Hellenic Journal οf Radiology

1. Introduction

The tumours that may lead to malignant biliary ob­structions (MBOs) are mainly adenocarcinoma of the pancreatic head and extrahepatic cholangiocarcinoma [1]. Other tumours that may lead to malignant jaundice are ampulary, gallbladder, and hepatocellular carcino­ma and lymph node strictures from gastric, pulmonary, breast and oesophageal cancer. When malignant jaun­dice occurs, the normal bile flow towards the duodenum is blocked. Although the complete cholestasis mecha­nism is not fully interpreted, the lack of bile in the duo­denum results in a systemic inflammatory reaction due to release of cytokines that targets specific organs hence leading to multi-organ failure. The deranged liver func­tion leads to Kupffer cells dysfunction and the lack of bile in bowel increases the membrane permeability and re­duces the bowel bacterial barrier, so promoting bacteri­al migration, initially to the portal and then to the sys­temic circulation [2-4]. If this situation is not corrected, uncontrollable sepsis occurs very quickly in the major­ity of cases.

2. Management of patients with MBOs

Patients with a MBO would either be considered opera­ble or inoperable according to the stage of disease at di­agnosis. Endoscopic drainage is usually the first approach in most of the centres. However this is not always techni­cally feasible, particularly for lesions located proximal­ly to the liver hilum [5]. Furthermore, when preoperative drainage is required there is an increased risk of infectious complications with the endoscopic approach [6]. The main contraindications for endoscopic retrograde cholangio­pancreatography (ERCP) are the presence of an oesopha­geal stricture, gastric volvulus, bowel perforation, unsta­ble cardiopulmonary status that would prohibit patient to be positioned supine and history of previous bowel sur­gery with change of the anatomy (Billroth II or Roux-en-Y loop). Therefore percutaneous transhepatic drainage ac­cess should be considered as the most appropriate treat­ment option [7]. Once the biliary tree is decompressed and biopsy of the underlying lesion is obtained, a multidisci­plinary team decides on whether to proceed with surgi­cal resection or palliative treatment. In the case of palli­ative approach, internalisation of the drains with the use of stents is required in order to reduce infection risk dur­ing chemotherapy. Stents ideally need to be patent for the whole patient’s life span, in order to avoid treatment in­terruption or re-intervention for cholangitis. The man­agement algorithm of a MBO patient is shown inTable 1.

3. Diagnostic Imaging

Initial imaging approach of MBOs is performed with trans abdominal ultrasound (US) that is expected to de­tect the dilated bile ducts and the presence of possible intrahepatic deposits. US is quick, accessible and of low cost. However some pitfalls may occur in the case of obese patients, when there is bowel interposition or in the case of paralysis of the right hemidiaphragm.

Computed tomography (CT) is usually performed not only to delineate the stricture but also to assess the pres­ence of intrahepatic and distal metastatic deposits. A triple phase CT scan is usually recommended for char­acterisation of the malignant stricture, even though ma­lignant biliary tumours are not expected to be enhancing avidly in the arterial phase. Adding a late contrast phase (6-15 min post injection) appears to increase the sensi­tivity of detection of adenocarcinomas [8].

Magnetic resonance imaging (MRI) plays also an im­portant role in the diagnosis of malignant biliary stric­tures, particularly if combined with magnetic reso­nance cholangiopancreatography (MRCP); the latter is a T2-weighted fast-spin echo sequence applied with pro­longed effective echo time (>240 milliseconds). The sig­nal from static fluid (biliary tree) is enhanced and the signal from the surroundings structures is suppressed (Fig. 1a). MRCP is able to locate the level of obstruction in 85-100% of the cases. In addition, with the use of con­ventional MRI sequences the sensitivity and the specific­ity for type of tumour diagnosis reach 86% and 98% re­spectively [9].

Malignant strictures involving the hilum are classified using the Bismuth-Corlette classification system based on the extension of the stricture into the intrahepatic ducts [10]. Bismuth type I strictures involve the proxi­mal common hepatic duct and spare the confluence be­tween the left and right ductal systems; type II strictures involve the confluence and spare the segmental hepatic ducts; types IIIa and IIIb involve either the right or left segmental hepatic ducts, respectively; and type IV stric­tures involve the confluence and both the right and left segmental hepatic ducts.

4. Percutaneous approach and drainage

The main indication for percutaneous transhepatic bil­iary drainage in a patient with a malignant biliary stric­ture is the elevation of bilirubin from a mechanical cause. The procedure may be performed on an emergency ba­sis if jaundice is combined with cholangitis and sepsis. Contraindications are mainly technical and are usual­ly relative like i.e. the coagulation status of the patient or the presence or not of ascites, with the exemption of complete infiltration of the liver parenchyma from wide­spread metastatic disease (Table 2).

