Hilar tumours

Tumours that involve the confluence of the bile ducts in the liver hilum provide a major therapeutic challenge. Adequate palliation requires relief of jaundice. Even though jaundice can be relieved if only 30% of the liver is drained, the presence of undrained bile ducts may result in pruritus and a continued risk of cholangitis and hepatic abscess. The biliary anatomy is defined by cholangiography, which today can be performed by magnetic resonance; the stricture is dilated, and plastic or metallic stents are endoscopically inserted over a guidewire. Patients with complex hilar strictures may benefit from the insertion of one or more stents, although there is debate about how many are necessary. The present article provides specific technical details, describes comparative trials of unilateral versus bilateral biliary drainage and explores new techniques that warrant further investigation.

T he management of malignant tumours involving the main hepatic confluence ('hilar tumours') remains a major clinical issue despite significant progress in the past few years. The bile ducts can be involved by primary cancer (cholangiocarcinoma or 'Klatskin tumour'), contiguous spread from gallbladder, pancreatic or hepatocellular carcinoma, or lymph node metastases from distant tumours. One should note that the proper term is 'hepatic confluence' and not 'bifurcation' of the bile ducts in the hilum, which reflects the direction of bile flow.
Only a minority of patients are candidates for surgical resection at the time of diagnosis (eg, 15% to 40% of those with cholangiocarcinoma), mainly because of the presence of either vascular invasion or metastases. Palliation of biliary or intestinal obstruction and pain has, therefore, become a major therapeutic goal. Achievement of palliation requires a thorough understanding of hepatobiliary anatomy ( Figure 1) as well as localization and staging of the tumour. The Bismuth-Corlette classification of hilar cholangiocarcinoma is based on the degree of involvement of the main and secondary hepatic confluences (1) ( Table 1). Representative cholangiograms are shown in Figure 2.
Because of its remarkable efficacy and low morbidity and mortality, endoscopic stenting has become the modality of choice for relieving obstructive jaundice in cases of inoperable hilar strictures. Furthermore, this technique can be used as a preoperative measure for potentially resectable lesions and for patients who might be candidates for adjuvant treatment, including photodynamic and radiation therapy (2)(3)(4)(5).

TECHNIQUE OF BILIARY STENTING
Complete cholangiography, which determines the precise location and extent of the hilar stricture, is the first step in planning effective biliary drainage. In previous years it was obtained by endoscopic retrograde cholangiopancreatography (ERCP). Nowadays, magnetic resonance cholangiopancreatography (MRCP) can provide detailed imaging with precise delineation of the hilar stricture without the need for an invasive procedure such as ERCP or percutaneous transhepatic cholangiography. There is reasonable agreement between the two techniques in determining the stage of cholangiocarcinoma, according to Bismuth-Corlette stage (Table 2) (6).
Biliary sphincterotomy is often undertaken, especially if it is anticipated that more than one stent will be inserted, but is not mandatory if the papillary orifice is sufficiently patent or has been mechanically dilated already. Sphincterotomy may also be precluded by the presence of a significant bleeding disorder.
A guidewire is then advanced through the stricture into the obstructed bile ducts. Hydrophilic guidewires are preferred because they are easier to pass through the stricture than Teflon-coated guidewires. When indicated, tissue sampling (either brush cytology or endoluminal forceps biopsy) should be carried out before stenting. The specificity of such techniques is impressive (approaching 100%) but their sensitivity ranges from only 15% to 70% (7). Tight strictures may be dilated before stenting, using bougies, coaxial dilating catheters or balloons (pneumatic dilation).
A plastic guiding catheter is then advanced over the guidewire and a plastic stent is inserted over the guidewirecatheter complex using another coaxial pushing catheter of the same size. Large bore (10 Fr) polyethylene or Teflon stents are generally used, because they have been shown to remain patent for significantly longer periods of time than do narrower devices (8). However, there is no significant advantage for 11.5 Fr over 10 Fr stents (9). Multiple plastic stents are required in cases of complex hilar strictures. In such cases, it is preferable to place the first stent into the left hepatic duct due to its acute angle resulting in more difficult stent placement. When inserting more than one stent, it is important first to insert all the necessary guidewires in the ducts chosen for drainage, to maintain access to each blocked duct after the first stent has been placed. Self-expandable metal stents (SEMS), which usually achieve a maximum diameter of 30 Fr, can be placed over the guidewire with or without preliminary dilation of the stricture. Most authors recommend the use of uncovered SEMS for hilar tumours, to avoid complete blockage of the other biliary branches.
High-volume centres enjoy a technical success rate for endoscopic stent placement of nearly 90% (10). Clinical success (ie, palliation of jaundice and pruritus) can be expected in more than 80% of patients in whom incomplete drainage of the bile ducts has been achieved (11,12).

