DOI:10.2214/AJR.05.0008
AJR 2006; 186:1580-1586
© American Roentgen Ray Society
Preoperative Evaluation of Resectability of Klatskin Tumor with 16-MDCT Angiography and Cholangiography
H. W. Chen1,
A. Z. Pan1,
Z. J. Zhen1,
S. Y. Su1,
J. H. Wang1,
S. C. H. Yu2 and
W. Y. Lau2
1 Department of Hepatobiliary Surgery, The First People's Hospital of Fo Shan,
Fo Shan 528000, Guang Dong, People's Republic of China.
2 Department of Surgery, Prince of Wales Hospital, The Chinese University of
Hong Kong, Shatin, Hong Kong SAR, People's Republic of China.
Received January 4, 2005;
accepted after revision July 7, 2005.
Address correspondence to H. W. Chen
(chwei{at}fsyyy.com).
Abstract
OBJECTIVE. The objective of our study was to evaluate prospectively
the preoperative use of 16-MDCT angiography and cholangiography in determining
the resectability of Klatskin tumors.
CONCLUSION. Preoperative MDCT angiography and cholangiography gave a
good assessment of the degree of vascular and biliary involvement of the
Klatskin tumor.
Keywords: biliary cholangiocarcinoma • cholangiography • Klatskin tumor • MDCT angiography
Introduction
In recent years, with the invention and development of the technology of
MDCT, increasing numbers of reports have been published on its use for
preoperative assessment of liver tumors and pancreatic tumors
[1,
2]. But few reports of its use
for the preoperative assessment of Klatskin tumors are available. We conducted
a prospective study on the use of 16-MDCT angiography and cholangiography in
the preoperative assessment of the resectability of Klatskin tumors, with the
goal of evaluating its accuracy and clinical value.
Subjects and Methods
General Information
From January 2002 to June 2004, 18 patients with Klatskin tumors were
recruited for preoperative assessment with MDCT angiography and
cholangiography. The diagnosis of Klatskin tumor was subsequently confirmed in
all 18 patients with histologic examination of the surgically resected or
biopsy specimens.
Thirteen men and five women participated in the study, with an average age
of 60.6 years and an age range of 38 to 73 years. The major presenting symptom
was painless progressive jaundice occurring for 1 week to 1 month, with a mean
of 15 days. The tumors were detected with sonography or MR
cholangiopancreatogram. All patients underwent laparotomy 10 to 14 days after
CT examination. Tumor resection was performed in four patients. Resection of
the right hemiliver and caudate lobe was performed in six patients. Resection
of the left hemiliver and caudate lobe was performed in two patients.
Palliative internal drainage was performed in two patients. U-tube external
drainage was performed in three patients; metallic biliary stenting was
performed in one patient.
Equipment
The equipment used was a 16-MDCT unit; imaging was performed with tube
voltage, 120 kV; tube current, 250 mA; scanning time, 0.5 sec; slice
thickness, 2.5 mm; volumetric reconstruction interval, 1.25 mm; pitch, 1.375;
and detector group thickness, 1.25 mm.
Study Method
All patients underwent standard and sonographically guided percutaneous
transhepatic biliary drainage before surgery. An 8.5-French pigtail catheter
was used in all patients. Unilateral drainage was performed in four patients.
Bilateral drainage was performed in 14 patients. The catheters were placed at
the intrahepatic bile ducts for external biliary drainage. The CT examination
was performed first, with a plain scan of the liver, followed by a triphasic
contrast-enhanced CT at 25 sec, 45 sec, and 65 sec after an injection of 120
mL of nonionic contrast medium (300 mg I/mL) given at the antecubital vein at
a rate of 3 mL/sec, to catch the early arterial phase, late arterial-early
portal venous phase, and hepatic venous phase, respectively. The liver was
scanned again when 30-40 mL of 10% meglumine iothalamate (Conray,
Mallinckrodt) was instilled into the biliary tree through the pigtail catheter
at 5 min after the triphasic CT, with the patients holding their breath.
