Staging of hilar cholangiocarcinoma by ultrasound ... - BIR Publications

1992, The British Journal of Radiology, 65, 871-877

Staging of hilar cholangiocarcinoma by ultrasound and duplex sonography: a comparison with angiography and operative findings By 'Christian Looser, MD, ttSteven C. Stain, MD, tHans U. Baer, MD, *Jiirgen Triller, MD and t§Leslie H. Blumgart, MD 'Institute for Diagnostic Radiology and tClinic for Visceral and Transplantation Surgery, Inselspital, University of Berne, CH-3010 Berne, Switzerland {Received 6 January 1992 and accepted 6 May 1992) Keywords: Proximal bile duct tumours, Cholangiocarcinoma, Klatskin tumours, Staging, Ultrasound, Angiography, Duplex sonography

Abstract. The pre-operative radiological assessment of proximal bile duct tumours is clinically important as resection may be limited by tumour extension along the bile ducts, into hepatic parenchyma or the adjacent vascular structures. Demonstration of the extent of biliary and vascular involvement can direct additional investigations and definitive treatment. 22 patients with hilar cholangiocarcinoma were studied pre-operatively by conventional ultrasound (US) and duplex sonography (DS). The extent of tumour infiltration and vascular involvement was compared with arteriography and operative findings. Bile duct dilatation and the level of obstruction was documented by US in 22 (100%), and the tumour was shown by US in 19 (86%). In these 19 patients, the extent of extraductal extension compared with operative findings was correct in 13, underestimated in two, and in four infiltration was massed. Vascular patency or involvement was correctly determined by DS in 19 (86%), and by arteriography in 18 (82%). In two of the three incorrect DS interpretations, lobar atrophy and contralateral hypertrophy distorted the hilar anatomy. US with DS is valuable in the pre-operative staging of proximal bile duct tumours in predicting ductal and vascular involvement.

Adenocarcinomas of the proximal bile ducts are rare tumours first reported by Altemeier et al in 1957, and have been associated with the eponym Klatskin tumours (Klatskin, 1965). In the United States, it is estimated that hepatobiliary tumours comprise 5% of newly diagnosed tumours in the gastrointestinal tract (Kreger et al, 1991). The autopsy incidence of proximal cholangiocarcinomas is 0.01-0.46% (Kirschbaum & Kosoll, 1941; Kuwayti et al, 1957; Sako et al, 1975). Typically, hilar cholangiocarcinomas are small with diameters of 1-3 cm. The extent and exact location within the hilum may be difficult to determine at operation, therefore it is important to have pre-operative staging (Voyles et al, 1983). Modern operative techniques combined with newer methods of imaging have permitted resection in up to 40% of patients in specialized hepatobiliary centres (Baer et al, unpublished data). This study was performed to assess the value of ultrasound (US) and duplex sonography (DS) in the pre-operative staging of proximal hilar cholangiocarcinoma. While US has been reported useful in the staging of proximal bile duct tumours, the use of DS to deterAddress correspondence to Jiirgen Triller, MD. |Present address: Department of Surgery, Rm 9900, LAC-USC Med. Center, 1200 N. State Street, Los Angeles, CA 90033, USA. §Present address: Department of Surgery, Memorial Sloan-Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA. Vol. 65, No. 778

mine vascular involvement has not been extensively studied. Materials and methods

Between 1986 and 1989, 22 patients (13 male, 9 female) with hilar cholangiocarcinoma were assessed pre-operatively by abdominal US with DS. The mean age was 61 years (range 41-75 years). The sonographic staging was performed with the Toshiba SSH 100 and Acuson 128 ultrasound machines, both of which had a pulsed duplex unit, and were equipped with sector and curved array sonographic heads of 3.5 MHz and 5.0 MHz. DS identified the hepatic artery, portal vein and their primary intrahepatic branches in all patients. In all patients, a selective arteriogram of the superior mesenteric artery and coeliac trunk was performed. The examinations were initially performed with conventional techniques using a 5.5 F Cobra catheter and non-ionic contrast medium (Omnipaque 350). If necessary, digital subtraction angiography was used. With the conventional technique, 60 ml of contrast medium was injected with a flow of 5-7 ml/s. The contrast medium was diluted 1:1 with normal saline for digital subtraction. Percutaneous transhepatic cholangiography was performed in 10 of the 22 patients. Arteriography was performed after the other pre-operative investigations (US, DS, computed tomography (CT), percutaneous transhepatic cholangiography (PTC), endoscopic retrograde cholangiopancreatography (ERCP)) and the angiographer was blind to the US and DS results. All 871

