placed  drains,  the  drains were not removed on postoperative day 5 but were kept in place until the biliary fistula had closed. If a fluid collection was demonstrated, percutaneous drainage of the bile collection was performed.

Postoperative  data:  i)  Histopathology  of the resected tumor. ii) Hospital stay. iii)Early and late post-operative complications.

Patient follow-up:  Patients were followed regularly in an outpatient setting every one to six months. The sites of disease recurrence (local and distant) were determined from imaging studies, including CT and MRI. Survival (overall and disease-free survival) were also estimated.

Data    were     presented     with     numbers,

percentage, arithmetic mean (X) and standard deviation (SD) and were analyzed with SPSS (version 16) statistical software. Disease free survival curve and overall survival curve were estimated  using  the  Kaplan-Meier  method. P values  less than  0.05  were considered  to be   statistically   significant.  A  multivariate Cox  proportional  hazards  model  was  used to evaluate which factors demonstrated an independent effect on postoperative estimated survival.

Results:

Between  March 2006 and  January  2013;

72  patients  were  admitted  for  resection  of a radiologically suspected resectable hilar cholangiocarcinoma.  Their mean age at diagnosis  was  57.76  ±10.23  years  (ranged from 41 to 69 years).

All patients presented with obstructive jaundice of varying degrees. Other presentations  are shown  in Table(l).  Fifty­ eight patients were Child grade A and 14 patients were  Child grade  B.  Sixty-five patients had non-cirrhotic livers and 7 patients had cirrhotic liver.

Data regarding  patients' age, sex, tumour extension, extent of resection, operative time, intra-operative bleeding (as recorded by the anaesthesiologist),  histological  grade, lymph node status, hospital stay (from the time of operation to discharge) and postoperative complications are shown in Tables (2,3,4).

Tumors  were  Bismuth-Corlette  type  I in

36 patients, type II in 24 patients, type Ilia in 5 patients and type  IIIb in 7 patients. All patients underwent surgical resection. Forty­ six  patients  underwent  bile  duct  resection with hepaticojejunostomy and regional lymph node dissection. Fourteen patients underwent bile duct resection and lymph node dissection with caudate lobectomy. Three patients underwent right hepatectomy, two patients underwent    extended    right    hepatectomy, four   patients   underwent   left  hepatectomy and three patients underwent extended left hepatectomy; all with hilar bile duct resection, caudate lobectomy and regional lymph node dissection. In the 12 patients who underwent major hepatic resections, the residual liver volume  was  more than  30%  in  10 patients and less than 30 in two patients.

Operative     estimated    blood     loss     was

greater in patients who underwent hepatic resection (26 patients; caudate lobectomy  in

14 and major hepatectomy in 12). In addition, all patients who underwent major hepatic resection required blood transfusion.

The number of lymph nodes removed was fewer than 12 in 17 patients and equal or more than 12 in 55 patients. The median number of lymph nodes removed was 17 (range 7-32). Lymphatic invasion  was negative  in 40 patients and positive in 32 patients.

The  lymph  node  ratio   (LNR),   defined

as the ratio of the number of lymph nodes harboring metastases to the total number of lymph  nodes removed  was less than 0.2  in

57 patients and equal or more than 0.2 in 15

patients.

Complete resection  (RO) was achieved  in

65 patients, while  7 patients  had incomplete resection. Two patients died of hepatic failure. Their   ages   were   59  and   66  years.   Death occurred at 7 and 35 days following extended left hepatectomy and right hepatectomy.

All   over   9   patients    (12.5)   developed

recurrence. One patient  underwent revisional hepaticojejunostomy following the development of  a left  hepatic  duct  stricture

1 year after  extended right  hepatectomy and bile duct excision. Subsequent histological examination showed  this to be due to tumour recurrence,  although   the  patient   was  alive and symptom free  15 months  after this reintervention.

For  all  72 patients, the  overall  estimated

survival rate  was  51.4%  at three  years  and

34.7%  at  five   years.   Forty   patients   with no lymph node metastases had a five-year estimated survival of 55% in comparison to a five-year  survival  of 9.4% for the 32 patients with  lymph   node  metastases  (p  =  0.015). Fifty-seven patients  with  LNR  less than  0.2 had  a five-year estimated survival  of 43.9% in comparison to  a five-year  survival  of 0% for the 15  patients  with  LNR  equal  or more than 0.2 (p = 0.005).