The data acquired were transferred to a workstation (Advantage Windows 4.1,
GE Healthcare) for reconstruction of the 3D images of the portal venous
system, hepatic arteries, and the biliary tree. The data were processed with
the techniques of volumetric reconstruction, multiplanar reconstruction, and
maximum intensity projection. Tumor resectability was assessed according to
the following criteria: the site and type of biliary obstruction, the
relationship of the tumor and blood vessels at the porta hepatis, evidence of
tumor infiltration of the second-order branch of the bilateral hepatic ducts,
evidence of tumor involvement of the proper hepatic artery together with the
left or right hepatic artery, evidence of involvement of both the right and
left branches of the portal vein, and evidence of distal metastasis or
retroperitoneal lymph node metastasis. The criteria used for diagnosing tumor
invasion of blood vessels or bile ducts was the detection of irregularity of
the endoluminal surface of the structure, straightening of the structure,
narrowing of the structure, or occlusion of the structure. The CT images were
interpreted by the consensus of three radiologists experienced in
hepatobiliary CT. The results of preoperative assessment of tumor
resectability were compared with the results of assessment at laparotomy.
Statistical analysis was performed with the kappa test.
Results
Table 1 shows the CT and
operative findings of all 18 patients.
Relationship of the Tumor to the Hepatic Artery and Portal Vein at the Porta Hepatis
Portal vein invasion was detected on MDCT angiography in 13 patients, and
these were all confirmed at laparotomy. Of the five patients in whom no portal
vein invasion was detected on MDCT angiography, four patients were confirmed
at laparotomy to have no portal vein invasion; one patient was confirmed to
have portal vein invasion. The diagnostic accuracy was 94.4% (17/18; 95%
confidence interval [CI], 72.7-99.9%). Hepatic artery invasion was detected on
MDCT angiography in 11 patients; all of these were confirmed at laparotomy. Of
the seven patients in whom no hepatic artery invasion was detected on MDCT
angiography, five patients were confirmed at laparotomy to have no hepatic
artery invasion; two patients were confirmed to have hepatic artery invasion.
The diagnostic accuracy was 88.9% (16/18; 95% CI, 65.3-98.6%).
Site and Type of Biliary Obstruction
The preoperative CT cholangiography showed that biliary obstruction had
occurred at the porta hepatis in all patients. Based on the Bismuth
classification (1975) [3] of
biliary obstruction, two patients had type I obstruction; two patients, type
II obstruction; four patients, type IIIa obstruction; two patients, type IIIb
obstruction; and eight patients, type IV obstruction. The operative findings
of the site and type of biliary obstruction were the same as the CT findings
in all patients. Diagnostic accuracy was 100% (95% CI, 81.5-100%).
Preoperative Assessment of the Resectability of Klatskin Tumor with CT
The tumors in 12 patients were considered resectable in the preoperative
assessment with CT. Of these, 11 tumors were actually resectable; therefore,
the accuracy of CT assessment of resectability was 91.7% (11/12; 95% CI,
61.5-99.8%). Of the six tumors considered to be unresectable with preoperative
CT, five were actually unresectable, and the other tumor became resectable
after the main portal vein was partially resected; therefore, the accuracy of
CT assessment of unresectability was 83.3% (5/6; 95% CI, 35.9-99.6%). When the
results of preoperative CT assessment of tumor resectability were compared
with the results of intraoperative assessment, which was used as the gold
standard, statistical analysis with the kappa test showed a kappa value of
0.784 (
>0), indicating that preoperative CT assessment had a good
correlation with the intraoperative assessment.
Discussion
The resectability of Klatskin tumors is determined by the following
factors: first, whether the second-order branch of the bilateral hepatic ducts
has been infiltrated by the tumor; second, whether the proper hepatic artery
has become involved together with the left or right hepatic arteries; and
third, whether the second-order branches of the main portal vein have become
involved simultaneously. If tumor resectability is accurately assessed
preoperatively, unnecessary operation can be avoided, and the relatively less
invasive nonoperative treatment options such as endoscopic or percutaneous
placement of biliary endoprostheses can be adopted
[4]. A direct and comprehensive
technique of clinical investigation for the preoperative assessment of
resectability of Klatskin tumor currently is still lacking. Liang et al.