C. Looser, S. C. Stain, H. U. Baer, J. Triller and L. H. Blumgart

patients were operated on shortly after the radiological investigations (within 7 and 10 days). The US and DS results were correlated with the arteriographic and operative findings. Results Ultrasound

Dilated bile ducts and the level of obstruction were identified by US in all 22 patients. 20 tumours involved the hepatic confluence, and two were located in the common hepatic duct. US identified a tumour mass in 19 of the 22 patients (86%). The tumour was not identified by US in three patients. In one of these patients, the tumour was intramural in a circular fashion, which did not produce a discrete mass. The second patient had perihepatic adhesions from a previous cholecystectomy, which caused a poor sonographic signal, and the final patient had excessive artefact from a previously placed biliary stent. Eight of the 19 sonographically identified tumours were isoechoic (42%) in comparison to the liver parenchyma. Eight were hyperechoic (42%), and three were hypoechoic (16%). The mean tumour diameter was 3 cm (0.5-6 cm). The patients were classified by operative findings. Group 1: Tumour growth confined to the bile duct wall or endolumen (n = 10). The sonographic criteria of the endoluminal tumours were proof of an intraductal mass lesion with clear demarcation of outline distinct from the bile duct wall. The US result was accurate in eight patients (80%) (Figs 1,2). The tumour mass was not identified in one patient. In an additional patient the tumour was visualized, but the extent of diffuse intramural growth was not seen. In this patient, US identified hepatic metastases, which allowed pre-operative planning of a palliative procedure.

Figure 1. Proximal cholangiocarcinoma (arrows) with only endoluminal extension into the left hepatic duct (LI). The right hepatic duct is free of tumour (RE). No evidence of transmural extension at the level of the common hepatic duct (CHD). 872

Figure 2. Endoluminal tumour (arrows) at the confluence of the right hepatic duct (RHD) and left hepatic duct (LHD). The hepatic artery (HA) and portal vein (VP) are not involved.

Group 2: Extraductal infiltration into the surrounding tissue (n = 12). This group was further divided into tumour growth that extended into the hepatic parenchyma (« = 9) or into the hepatoduodenal ligament (» = 3). Identification of extraductal tissue was regarded as tumour. Criteria of the extraductal tumour infiltration were heterogeneity of the normal bile duct wall structure with spreading of isoechogenic tumour mass contiguous with the periportal fat tissue adjacent to the liver parenchyma. Loss of demarcation of the extraductal tumour tissue of the liver parenchyma was regarded as a direct infiltration and not as tumour-related mass effect. The infiltration of the liver tissue was correctly described in four of nine patients (45%) (Fig. 3). The depth of invasion was underestimated in two (22%) and

Figure 3. Large cholangiocarcinoma (TU) extending into segment IV with infiltration along the bile ducts beyond the bifurcation (arrow) of the right hepatic ducts (RHD) and left hepatic duct (LHD). The British Journal of Radiology, October 1992

Staging of hilar cholangiocarcinoma

Figure 4. Intraluminal tumour (large arrows) is seen common hepatic duct (CHD) located anterior to the vein (PV) and hepatic artery (HA). Enlarged regional node (small arrows) anterior to the inferior vena cava histology showed reactive hyperplasia.