In the multivariate analysis, the following factors  were independent prognostic  survival factors  (1) Child's grade, (2) residual disease, (3) lymph node metastasis and (4) lymph node ratio (LNR). Other factors, such as operative time, intra-operative bleeding,  histological grade,  and number  of lymph  nodes  removed were  not  significant  predictors of  estimated survival.

Discussion

This study included 72 patients with a radiologically suspected resectable  hilar cholangiocarcinoma. The first sign of hilar cholangiocarcinoma in  the  studied  patients was jaundice. Many patients  had one or more nonspecific complaints, such as abdominal pain,  general  malaise, anorexia, and  weight loss.  Although  most  patients  eventually became  jaundiced, those  with  ipsilateral or segmental involvement may  have  abnormal

liver function tests and even  pruritus  without jaundice.26 Cholangitis is rarely  a presenting feature  in patients  with  cholangiocarcinoma in  the   absence   of  prior  biliary   intubation. Most    patients    with   hilar     strictures    and jaundice have cholangiocarcinoma. However, alternative  diagnoses   can   be   expected   in

10 to 15%  of patients,  the  most  common  of

which  are gallbladder carcinoma, Mirizzi syndrome,  and  idiopathic  benign   focal stenosis  (malignant masquerade).26

Benign      strictures      of     the     proximal biliary tree are uncommon, and hilar cholangiocarcinoma must remain the leading diagnosis until definitively disproved. In most cases, this cannot be done without exploration. Relying  on the results of percutaneous needle biopsy or biliary brush cytology  is dangerous, because  the results are often misleading, and one  may  miss  the  opportunity to  resect  an early cancer.

In the present study, tumors were Bismuth­

Corlette  type  I in 36  patients,  type  II  in 24 patients,  type  Ilia  in 5 patients  and type  IIIb in 7 patients.  Bismuth-Corlette classification has   been   used   to   define   the   longitudinal tumor   spread   in  one  dimension  along  the bile  duct  and  it does  not  incorporate radial tumor  growth. 27  Surgical  candidates cannot be  determined solely  by  this  classification, and it is not indicative of survival.28 Another aspect  to consider  is that longitudinal spread pattern  of  a tumor   can  be  related  to  gross morphology.29   Papillary  tumors   frequently present with long-range mucosal spread, while infiltrating tumors  tend to show  subepithelial extension. The  subepithelial infiltration  may readily be depicted  on CT or MRI by showing thickening or increased enhancement of the ductal   wall,   but  the   mucosal   spread   may hardly  be visible  on  CT or  MRI.  Therefore, determination of longitudinal spread  must be made  more  cautiously when  a  papillary or polypoid  tumor  is seen on imaging.  Abe Met aP0 presented an illustrative case in which the utility   of  choledocoscopy  is  demonstrated. In   that   case,   cholangiography  showed   a polypoid   tumor   in   the   middle    CBD,   but choledochoscopy   demonstrated   multifocal superficial spreading tumors  along the  entire

Figure {1): En bloc resection of Klatskin tumour {Bismuth-Corlette type I) shows the proximal right and left ducts {RD,LD), the gastroduodenal artery {GDA), proper hepatic artery {HA), left and right hepatic arteries {RHA,LHA), portal vein {PV),and inferior  vena cava {IVC).

Figure {2): (it) En bole resection of Klatskin tumor {Bismuth-Corlette  type II) with caudate lobectomy. {B) Roux-en-Y hepaticojejunostomy.

Figure {3): {it, B) Extended left hepatectomy for Klatskin tumor {Bismuth-Corlette type Illb).

Figure {4): {A, B) Extended right hepatectomy  for Klatskin tumor {Bismuth-Corlette type lila).

bile duct necessitating more extensive surgery than was expected from the cholangiography alone.