HOW MUCH DRAINAGE IS ENOUGH?
The goal of palliation in cases of obstructive jaundice is to relieve cholestasis, which requires complete drainage of the intrahepatic ductal system. In practice, stent placement that drains 30% of the liver may alleviate jaundice but not cholestasis. The remaining obstructed ducts provide a continued risk of cholangitis and hepatic abscess (13).
The number of stents needed to drain a malignant hilar stricture should be equal to the number of obstructed ducts at the primary and secondary (right or left) hepatic confluence. For example, three stents would be required for a Bismuth-Corlette type III stricture. Moreover, biliary anatomy may complicate efforts to obtain endoscopic drainage: anatomical     variants involving the right hepatic duct system (eg, right sectorial duct originating from the left hepatic duct) occur in at least 40% of the general population. Some authorities recommend that only a single hepatic lobe be drained, claiming no differences in outcome (11,12). A recent randomized trial undertaken by De Palma et al (14) addressed this issue. They assigned 157 consecutive patients with malignant hilar obstruction to either unilateral (group A) or bilateral (group B) hepatic duct drainage. The two groups were comparable in terms of tumour origin and Bismuth-Corlette stricture type. Results are summarized in Table 3. In the intention-to-treat analysis, patients in group A had significantly higher rates of successful stent insertion and successful drainage and a lower incidence of early complications, mainly because of fewer cases of cholangitis. The rates of late complications (in patients who had survived at least 30 days) and median survival were comparable in the two groups. In the per-protocol analysis (which evaluated only patients in whom stent placement had been successful), there were no significant differences in rates of effective drainage, complications or survival. The results of this study suggested that there was no advantage to attempting bilateral drainage.
The De Palma study (14), however, is far from being free of bias. For example, approximately 30% of the patients in both groups had a Bismuth-Corlette type I stricture, which clearly does not require bilateral drainage. Detailed subgroup analysis of these patients was not undertaken, so it is not possible to determine how their inclusion may have affected the results. Furthermore, the authors did not compare complete with incomplete drainage, but only one versus two stents, even in patients with Bismuth-Corlette type III strictures (which require three stents to achieve complete drainage). The high rate of early and late cholangitis in patients with complex strictures could be attributed to the presence of injected and undrained bile ducts.
Chang et al (10) found that bilateral drainage is associated with lower sepsis-related morbidity and mortality and longer mean survival, at least when there has been contrast injection into both lobes. A previous study also demonstrated longer survival in patients who have undergone complete drainage of the biliary tree (13).
Placement of one or more metallic stents is an alternative option for patients with complex hilar obstruction. A recent study (15) showed that the insertion of unilateral (in 28 patients) or bilateral (in nine patients) uncovered metallic stents alone provided definitive palliation in 69% of patients with hilar cholangiocarcinoma, with an overall median survival of 149 days. Even better results have been reported in a recent series of 61 patients who underwent unilateral placement of SEMS: successful drainage was achieved in 96.7%, late stent occlusion occurred in 22.9% and median survival was 140 days (16). It is still unknown, however, if complete drainage with metallic stents provides a superior clinical outcome.

FUTURE DIRECTIONS AND UNRESOLVED ISSUES
Hintze and colleagues (17) (iridium-192, 30 Gy to 50 Gy) to 12 patients. Eight of the patients experienced mild to moderately severe gastrointestinal toxicity, and two patients who received brachytherapy developed duodenal ulceration. Of the 12 patients who received brachytherapy, four achieved a clinical response (two complete and two partial). The median survival was 21.2 months, and two of the patients with unresectable disease survived longer than five years.

CONCLUSIONS
More investigation is required to refine endoscopic and other techniques, and adapt them to the specific characteristics of the underlying disease. We still need to determine whether complete drainage of the obstructed biliary tree, using multiple stents, is necessary to achieve adequate palliation, because published studies have yielded conflicting results. Preliminary evidence suggests that incomplete drainage may reduce morbidity from cholangitis. The role of MRCP should be defined. The relative merits of uncovered metallic versus plastic stents must be explored. Finally, more work needs to be done with adjunctive techniques, including intraluminal brachytherapy, photodynamic therapy and high-frequency ultrasonography.
Can J Gastroenterol Vol 18 No 7 July 2004 453