in the portal lymph (IVC);

the positive predictive value of tumour infiltration if the tumour can be visualized was 77% (15 of 19 patients). There were no US demonstrations of parenchymal or hepatoduodenal ligament infiltration not confirmed at operation. Only one patient had liver metastases identified at operation. These had been identified preoperatively by US. Irregular nodular structure of an enlarged lymph node was regarded as representing tumour. Seven patients had enlarged hilar lymph nodes at operation. Three of these nodes showed reactive hyperplasia, and four were proven histologically to be metastatic. US visualized two of the seven as being enlarged (Fig. 4). In 16 out of 22 patients (72%) the extent of intraluminal tumour spread into the right or left hepatic duct was judged correctly. Contiguous infiltration into subsegmental ducts was not assessed sonographically, as unilateral intrahepatic parenchymal infiltration was an operative indication for an extended hepatic resection. Duplex sonography

missed in three (33%). In two of the latter patients, the tumour was not shown. The sonographic criteria for invasion of the hepatoduodenal ligament was an isoechogenic structure extending from the tumour or longitudinal spread along the confines of the hepatoduodenal ligament with obliteration of the ligamentous fat stripe. Invasion of the hepatoduodenal ligament was accurately visualized in one of three patients (33%) (Fig. 4); the tumour extension was missed in the other two (67%). In summary, US correctly identified the depth of local tumour extension in 13 of 22 patients (59%) and in 13 of 19 patients in whom the tumour was seen (68%). By inclusion of the additional two patients in whom parenchymal infiltration was identified but underestimated,

DS was generally performed in the lateral position for longitudinal examination of the hepatic and portal vessels. All 22 patients also had hepatic arteriography and late-phase portography. For extrahepatic tumours, vascular stenosis increased the medianflowspeed(jet stream phenomenon) by comparison with the pre-stenotic areas. Post-stenotic turbulences were identified. The medium velocities in the segmental branches of the portal vein within intrahepatic tumours were significantly lowered by comparison with the segmental branches not involved with tumour. Post-stenotic turbulence or jet stream phenomenon at the level of the stenosis cannot be seen intrahepatically because venous blood flow in the segmental branches of the portal vein is too low. Intrahepatic portal vein occlusion can be identified because of the missing blood

(a)

(b)

Figure 5. Proximal cholangiocarcinoma with stenosis and encasement of the portal vein, (a) DS shows jet phenomenon at the level of the stenosis with turbulence and increased central hepatopedal flow, and elevated median venous flow velocity. High-grade stenosis (arrows) of the portal vein (PV). (b) Portal venography after coeliac trunk injection reveals stenosis (arrow) of the main portal vein below the hilum. A biliary endoprosthesis is in place (e). Vol. 65, No. 778

873


(a) (b) Figure 6. Left portal vein occlusion due to tumour infiltration with subsequent left lobar atrophy, (a) DS demonstrates absent flow in the left portal vein (large arrow) due to tumour infiltration (small arrows), (b) Portal venography after coeliac trunk injection shows occlusion just beyond the origin of the left portal vein (arrow).

flow signal distal to the tumour with a Doppler angle of less than 60°. In 13 patients, there was no tumour involvement by DS. Arteriography and portography were also normal. These patients had no vascular involvement at operation. Nine patients had vascular involvement confirmed by operation; duplex accurately described six of these. In two of the six, there was extrahepatic portal vein stenosis, also identified by portal venography (Fig. 5). DS identified two patients with unilateral portal vein occlusion also identified at portography and confirmed at operation (Fig. 6). DS correctly identified two stenoses of a unilateral intrahepatic portal vein branch confirmed at operation. One was missed by angiography, and was misinterpreted as an occlusion. The duplex findings were incorrect in three patients. Patient 1. DS of the hepatic artery and the portal vein

were thought to be normal. Hepatic arteriography revealed hepatic artery encasement. Portal venography was interpreted as normal. At operation, localized tumour encasement of the portal vein and proper hepatic artery caudal to the hilar plate was found. Patient 2. DS showed a normal hepatic artery and a stenosis of the left branch of the portal vein. Arteriography and portal venography were normal. At operation, the apparent compression was due to right lobe atrophy with compensatory left lobar hypertrophy. Patient 3. DS visualized a normal hepatic artery and what was thought to be a normal portal vein. The hepatic arteriogram and portal venography were also normal. There was marked left lobar atrophy which distorted the hilum. The bile duct could not be dissected free from the involved portal vein and a palliative biliary enteric bypass was performed. Vascular involvement was accurately predicted by DS