Matsuo   K  et  al24   provides  convincing

evidence that a preoperative clinical T staging system, originally proposed by Blumgart and subsequently modified, effectively predicts resectability   and   likelihood   of   metastatic

Table (1): The different presentations encountered in patients with hilar cholangiocarcinoma

Table (2): Patients' age, sex, liver disease, tumor extension and associated co-morbidities.

disease. Intheircohortof380patientsmanaged over an 18-year period, approximately 60% of all patients had unresectable disease, either at presentation or exploration. Although locally advanced  disease  was  an  important  factor, the  most  common  reason  for  irresectability was metastatic disease. In their analysis, the resectability  rate was  64.3%  in T1 tumors,

41.3% in T2 tumors, and 1.3 (1 patient) in T3 tumors.  Resectability,  the  probability  of an

RO resection, and the likelihood of metastatic disease   correlated   significantly   with   the clinical  T stage.  They found  that  complete resection with negative margins almost never occurred for T3 tumors but was much more frequent   in  T1  (44.1%)   and  T2  (36.3%) tumors.  The  presence  of  distant  metastatic disease increased with more locally advanced and  higher   clinical  T  stage  tumors,   with

48.7%  of  T3  tumors  harboring   metastatic

Table (3): Extent  of resection,  operative time,  intra-operative bleeding,  histological  grade, lymph node status and hospital stay.

disease, followed by 34.8% of T2 and 19.6%

of T1lesions.

In an analysis of 225 patients  with hilar cholangiocarcinoma,  resectability was nearly

60% in T1 tumors, 31% in T2 tumors, and 0%

in T3 tumors.  Survival  also  decreased  with increasing clinical T stage.23

Positron emission tomography  has shown a high sensitivity for diagnosing biliary malignancy. Its limitation is that the patients

with biliary tract infections or inflammatory processes in the biliary tree (as in PSC) can have false positive results. Its best use may be as a diagnostic tool after resection to discover recurrence.31

All  patients  m  the  present  study underwent  surgical  resection.  Forty-six patients underwent bile duct resection with hepaticojejunostomy   and   regional    lymph node dissection. Fourteen patients underwent bile duct resection and lymph node dissection with caudate lobectomy. Three patients underwent right hepatectomy, two patients underwent    extended    right    hepatectomy, four   patients   underwent   left  hepatectomy and three patients underwent extended left hepatectomy; all with hilar bile duct resection, caudate lobectomy and regional lymph node dissection.  In the 12 patients who underwent major hepatic resections, the residual liver volume was more than 30% in 10 patients and less than 30% in two patients. The extended resections that enable a better oncological clearance have become achievable as a result of the major advances in surgical techniques and preoperative  and postoperative  care, which have reduced morbidity and mortality after major hepatic resection.

The role of preoperative  biliary drainage in jaundiced patients remains controversial. Most patients undergo biliary drainage prior to  referral for resection,  despite the  lack of data showing a benefit. The presence of cholangitis mandates biliary decompression, but there is no proof that routine biliary drainage  in  all patients  facilitates  resection or  reduces  postsurgical  morbidity. 32,33  On the contrary, the available data would suggest

that biliary stents are associated with greater postoperative   infection     complications.34

Whether major hepatic resection in the face of biliary obstruction is associated with a greater risk of liver failure or other complications remains  an  open  question.35  In  patients  of this study, preoperative biliary stenting is not recommended as it makes dissection more difficult and time-consuming and at the same time  decompresses  the  biliary tree reducing its   diameter   and   making   further   biliary­ enteric anastomosis more difficult.

In the present study, although imaging studies  identified  many  patients  with resectable  disease,  a  significant  proportion were  found   to   have  unresectable   disease only  at the time  of laparotomy.  Nearly  one third   of  patients  had   unresectable  tumors at presentation. However, of the remaining patients  with  potentially  resectable  lesions, only 50% underwent resection. As a result, staging  laparoscopy  has  been  increasingly used in an effort to reduce the incidence of unnecessary open explorations. In a recent analysis of 56 patients with potentially resectable tumors based on radiological findings, laparoscopy identified unresectable tumors in 14 patients (25%).36 The yield was significantly  higher for patients with clinical T2/T3  tumors  (12  of  33,  36%)  compared with those with clinical T1 tumors (2 of 23,

9%), which is almost certainly related to the higher incidence  of metastatic  disease in the former  group.   Laparoscopy   detected  most patients  with  peritoneal  or liver  metastases but failed to  detect  all locally  unresectable tumors.   Despite   this   limitation,   however, laparoscopic staging appears to have a role in

these patients.