[5] reported the use of color
Doppler sonography in the preoperative assessment of the relationship between
tumors and blood vessels at the porta hepatis. The diagnostic accuracy for the
relationship between tumor and portal vein was 81.82%, and the diagnostic
accuracy for the relationship between tumor and portal vein was 77.27%,
indicating that color Doppler sonography is valuable for the preoperative
assessment of the resectability of Klatskin tumors
[5]. In the last few years, MR
cholangiopancreatography (MRCP) has become a rapidly developing technology for
the noninvasive imaging of the pancreatic and biliary ductal system, replacing
endoscopic retrograde cholangiopancreatography (ERCP) and percutaneous
transhepatic cholangiography (PTC) as the first-line investigatory method for
biliary obstruction [6,
7]. The diagnostic accuracy of
MRCP in locating the site of biliary obstruction and in showing the cause of
biliary obstruction was 100% and 95%, respectively
[8]. Selective angiography is
uncommonly performed today because it is an invasive procedure requiring
arterial catheterization with a risk of complications, and it only provides an
indirect portovenogram. Although each of the investigatory techniques just
described carries its own merits, they can only evaluate the condition of an
isolated vascular or biliary system; they cannot provide information on the
relationship among the tumor, important vessels, and bile ducts at the porta
hepatis. Therefore, a proper assessment of tumor resectability requires a
combination of the results of several different investigations.
MDCT angiography is a vascular imaging technology that involves the
continuous volumetric acquisition of anatomic, pathologic, and physiologic
data by an MDCT scanner at the peak arterial phase of contrast enhancement,
computerized postacquisition data processing, and ultimately reconstruction of
a 3D image of the target vessels. The introduction of MDCT angiography not
only significantly increases the accuracy of the diagnosis and localization of
Klatskin tumor, it also provides a unique advantage in preoperative tumor
staging and assessment of tumor resectability by 3D vascular imaging
[9,
10].
Two major clinical applications of MDCT angiography have proved
valuable.
Precise Demonstration of the Relationship of the Tumor, Hepatic Artery, and Portal Vein with 3D Reconstruction of CT Angiography
In the present series, the diagnostic accuracy of MDCT angiography in the
preoperative assessment of the relationship between tumor and portal vein was
94.4%, and the diagnostic accuracy in the assessment of the relationship
between tumor and hepatic artery was 88.9%. These results were superior to
those of color Doppler sonography
[5]. In patient 1 of the
present series, MDCT angiography showed obvious evidence of tumor invasion of
the right and left branches of the portal vein and the right hepatic artery
that were of reduced caliber and narrowed, whereas the left hepatic artery was
intact (Figs. 1B and
1C). At laparotomy, tumor
invasion of the right hepatic artery, the right and left branches of the
portal vein, and portal vein bifurcation was confirmed. Right hemihepatectomy
and resection of the tumor-involved portal bifurcation were performed,
followed by end-to-end anastomosis of the left branch and the main trunk of
the portal vein. In patient 2, MDCT angiography showed that the right branch
of the portal vein and the right hepatic artery were uninvolved by the tumor;
however, the peripheral end of the left branch of the portal vein and the left
hepatic artery were irregularly narrowed with reduced caliber (Figs.
2B and
2C). At laparotomy, these
findings were confirmed. Moreover, tumor invasion of the left hepatic duct and
caudate lobe bile duct was also noted (Fig.
2A). Left hemihepatectomy and resection of the caudate lobe were
therefore performed. In patient 3, MDCT angiography showed obvious narrowing
of the left branch of the portal vein, irregularity of the anterior wall of
the right branch of the portal vein, together with narrowing and reduced
caliber of the bilateral hepatic arteries, indicating tumor invasion of all
these structures (Figs. 3B and
3C). At laparotomy, tumor
invasion of the left and right branches of the portal vein and the left and
right hepatic arteries was found. The tumor was unresectable. A U-tube was
then placed for biliary drainage. U-tube drainage is a surgical intubation
procedure indicated for those patients with unresectable tumors found at
laparotomy. The tube, consisting of a transhepatic upper limb and a subhepatic
lower limb extending out of the patient's body surface and connected to each
other in a U-shaped close loop, provides a large-bore channel for palliative
drainage of the obstructed ducts
[11,
12].