Table I. Results of duplex sonography Duplex sonography

Arteriography

Operative findings

HA and PV normal (13)



PV stenosis (2) PV branch occlusion (2) PV branch stenosis (2)

PV stenosis (2) PV branch occlusion (2) *Missed by angiography (1) *Occlusion (1)

PV stenosis (2) PV branch occlusion (2) PV branch stenosis (2)

*HA normal; *PV normal (1) •Left PV stenosis; HA normal (1) *PV normal; HA normal (1)

*PV normal; HA encasement (1) HA and PV normal (1) *PV normal; HA normal (1)

HA and PV encasement HA and PV normal HA normal; PV involved

Columns refer to radiological or operative findings. HA, hepatic artery; PV, portal vein. ""Incorrect radiological diagnosis of either the hepatic artery or portal vein. Numbers in parentheses refer to numbers of patients. 874

The British Journal of Radiology, October 1992

Staging of hilar cholangiocarcinoma

in 19-patients (86%), and by arteriography in 18 of the 22 patients (82%) (Table I). In two of the three failures by DS, the atrophy with hypertrophy produced hilar distortion, which prevented accurate delineation of the anatomy. 11 of the 22 patients (50%) underwent local tumour excision (Voyles et al, 1983) or hepatic resection (Kirshbaum & Kosoll, 1941). Palliative biliary enteric bypass decompression was performed in 11 (50%). Discussion

In the early 1980s, PTC, hepatic arteriography and portal venography were regarded as the gold standards of imaging proximal bile duct tumours (Gibbons et al, 1983; Voyles et al, 1983; Beazley et al, 1984). US was used to demonstrate dilated intrahepatic ducts, and is still the first radiological investigation employed in the assessment of obstructive jaundice (Thomas et al, 1982; Scharschmidt et al, 1983; Saint Marc Girardin et al, 1985). In the clinical staging of bile duct tumours, US will demonstrate dilated intra- and extrahepatic bile ducts, as well the level of obstruction. In the current series, this was possible in 100% of the cases. The diagnosis of a bile duct lesion or a hilar mass as the cause of obstruction was made in 86% of our patients. Gibson et al (1986) and Yeung et al (1988) reported similar accuracies of 88% and 86%, respectively. The high sensitivity is possible due to the high resolution of modern ultrasound equipment. Tumours localized to the bile ducts were found intraoperatively in 10 patients. US identified tumour limited to endoluminal or mural bile duct involvement in eight patients (80%) (Figs 1, 2). These tumours, if confined to the hilar area, are more likely to be resectable by local hilar resection. Sclerosing tumours that infiltrate into the surrounding tissue or along the bile ducts are more difficult to assess sonographically and to remove surgically. Tumours growing extraductally can infiltrate into the hepatoduodenal ligament or into the liver parenchyma. Infiltration of the hepatoduodenal ligament was identified by US in only one of three patients. If a small part of the tumour lies extramurally, this often cannot be distinguished from hilar lymph nodes which may or may not be enlarged. The slow-growing nature of the lesions may not produce a hypoechoic rim of surrounding oedema, which can discern tumour infiltration from the surrounding structures. The identification of spread into the peritumoral inflammatory reaction is not sonographically possible. Infiltration into the liver parenchyma was identified in six of nine patients, although underestimated in two. This finding may be due to the echogenicity of the tumours. Nearly half of the tumours were isoechoic, but sonomorphological structure was different compared with the surrounding liver tissue, and the delineation of the parenchymal invasion of an isoechoic tumour is difficult if a hypoechoic rim of oedema is not present. If the tumour is large, the primary site of origin and depth Vol. 65, No. 778