Complete resection  (RO) was achieved in

65 patients, while 7 patients had incomplete resection.  In  patients  with  potentially resectable tumors, there is no doubt that the primary goal of surgery should be a complete resection  with  histologically  negative resection   margins   (i.e.,   a   RO  resection), which at a minimum requires resection of the extrahepatic biliary apparatus and subhilar lymphadenectomy. However, there are now substantial data to suggest that en bloc partial hepatectomy  is also required  in most  cases. The results of recent studies show a parallel between the number of patients undergoing partial  hepatectomy  and  those  having negative resection margins which is a potent predictor of outcome_23,37-46 In addition, tumors involving the left hepatic duct almost always involve the main caudate duct and usually require a complete caudate resection as wel1.47 Extensive resections for hilar cholangiocarcinoma  have  been   associated with   significant   morbidity,   and   mortality rates, even at high-volume centers, are on the order of 5 to 10%. Infective complications are particularly common and often play a central role in postoperative mortality.23

In this study, the number of lymph nodes removed was fewer than 12 in 17 patients and equal or more than 12 in 55 patients. The 40 patients with no lymph node metastases had better five-year survival than the 32 patients with  lymph  node  metastases  (55%  versus

9.4%).  Fifty-seven  patients  with  LNR  less than  0.2 had  a five-year  survival  of  43.9% in  comparison   to  a  five-year  survival   of

0% for the  15 patients  with  LNR  equal  or

more  than  0.2.  Some  authors  established  a cut-off  point  of 12 for the  number  of  LNs removed  in  patients  with pancreaticobiliary carcinoma.48,49  Similarly, in this  study, we adopted  the  same  cut-off  point.  Numerous studies have demonstrated  that patients with lymph  node  metastases  have   significantly worse survival   rates   than   patients   with node-negative              disease.     The    number     of lymph  node  metastases   was  an  important prognostic factor in patients with extrahepatic cholangiocarcinoma    after     resection.50,51

The lymph node ratio (LNR), defined as the ratio of the number of lymph nodes bearing metastases   to  the  total  number   of  lymph nodes   removed,    has   been   demonstrated to  be  a  more  important   prognostic  factor in  gastrointestinal  carcinomas  than  the presence  of absolute number of lymph node metastasis.S2  The   evaluation   of   LNR   is useful  for  several  reasons.  LNR  combines data regarding the number of lymph nodes removed    with   the    number    of   positive nodes, providing information regarding the adequacy of lymphadenectomy. Moreover, LNR is a simple method to stratify patients more accurately  via a method that considers both the biology of the disease (number of positive nodes) and the adequacy of lymph nodes dissection (number of excised nodes). Metastatic  disease to  regional  lymph  nodes is not uncommon in patients with hilar cholangiocarcinoma.    In   a   recent   review of  110  patients,   Kitagawa   et  aP3   found that 47% had no involved nodes, 35% had regional  lymph  node  metastases,  and  17% had regional and para-aortic node metastases. There was a significant survival difference based on nodal status. Node-negative patients had 3- and 5-year survival rates of 55% and

30%,  respectively,  compared  with  32  and

14.7%, respectively, for those with regional nodal  metastases  and  12.3%  in  both  cases for those with para-aortic node metastases.53

Kawai et al48 demonstrated that LNR was an

important  prognostic  factor  after  resection of middle and distal cholangiocarcinoma. Extended    lymph    node    dissection    with careful  examination  for  metastases  allows for   a   more   accurate   evaluation   of   LN status and should be performed whenever feasible  as it offers  a survival  advantage  in hilar cholangiocarcinoma.  In this series, metastatic  disease  to  lymph  nodes  beyond the hepatoduodenal ligament (celiac, para­ aortic, and so on) was a contraindication  to resection.

Complete resection was associated with a

better survival. However, some patients were not candidates for resection, and these patients were excluded from the study after palliating the biliary obstruction.  Percutaneous  biliary

drainage   and  subsequent   placement   of  a self-expandable     metallic           endoprosthesis was         the    preferred           approach          and       was successfully    performed   in   most   patients with irresectable  HC. Endoscopically  placed stents  were usually ineffective for  proximal biliary obstruction.  Biliary stenting  was not appropriate  in patients with an atrophic lobe as it did not relieve jaundice and was avoided, unless performed to control infection. Patients found to have unresectable, locally advanced tumors   at   operation   were   candidates   for intrahepatic         biliary-enteric               bypass.             Not infrequently,  unresectability  was  discovered only after an extensive dissection,  including transection of the distal bile duct, and biliary­ enteric  anastomosis  was then established  to restore continuity. In such cases, the segment III  duct  was  often  used, although  the  right anterior  or  posterior  sectoral  hepatic  ducts were also used in some cases. In a review of

55 consecutive bypass procedures in patients

with malignant hilar obstruction, the authors found segment III bypass yielded the best results, with a 1-year bypass patency rate of

80%.54 An  advantage  of this  approach  was

that the anastomosis was away from the tumor and less susceptible to recurrent obstruction due to disease progression.