View larger version (102K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1B —Preoperative 3D reconstructed CT angiography and cholangiography of
73-year-old man. Reconstructed hepatic arteriogram showed evidence of invasion
of right hepatic artery (RHA) with reduced caliber and narrowed lumen.
|
|
View larger version (109K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1C —Preoperative 3D reconstructed CT angiography and cholangiography of
73-year-old man. Reconstructed portovenogram showed tumor located superior to
portal bifurcation. Evidence of tumor invasion of right branch of portal vein
(right portal vein [RPV]) was noted, with truncation and disappearance of its
peripheral end. Tumor invasion was also noted at anterior wall of origin of
left branch of portal vein (left portal vein [LPV]). This finding was
confirmed at surgery.
|
|
View larger version (79K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2B —Preoperative 3D reconstructed CT angiography and cholangiography of
69-year-old man. Reconstructed hepatic arteriogram showed normal proper
hepatic artery and right hepatic artery (RHA) (arrow). Peripheral
segment of left hepatic artery (LHA) was of reduced caliber and narrowed
(arrow).
|
|
View larger version (96K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2C —Preoperative 3D reconstructed CT angiography and cholangiography of
69-year-old man. Reconstructed portovenogram showed normal main trunk and
right branch of portal vein (RPV) (arrow). Peripheral end of left
branch of portal vein was irregularly narrowed (LPV) (arrow).
|
|
View larger version (64K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 2A —Preoperative 3D reconstructed CT angiography and cholangiography of
69-year-old man. Catheter cholangiogram of right biliary ductal system showed
evidence of dilatation of right intrahepatic bile ducts. There was contrast
filling defect at confluence of right and left hepatic ducts and common
hepatic duct (arrow). Filling defect was more obvious on left side.
Left hepatic duct was not outlined with contrast agent but common hepatic duct
was outlined, indicating porta hepatis tumor was located at junction between
common hepatic duct and left hepatic duct.
|
|
View larger version (107K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3B —Preoperative 3D reconstructed CT angiography and cholangiography of
38-year-old woman. Reconstructed hepatic arteriogram showed tumor invasion of
left (LHA) and right (RHA) hepatic arteries that were of reduced caliber and
narrowed (arrows).
|
|
View larger version (94K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3C —Preoperative 3D reconstructed CT angiography and cholangiography of
38-year-old woman. Reconstructed portovenogram showed evidence of tumor
encasement and occlusion of left branch of portal vein (LPV) (arrow)
and tumor invasion of anterior wall of right branch of portal vein (RPV), with
mild surface irregularity (arrow).
|
|
View larger version (117K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 1A —Preoperative 3D reconstructed CT angiography and cholangiography of
73-year-old man. Catheter cholangiogram of bilateral system performed with
percutaneous transhepatic biliary drainage catheter showed moderate dilatation
of left (LHD) and right (RHD) hepatic ducts. Underfilling was noted at
confluence of intrahepatic ducts and common bile duct (CBD). RHD was more
obviously affected and probably completely occluded. Obvious narrowing of LHD
was noted at confluence, although contrast agent could still pass into common
hepatic duct.
|
|
View larger version (92K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 3A —Preoperative 3D reconstructed CT angiography and cholangiography of
38-year-old woman. Catheter cholangiogram with pigtail catheters placed within
bilateral system showed dilatation of left (LHD) and right (RHD) hepatic bile
ducts, with left ducts more severely affected. There was underfilling at
confluence of LHD and RHD and common bile duct (CBD). Tumor invasion was noted
in right and left hepatic ducts (arrows), although contrast agent can
still partially pass into common hepatic duct. GB = gallbladder.
|
|
View larger version (99K):
[in this window]
[in a new window]
[as a PowerPoint slide]
|
Fig. 4A —Cholangiograms of 80-year-old man. Preoperative CT cholangiogram
showed stenosis of common hepatic duct (CHD) and proximal segment of posterior
right hepatic duct (PRHD), signifying tumor invasion at these sites and
Bismuth classification IIIa. ARHD = anterior right hepatic duct; CBD = common
bile duct; GB = gallbladder.
|
|
Precise Demonstration of the Site and Extent of Biliary Obstruction by CT Catheter Cholangiogram and 3D Reconstruction of the Biliary Tree
Three-dimensional biliary tree imaging allows accurate preoperative Bismuth
staging and provides important information for planning the approach to
biliary drainage. In the present series, the accuracy of preoperative staging
was 100%. The authors found that CT cholangiograms have the advantages of high
spatial resolution, precise visualization of anatomic structures, and
multidirectional assessment of biliary branches, and therefore they provide
important information for localization and staging of the tumor (Figs.