Assessment of the coagulation status of the patient is of paramount importance prior to the procedure. In case of deranged clotting, blood products (vitamin k, fresh fro­zen plasma or platelet transfusion) may be administered. In case of presence of ascites, percutaneous drainage may be performed prior to accessing of the biliary tree. The procedure is usually performed under local anaesthesia (lidocaine 2%) and conscious sedation using 1-8 mg of mi­dazolam and 50-200μg of fentanyl. Antibiotic prophylaxis (i.e. with cefuroxime 750 mg) may be administered before the procedure and continued for up to 5 days after the pro­cedure according to the operator’s preference.

Percutaneous approach is performed by a transhe­patic puncture with a rather thin (21-22 G) Chiba nee­dle under US or fluoroscopic guidance. When access to a bile duct is obtained a minor amount of diluted con­trast is usually injected to confirm correct position. Oku­da et al. first described the fluoroscopically guided tech­nique, in 1974 [11]. When the thin needle was positioned in the biliary system a cholangiogram was performed and a second puncture followed to a duct that was con­sidered adequate in terms of angulation and size. The second puncture was performed with a 5 Fr needle cath­eter and a 0.035” inch wire was advanced in the biliary tree. With the use of US guidance, once the thin needle is in the biliary system a 0.021” inch wire is advanced in the biliary tree. The system is upsized to 6 Fr with­out the need of a second puncture in this case. When the 6 Fr catheter is advanced a cholangiography is per­formed with diluted iodinated contrast (Fig. 2a) [4]. For common bile duct (CBD) lesions, right-sided approach is preferred, except in cases of ascites or colon interposi­tion (Fig. 1b). Right side route provides a more straight way for wire and catheter manipulation and keeps op­erator’s hands away from x-ray beam (Fig. 2b). Left sideapproach on the other hand appears less painful for the patient as there is no route through intercostal spaces. For hilar lesions the puncture site should be decided af­ter studying tumour location, extent and liver lobe infil­tration, potential lobe atrophy and biliary tree anatom­ical variability [12].

In case of opacification of multiple obstructed bile ducts, the operator should try to drain as many of the opacified ducts as possible in order to avoid bacteri­al contamination and post procedural contrast related cholangitis [13].

In the case of MBOs, percutaneous transhepatic biliary drainage (PTBD) follows in order to decompress the ob­structed bile duct system [14]. The drainage of the bile ducts is usually performed with a small 8 Fr plastic mul­ti-hole pigtail catheter. In the cases where the lesion has not been biopsied or in the case where the biliary tree is infected, external drainage catheter placement is sug­gested. The catheter is secured to the skin with sutures. Self-locking catheters are preferred in order to minimise the dislocation risk. In cases of complex hilar strictures, placement of multiple external biliary catheters may be necessary to achieve complete drainage. The types of available drains are described inTable 3a.

Biopsy of the lesion may be either obtained with endo­scopic ultrasound - guided fine-needle aspiration (EUS-FNA) or with the use of endobiliary forceps. EUS-FNA is effective in obtaining cytological samples for Bismuth I-II lesions, however for hilar lesions this access is less ef­fective [15, 16]. In addition, the cytologic sample may not always be diagnostic and core biopsy of the lesion may be required. Endobiliary biopsy may be obtained with theuse of biopsy forceps overcoming the problem of inade­quate sampling that is encountered with FNA [17]. The forceps may be advanced either endoscopically or per­cutaneously. In the case of percutaneous insertion a 7Fr sheath is used as access and a security wire is used and placed across the stricture. The forceps are placed on the side of the wire and usually 3-4 samples are obtained and placed in a formalin suspension.

If there is already confirmation of malignancy and there is intention to proceed with a stent the obstruc­tion can be crossed with a hydrophilic wire and the per­cutaneous biliary catheter can be pushed through the stenosed /obstructed duct, so that bile is draining both internally towards the duodenum and externally (inter­nal- external drainage) (Fig. 3a) [4, 5]. By crossing the stricture a “rendezvous” procedure with endoscopic ap­proach may also be performed particularly for Bismuth III and IV lesions.

5. Use of biliary stents

In case of inoperable tumours, internalisation of the drain is required in order to be able to either administer palliative chemotherapy or to offer a satisfactory quality of life [18-20]. This is achieved with the use of stents that may be plastic, bare metallic or covered metallic (Table 3b). Prognostic indicators for stent placement are thegeneral status of the patient, the performance score of oncologic measures (i.e. the Karnofsky performance sta­tus scale) and the status of the organ as indicated by the liver biochemistry.