of infiltration into the surrounding tissue is difficult to recognize sonographically (Fig. 3). The differential diagnosis should include hepatic metastases, or primary hepatocellular carcinoma invading the hilum. We attribute little clinical significance to sonographically identified lymph nodes. US is poor in this regard, as only two of seven enlarged nodes were seen by US (Fig. 4). Even at operation, only four of seven actually contained metastatic tumour. Haematogenous metastases from cholangiocarcinoma are unusual (Kirshbaum & Kosoll, 1941; Kuwayti et al, 1952). Only one patient was found to have liver metastases, which were documented by US, allowing planning of a palliative procedure. Conventional US can provide an accurate demonstration of bile duct dilatation and of the level of obstruction (Baron et al, 1982; Machan et al, 1986; Karstrup, 1988). The sensitivity of tumour identification is also reasonably high, but the accuracy of US in visualization of the ductal extension of tumour or extraductal infiltration is less clear. US has been used to predict the proximal tumour extent by involvement of ducts above the secondary bifurcations. Gibson et al (1986) found US to be more accurate than CT in this regard (32% versus 0%), but both modalities were poor by comparison with PTC. Here, US correctly predicted the extent of intrahepatic ductal spread in only 70% of patients. We therefore believe direct cholangiography is usually necessary to identify proximal tumour extent. In the absence of metastases, another obstacle to successful resection is vascular involvement. Modern operative techniques have permitted major hepatic resection for hilar tumours with low morbidity and mortality (Langer et al, 1985; Bengmark et al, 1988; Pinson & Rossi, 1988). Vascular reconstruction may be employed in selected cases (Tsuzuki et al, 1983; Blumgart et al, 1984; Sakaguchi & Nakamura, 1986). Our one perioperative death occurred after Hgation of the hepatic artery and subsequent necrosis. Arterial ligation is poorly tolerated by the jaundiced liver, and we now would recommend arterial reconstruction if such a resection is undertaken (Thomas et al, 1990). If ductal infiltration is unilateral, the limits of resectability depend on the state of the hepatic and protal vasculature. Our criteria for unresectability are: 1. Occlusion of the main portal vein or hepatic artery (except in unusual circumstances). 2. Bilateral involvement or primary hepatic arterial or portal venous branches. 3. Combination of unilateral vascular involvement with cholangiographic evidence of extensive contralateral spread. 4. Hepatic or nodal metastases. The sonographic accuracy of predicting resectability was 60%. If DS can reliably predict unresectability based on the involvement of the hepatic and portal vessels, arteriography would not be necessary in the clearly unresectable patients. In our initial experience in proximal bile duct tumours, we believe DS to be fairly 875


accurate (86%), and comparable to arteriography in the assessment of hilar vessels (Figs 5, 6). It is difficult by angiography or at operation to distinguish compression from actual tumour extension, which will not permit dissection from the adjacent vessels. In the two patients in whom vascular involvement was missed by DS, arteriography was also incorrect. The one false positive duplex sonogram of portal vein stenosis was due to compression, but from the contralateral hypertrophied lobe and not from tumour extension. In vascular encasement by tumour without stenosis, DS does have an advantage over the grey-scale scan, as there is no visible change of the flow velocity or flow profile. Close to the vessel wall, inhomogeneity offlowis to be expected, caused by the loss of elasticity of the vessel wall adjacent to the tumour. It is possible that colour-flow Doppler sonography may be more accurate. Two of the three incorrect diagnoses were made when lobar atrophy and contralateral hypertrophy was present. The atrophy-hypertrophy complex is most often caused by long-standing unilateral bile duct or portal vein obstruction (Longmire & Tompkins, 1975; Czerniak et al, 1986). In Klatskin tumours, the latter is more common. The compensatory hypertrophy causes an axial rotation of the portal triad, which makes radiological imaging and operative resection difficult. The presence of an atrophy/hypertrophy complex indicates a marked progression of the local tumour growth (Hadjis et al, 1989), and careful evaluation of the affected portal venous structures with both DS and angiography is necessary. DS is important in the demonstration of intrahepatic branches of the portal vein as distinct from dilated bile ducts, which may be difficult with grey-scale imaging alone. DS gives information about relevant perfusion of the two liver lobes independently of tumour spread. Conclusions