4A,
4B
and4C). In patient 1, CT
cholangiogram clearly showed that the biliary tumor at the porta hepatis was
located at the right hepatic duct and common hepatic duct, with the origin of
the second-order branch of the left hepatic duct uninvolved
(Fig. 1A). Based on this
information, a right hemihepatectomy was planned preoperatively to preserve
the left hemiliver. The preoperative findings were confirmed
intraoperatively.
In patient 2, CT cholangiogram showed that the tumor at the porta hepatis
was mainly located at the left hepatic duct and common hepatic duct, with the
origin of the second-order branch of the right hepatic duct uninvolved. Left
hemihepatectomy was planned preoperatively. The findings were confirmed at
laparotomy.
The proportion of circumferential contact of the tumor with a blood vessel
or bile duct is important information and should have been included in the
diagnostic criteria of tumor invasion. It was found retrospectively that
neglecting circumferential contact by the tumor was a major cause of error
leading to the underdiagnosis of tumor invasion, and the structure is most
probably invaded if it is surrounded by tumor for half or more of its
circumference.
In conclusion, 3D MDCT angiography and cholangiography are useful
diagnostic tools for delineation of the relationship between the tumor and the
important vessels at the porta hepatis and for tumor staging. They therefore
provide important information on the preoperative assessment of tumor
resectability and the extent of surgery. It is believed that this
investigatory technique shows the relationship of the tumor, blood vessels,
and bile ducts with a higher definition and a more direct and comprehensive
perspective when compared with the other currently available investigatory
methods and therefore deserves a wider clinical application.
References
- Ling H, Guan Y, Ding B, Lin X, Zhang H, Chen K. Vascular
involvement in pancreatic carcinoma: preoperative assessment by multislice CT
angiography. Clin J Radiol 2002;36
: 609-612
- Wigmore SJ, Redhead DN, Yan XJ, et al. Virtual hepatic resection
using three dimensional reconstruction of helical computed tomography
angioportograms. Ann Surg 2001;233
: 221-226[CrossRef][Medline]
- Bismuth H, Corlette MB. Intrahepatic cholangioenteric anastomosis
in carcinoma of the hilus of the liver. Surg Gynecol
Obstet 1975; 140:170
-178[Medline]
- Nimura Y, Kamiya J, Kondo S, et al. Aggressive preoperative
management and extended surgery for hilar cholangiocarcinoma: Nagoya
experience. J Hepatobiliary Pancreat Surg2000; 7:155
-162[CrossRef][Medline]
- Liang P, Hao FM, Zhou NX, et al. Preoperative evaluation of
resectability of hilar bile duct carcinoma by color Doppler flow imaging and
duplex Doppler. Clin J Surg 1994;32
: 259-261
- Manfredi R, Masselli G, Maresca G, et al. MR imaging and MRCP of
hilar cholangiocarcinoma. Abdom Imaging2003; 28:319
-325[CrossRef][Medline]
- Vaishali MD, Agarwal AK, Upadhyaya DN, et al. Magnetic resonance
cholangiopancreatography in obstructive jaundice. J Clin
Gastroenterol 2004; 38:887
-890[Medline]
- Zhou JX, Liang BL, Xu LY, Huang SQ. MR cholangiopancreatography and
MR imaging in the diagnosis of extrahepatic cholangiocarcinoma.
Clin J Oncol 2004;26
: 421-423
- Tillich M, Mischiroger HJ. Preisegger KH, et al. Multiphasic
helical CT in diagnosis and staging of hilar cholangiocarcinoma.
AJR 1998; 171:651
-658[Abstract/Free Full Text]
- Cha JH, Han JK, Kim TK. Preoperative evaluation of Klatskin tumor:
accuracy of spiral CT in determining vascular invasion as a sign of
unresectability. Abdom Imaging 2000;25
: 500-507[CrossRef][Medline]
- Terblanche J, Saunders SJ, Louw JH. Prolonged palliation in
carcinoma of the main hepatic duct junction. Surgery1972; 71:720
-731[Medline]
- Krige JEJ, Terblanche J. Dilatation and transhepatic intubation:
the U tube. In: Terblanche J, ed. Hepatobiliary malignancy: its
multidisciplinary management. London, UK: Edward Arnold,1994
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Top
Abstract
Introduction