There  were  two  perioperative  deaths  in

this series. Death occurred at 7 and 35 days following   extended   left  hepatectomy   and right hepatectomy. The cause of death was hepatorenal failure  in the two cases. Several groups have explored the possible role of preoperative portal vein embolization; the rationale  for  this  is to  induce  hypertrophy of the future liver remnant prior to surgery, thereby potentially reducing the risk of postoperative hepatic failure.S5-57 Although this  technique  may  be  of  some  value,  the lack of compelling controlled data makes it difficult to advocate its routine use.

All over 9 patients  (12.5) developed

recurrence. In this study, the overall estimated survival  rate was  51.4% at three  years  and

34.7%  at  five  years.   Following  resection,

5-year  survival  rates  were  significantly greater in patients who underwent a RO resection.  Furthermore,  survival  in  patients

with  histologically   involved   resection margins  (R1) was  little better  than survival in those with unresectable,  locally advanced tumors  was. Although  5-year  survival  after a RO resection was approximately  38.5%  in this series, it is notable that cancer recurrence after 5 years was not uncommon.

This study is limited by the small number

of patients. The results ofthis study will need to be confirmed in a multi-institutional cohort of patients.

Conclusion:

Cholangiocarcinoma  remams  a devastating disease. The treatment for cholangiocarcinoma remains a challenge because   of  the  aggressive   nature   of  the disease  and  the  absence  of  effective treatments besides surgical resection. Most patients have unresectable tumors at the time of diagnosis and have a dismal prognosis. Complete resection (RO) is the only treatment that  offers  any  hope  of  long-term  survival but is possible in few patients. Furthermore, even   after   resection,    disease   recurrence is  common.   Lymph  node  ratio   (LNR)   is a powerful independent prognostic factor predicting survival in patients with hilar cholangiocarcinoma after surgical resection. The development of diagnostic modalities (tumour  markers,  cytology  and  radiology) are of the utmost importance to identify these patients at an early stage to preserve radical surgery  possible.  Adjuvant  therapy  has  not been shown to have a role in this disease.

Reference:

1-  Klatskin G: Adenocarcinoma of the hepatic duct at its bifurcation within the porta hepatis. An unusual tumor with distinctive clinical and pathologic features. Am 1Med

1965; 38: 241-256.

2-   Khan SA, Thomas HC, Davidson BR, Taylor­ Robinson SD: Cholangiocarcinoma. Lancet

2005;366: 1303-131

3-   LazaridisKN,GoresGJ:Cholangiocarcinoma.

Gastroenterology2005; 128: 1655-1667.

4-   Curley SA:  High  malignant biliary tract

obstruction. In: Mastery of Surgery. Fischer JE, Bland KI (Editors); Philadelphia: Lippincott Williams & Wilkins (Publisher);

5th edn.; 2007; 1165-1174.

5-    Lau   SHY,  Lau   WY:  Current   therapy   of hilar    cholangiocarcinoma.    Hepatobiliary Pancreat Dis Int2012; 11: 12-17.

6-    Patel T: Cholangiocarcinoma. Nat Clin Pract

Gastroenterol Hepatol2006; 3: 33-42.

7-    de  Groen  PC,   Gores  GJ,  La  Russo  NF, Gunderson LL, Nagorney DM: Biliary tract cancers. NEngl1Med1999; 341: 1368-1378.

8-    Jarnagin WR,Shoup M: Surgicalmanagement

of  cholangiocarcinoma.  Seminars in  Liver

Disease 2004; 24: 189-199.

9-    WeinbrenK, Mutum SS: Pathological aspects of cholangiocarcinoma. 1Pathol1983;139:

217-238.

10- Nakanuma  Y, Sato  Y, Harada  K, et al: Pathological classification of intrahepatic cholangiocarcinoma based on a new concept. World1Hepatol2010;2: 419-427.

11-  Komuta  M, Govaero  0, Vandecaveye V, et

al: Histological diversity in cholangiocellular carcinoma reflects the different cholangiocyte phenotypes. Hepatology2012;55: 1876-1888.