Metallic biliary stents have been proved as the best palliative treatment of non-resectable malignant ob­structive jaundice, allowing longer patency rates than plastic endoprostheses [21]. The technique is safe, with a high technical success rate of over 97% and low com­plication rate [22, 23] and procedure related mortality is reported between 0.8 and 3.4%. Early complications rate within the first 30 days is about 2% while the late com­plications rate can reach 16% [23]. Mean overall prima­ry stent patency is reported 120 days, but mean overall secondary stent patency is 242.2 days [22].

An integration process accompanies the deployment of metal endoprostheses within the biliary tree into the surrounding malignancy, which after a period leads to the formation of tissue through the stent’s struts that fi­nally blocks the stent. This is the so-called “tumour in­growth”. Ingrowth is the result of tumour growth but also of the accumulation of biliary sludge and bacterial biofilm. The motility of bare stents plays a crucial role on the latter aspect of ingrowth as “fractures” of the stents occur and lead to dysfunction. This effect occurs more frequently with laser cut metallic stents therefore the use of woven stents is suggested.

The malignant proliferation can sometimes occur in the proximal end of the stent and not through the mesh. This is called “tumour overgrowth” and also gradually blocks the stent lumen [24]. In such cases, patient requires new percutaneous intervention that leads to placement of a second stent through the occluded stent. In some cases stent occlusion is due to bile sludge and needs “cleaning” with a simple angioplasty balloon (Fig. 3b-c).

Hausegger et al. [25] in the early days of use of metal­lic stents in the biliary system analysed the histological changes after the deployment of stainless steel endopros­theses in the biliary tree for the treatment of malignant biliary disease. In fourteen cases histological examina­tion was performed after autopsy and in one case a sur­gical specimen was analysed after tumour resection. In the analysed specimens, the stent was inserted in a peri­od ranging from 5 days to 21 months. For the initial peri­od of the first 15 days, histology revealed that the cuboid epithelium of the biliary tree is completely destroyed in the areas that were in contact with the stent. There were moderate inflammatory changes in the sub muco­sa with minor lymphatic infiltration and oedema. The internal layer of the stent was covered by non-specificgranulomatous tissue. The fibrous tissue and the tumour were displaced. No tumour cells were recognised within the biliary tree lumen and no signs of acute inflamma­tion were noticed. In the period between the 2nd and the 12th month the endoprosthesis was gradually integrat­ed to the surrounding tissue by a layer of granulomatous tissue and tumour ingrowth. In a similar study Boguth et al. [26] describe similar histological changes that oc­cur in the first 3 to 6 months and lead to the occlusion of the bare metallic endoprosthesis. Ingrowth through the mesh of the bare stent occurs in all patients that survive more than 6 months and re-intervention is usually re­quired. The patients may present with cholangitis and a new procedure and a new stent placement is usually nec­essary to resolve the situation.

There is usually a difference in timing of clinical ex­pression of symptoms that is related to the location of the tumour. Intrahepatic lesions tend to give symptoms later than extrahepatic ones. This is due to the fact that the only symptoms expected are those related to the ma­lignant obstruction of the biliary tree and in case of in­trahepatic tumour, there are several collateral drains that may be used until complete occlusion occurs [27].

6. Covered stents

In order to reduce stent’s dysfunction from tumour in­growth, covered metallic stents were developed in the last decade. Various authors tested several coverage ma­terials with a different range of results [27]. The initial­ly used covered stents were “home made” by applying a coverage membrane on the available bare stents. Sai­to et al. in 1994 used biliary Gianturco-Roesch Z-stents covered with a Gore-Tex membrane [28] and reported satisfactory medium- to long-term results in a study of six patients. Thurnher at al. reported in 1996 their ex­perience with the first type of covered Wallstents [29]. The coverage was a 0.015 mm thick polyurethane mem­brane that was also used from Rossi et al. in 1997 [30] and Hausseger et al. in 1998 [31]. Both investigator groups re­ported that the 0.015 mm polyurethane membrane was eroded from tumour and gastric juice. Similar results were also presented from Kanasaki et al. in 2000, where nitinol Strecker stents were used with the same coverage [32]. A 0.035 mm polyurethane membrane was used in homemade covered Gianturco-Roesch Z-stents and spi­ral Z-stents from Miyayama et al. in 1997, with better re­sults [33]. Han et al. reported a 71% patency at 20 weeks using a 0.030 mm thick polyurethane membrane in cov­ered Niti-S stents [34]. Isayama et al., using 0.040-0.050 mm thick polyurethane-covered Wallstents, did not re­port tumor ingrowth [35] and presented even improved results with a 0.050-0.060 mm polyurethane membrane using covered Diamond stents [36].