In the staging of proximal cholangiocarcinomas, the advantages of US and Doppler US are as follows. The level of the biliary obstruction can clearly be determined. Ductal tumour spread is underestimated in 30% of cases. Direct cholangiography is still necessary. Peritumoral infiltration is seen, but may be underestimated, especially extrahepatic extension along the hepatoduodenal ligament. Recognition of enlarged local lymph nodes is meaningless, since tumour cannot be distinguished from hyperplasia. Final assessment of resectability should be performed intraoperatively, if criteria for unresectability are not clearly identified pre-operatively. The addition of DS to conventional US is valuable. DS is equal to angiography in the identification of vascular involvement. DS may assist in the selection of patients for potential resection and, more importantly, non-invasively identify those patients clearly unresectable who might be considered as candidates for nonoperative palliation (Carrasco et al, 1985; Irving, 1989; Neuhaus et al, 1991). 876

References ALTEMEIER, W. A., GALL,

E. A., ZINNINGER,

M.

M. &

HOXWORTH, P. I., 1957. Sclerosing carcinoma of the major intrahepatic ducts. Archives of Surgery, 75, 450-460. BARON, R. L., STANLEY, R. J., LEE, J. K. T., KOEHLER, R. E., MELSON, G. L., BALFE, D. M. & WEYMAN, P. J., 1982. A

prospective comparison of the evaluation of biliary and obstruction using computed tomography ultrasonography. Radiology, 145, 91-98. BEAZLEY, R. M., HADJIS, N., BENJAMIN, I. S. & BLUMGART,

L. H., 1984. Idiopathological aspects of high bile duct cancer. Experience with resection and bypass surgical treatments. Annals of Surgery, 199, 623-636. BENGMARK, S., ECKBERG, H., EVANDER, A., KLOFVER-STAHL, B.

& TRANBERG, K. G., 1988. Major liver resection for hilar cholangiocarcinoma. Annals of Surgery, 207, 120-125. BLUMGART, L. H., BENJAMIN, I. S., HADJIS, N. S. & BEAZLEY,

R. T., 1984. Surgical approaches to cholangiocarcinoma at the confluence of the hepatic ducts. Lancet, i, 66-70. CARRASCO, C. H., WALLACE, S., CHARNSANGAVEL, C , RICHLI, W., WRIGHT, K. C , FANNING, T. & GIANTURCO, C , 1985.

Expandable biliary endoprosthesis: an experimental study. American Journal of Roentgenology, 145, 1279-1281. CZERNIAK, A., SOREIDE, O., GIBSON, R. N., HADJIS, N. S., KELLEY, C. J., BENJAMIN, I. S. & BLUMGART, L. H., 1986.

Liver atrophy complicating benign bile duct strictures. Surgical and interventional radiologic approaches. American Journal of Surgery, 152, 294-300. GIBBONS, C. P., GRIFFITHS, G. J. & CORMACK, A., 1983. The

role of percutaneous transhepatic cholangiography and greyscale ultrasound in the investigation and treatment of bile duct obstruction. British Journal of Surgery, 70, 404-496. GIBSON, R. N., YEUNG, E., THOMPSON, J. N., CARR, D. H., HEMINGWAY, A. P., BRADPIECE, H. A., BENJAMIN, I. S.,

BLUMGART, L. H. & ALLISON, D. H., 1986. Bile duct

obstruction: radiologic evaluation of the level, cause and tumor resectability. Radiology, 160, 43-47. HADJIS, N. S., ADAM, A. A., GIBSON, R., BLENKH/RN, J. I., BENJAMIN, I. S. & BLUMGART, L. H., 1989. Nonoperative

approach to hilar cancer determined by the atrophyhypertrophy complex. American Journal of Surgery, 157, 395-399. IRVING, J. D., ADAM, A. A., DICK, R., DONDELINGER, R. F.,

LUNDERQUIST, A. & ROCHE, A., 1989. Gianturco expandable

metallic biliary stents: results of a European clinical trial. Radiology, 172, 321-326. KARSTRUP, S., 1988. Ultrasound diagnosis of cholangiocarcinoma at the confluence of the hepatic ducts (Klatskin tumours). British Journal of Radiology, 61, 987-990. KIRSHBAUM, J. D. & KOSOLL, D. D., 1941. Carcinoma of the

gallbladder and extrahepatic bile ducts. Surgery, Gynecology and Obstetrics, 93, 740-754. KLATSKIN, G., 1965. Adenocarcinoma of the hepatic ducts at its bifurcation within the porta hepatis. American Journal of Medicine, 38, 241-256. KREGER, B. E., SPLANSKY, G. L. & SCHATZKIN, A., 1991. The

cancer experience in the Framingham Heart Study Cohort. Cancer, 67, 1-6. KUWAYTI,

K.,

BAGENSTOSS, A.