12-  Blechacz B, Gores GJ: Cholangiocarcinoma: Advances in pathogenesis, diagnosis, and treatment. Hepatology2008; 48: 308-321.

13-  Nagorney  DM,  Donohue  JH,  Farnell  MB, et al: Outcome after curative resections of cholangiocarcinoma. Arch Surg 1993;  128:

871-879.

14- Bismuth  H,  Corlette   MB:   Intrahepatic cholangioenteric anastomosis  in  carcinoma of the hilus of the liver. Surg Gynecol Obstet

1975; 140: 170-178.

15-  Greene  FL,  Page  PD,  Fleming  ID,  et  al: AJCC  cancer  staging  manual.  6th  edition. New York: Springer-Verlag; 2002.

16-  International Union Against Cancer (UICC):

TNM classification of malignant tumors.  (Jh

edn. 2002; 81-89.

17-  Chung   YU,   Kim    MJ,   Park    YN,   Lee YH, Choi JY: Staging of extrahepatic cholangiocarcinoma.  Eur Radial 2008;  18:

2182-2195.

18-  Jang  JY, Kim  SW, Park  DJ,  et  al: Actual long-term outcome  of extrahepatic  bile duct cancer  after   surgical  resection.   Ann  Surg

2005; 241: 77-84.

19-  Todoroki  T, Kawamoto  T, Koike  N,  et  al: Radical resection ofhilar bile duct carcinoma and  predictors  of survival.  Br 1Surg 2000;

87: 306-313.

20- Tsao n, Nimura Y, Kamiya J, Hayakawa  N, Kondo  S, Nagino M, et al: Management  of

hilar cholangiocarcinoma: Comparison of an

American  and  a  Japanese  experience.   Ann

Surg2000; 232: 166-174.

21- Seyama Y, Makuuchi M:  Current surgical treatment   for   bile   duct   cancer.   World 1

Gastroenterol2007; 13: 1505-1515.

22- van   Gulik   TM,  Gouma   DJ:   Changing perspectives in the assessment of resectability ofhilar cholangiocarcinoma. Ann Surg Oneal

2007; 14: 1969-1971.

23-  Jarnagin  WR, Pong Y, DeMatteo  RP, et al: Staging, resectability, and outcome in 225 patients with hilar cholangiocarcinoma. Ann Surg2001; 234: 507-519.

24-  Matsuo  K,  Rocha  FG,  Ito  K,  et  al:  The Blumgart  preoperative   staging  system  for hilar            cholangiocarcinoma:    Analysis     of resectability and outcomes in 380 patients. 1

Am Coil Surg2012; 215: 343-355.

25- Cheng    Q,    Luo    X,    Zhang    B,   et    al: Predictive factors for prognosis of hilar cholangiocarcinoma: Post-resection radiotherapy  improves  survival.  Eur 1Surg Oncol2007; 33: 202-207.

26-  Wetter  LA,  Ring  EJ,  Pellegrini  CA,  Way LW: Differential  diagnosis of sclerosing cholangiocarcinomas of the common hepatic duct (Klatskin tumors). Am1Surg1991; 161:

57--62.

27- Bismuth   H,  Nakache   R,  Diamond   T: Management  strategies in resection for hilar cholangiocarcinoma.  Ann  Surg 1992;  215:

31-38.

28- Bold    RJ,    Goodnight    JE    Jr:    Hilar cholangiocarcinoma:         Surgical          and endoscopic approaches.  Surg Clin North Am

2004; 84: 525-542.

29-  Sakamoto E, Nimura Y, Hayakawa N, et al: The pattern of infiltration at the proximal border of hilar bile duct carcinoma: A histologic analysis of 62 resected cases. Ann Surg 1998; 227: 405-411.

30- Abe  M,  Kondo   S,  Hirano   S,  et  al: Superficially spreading cholangiocarcinoma. Int 1Gastrointest Cancer2005; 35: 89-94.

31-  Friman    S:   Cholangiocarcinoma:  Current

treatment  options. Scand 1Surg 2011; 100:

30-34.

32-  Lai EC, Mok FP, Fan ST, et al: Preoperative endoscopic       drainage       for      malignant obstructive  jaundice.  Br 1 Surg 1994;  81:

1195-1198.

33- Pitt HA, Gomes AS, Lois JF, et al: Does preoperative percutaneous biliary drainage reduce  operative  risk  or  increase  hospital cost? Ann Surg 1985; 201: 545-553.