H.,

STAUFFER,

M.

H. &

PRIESTLEY, J. T., 1957. Carcinoma of the major intrahepatic and the extrahepatic bile ducts exclusive of the papilla of Vater. Surgery, Gynecology and Obstetrics, 104, 357-366. LANGER, J. C , LANGER, B., TAYLOR, B. R., ZELDIN, R. &

CUMMINGS, B., 1985. Carcinoma of the extrahepatic bile ducts: results of an aggressive surgical approach. Surgery, 98, 752-759. LONGMIRE, W. P. JR & TOMPKINS, R. K., 1975. Lesions of the

The British Journal of Radiology, October 1992

Staging of hilar cholangiocarcinoma segmental and lobar hepatic ducts. Annals of Surgery, 182, 478-495. MACHAN,

L.,

MULLER,

L. & COOPERBERG, P.

L.,

1986.

Sonographic diagnosis of Klatskin tumors. American Journal of Roentgenology, 147, 509-512. NEUHAUS, H., HAGENMULLER, E., GRIEBEL, M. & CLASSEN, M.,

1991. Percutaneous cholangioscopic or transpapillary insertion of self-expanding biliary metal stents. Gastrointestinal Endoscopy, 37, 31-37. PINSON, C. W. & Rossi, R. L., 1988. Extended right hepatic lobectomy, left hepatic lobectomy, and skeletonization resection for proximal bile duct cancer. World Journal of Surgery, 12, 52-59. SAINT MARC GIRARDIN, M. F., LEMINOR, M., ALPEROVITCH, A., ROUDOT-THOROVAL, F., METREAU, J. M. & DHUMEAUX,

D., 1985. Computer-aided selection of diagnostic tests in jaundiced patients. Gut, 26, 961-967. SAKAGUCHI, S. & NAKAMURA, N., 1986. Surgery of the portal

vein in resection of cancer of the hepatic hilus. Surgery, 99, 344-349. SAKO, S., SEITZINGER, G. L. & GARSIDE, E., 1957. Carcinoma of

SCHARSCHMIDT, B. F . , GOLDBERG, H . I. & SCHMID, R., 1983.

Approach to the patient with cholestatic jaundice. New England Journal of Medicine, 308, 1515-1519. THOMAS, M. J., PELLEGRINI, C. A. & WAY, L. W., 1982.

Usefulness of diagnostic tests for biliary American Journal of Surgery, 144, 102-108.

obstruction.

THOMAS, P. G., BAER, H. U., MATTHEWS, J. B., GERTSCH, P. &

BLUMGART, L. H., 1990. Post-operative hepatic necrosis due to reduction in arterial flow during surgery for chronic biliary obstruction. Digestive Surgery, 7, 31-35. TSUZUKI, T., OGATA, Y., IIDA, S., NAKANISHI, T., TAKENAKA, Y.

& YOSHII, H., 1983. Carcinoma of the bifurcation of the hepatic ducts. Archives of Surgery, 118, 1147-1151. VOYLES, C. R., BOWLEY, N. J., ALLISON, D. J., BENJAMIN, I. S.

& BLUMGART, L. H., 1983. Carcinoma of the extrahepatic biliary tree. Radiologic assessment and therapeutic alternatives. Annals of Surgery, 197, 188-194. YEUNG, E. Y. C , MCCARTHY, P., GOMPERTZ, R. H., BENJAMIN, I. S., GIBSON, R. N. & DAWSON, P., 1988. The radiologic

appearance of hilar cholangiocarcinoma (Klatskin tumours). British Journal of Radiology, 61, 991-995.

the extrahepatic bile ducts: review of the literature and report of six cases. Surgery, 41, 416—437.

Vol. 65, No. 778

877