During the last ten years, covered stents with a cover­age membrane from expanded poly-tetra-fluoro-ethyl­ene/ fluorinated-ethylene-propylene (ePTFE/FEP) were developed and are available in the market (Viabil©, W. L. Gore & Associates, Flagstaff, AZ, USA) [24]. Randomised controlled trials were performed with the use of such covered devices that have shown the superiority of these covered stents in comparison to uncovered stents in spe­cific patient population(Fig. 2a-b) [37, 38].

Covered stents are not suitable for all patients with ma­lignant jaundice. They have to be reserved for patients with a reasonable expected survival and not used in the case of advanced disease [39]. Furthermore, anatomical features have to be taken into consideration including the stricture site, location and patency of the intrahe­patic, cystic and pancreatic ducts. Usually, only Bismuth type I strictures are suitable for covered stent placement, whereas specific covered stents may be placed in some cases of type II. The covered portion should not be ad­vanced in the intrahepatic ducts in order to avoid chol­angitis. For this purpose, covered stents with side holes have been developed. The “holed” region does not pre­vent tumour ingrowth and it is also not extending prox­imally enough to prevent tumour overgrowth as a bare stent would. Nevertheless, side holes permit placement in anatomically complicated cases avoiding cholangitis or cholecystitis. The same principal should be followed for the cystic duct but less for the pancreatic, since pan­creatitis may less frequently occur and the location of pancreatic duct is not a true limit in stent placement.

7. Stent placement techniques

Stent placement can be performed as one- (so called “pri­mary stenting”) or two- (or more) step procedure. The factors that would influence the type approach are mul­tiple but mainly consist of the presence or not of diag­nosis of malignancy, the presence or not of biliary sepsis and technical issues such as intra-procedural bleeding or bile leak.

If a more than one step approach is decided then a bil­iary drainage catheter may be left in situ for 1-2 weeksbefore a stent is inserted. Adam et al. [40] introduced the concept of “primary stenting” in 1991 (Fig. 1b-c). They suggested that self-expandable metallic stents could be placed in patients with a better prognosis as a one-step procedure, aiming to reduce total hospitalisation time. However, at the time operators were inclined to check with a cholangiogram 1-3 days later. Therefore they suggested leaving in situ a 5 Fr angiographic cath­eter as access for the cholangiographic control (Fig. 1c). The catheter was retrieved, after stent’s expansion was confirmed and complications were excluded (Figure 4c-d). This is a practice that many operators still follow nowadays, even though true “primary stenting” con­sists of obtaining access, crossing the lesion, deploying a stent and plugging the tract in one session (Fig. 1d). To achieve that, diagnosis needs to be obtained prior to the biliary procedure i.e. previous EUS-FNA and the whole biliary procedure needs to be uneventful with­out bleeding or biliary leak. Intra-procedural remodel­ling of the stent is required to ensure that expansion is satisfactory. If so the tract can be plugged with gel­foam pellets and the patient can be discharged the next day. The whole procedure may last 1-2 hours, depend­ing on the degree of anatomical complexity. The proce­dure may be performed under deep sedation and anal­gesia or general anaesthesia [41]. Hospitalisation time depends on the complexity of the case and the manoeu­vres performed. In general, patient is kept in hospital until he/she fully recovers. Immediate post-procedure monitoring consists of haemodynamic assessment and puncture site review.

In the past years, there was discussion about the use­fulness of bilateral versus unilateral lobe drainage and stenting. In case of bilateral drainage, stents may ei­ther be positioned by puncturing separately the right and left side ducts and catheterising separately the CBD (Y configuration), or may be placed from a single side puncture after catheterising the other side ducts and the CBD form the same side (T configuration). There is still a degree of controversy as to whether partial or complete biliary drainage should be done. Inal et al. [42] studied 138 patients with hilar malignant strictures that received unilateral or bilateral stenting. Only patients with type IV lesions appeared to benefit from bilater­al stenting, whereas for those with Bismuth type II and III there was no benefit in terms of patency. Although the cumulative stent patency seemed to be better af­ter bilateral than unilateral drainage approach, there is, based on the available literature, not enough data to support bilateral drainage for malignant hilar obstruc­tion [43]. Older and recent studies show that partial-liv­er drainage achieves results as good as those after com­plete liver drainage with significant improvements in quality of life and reduction of the bilirubin level [22, 44, 45]. Therefore the insertion of more than one stent would not appear justified as a routine procedure in pa­tients with biliary bifurcation tumours.

Another frequently asked question was if we should stent the sphincter of Oddi in every case, even if the tumour is ending higher than the level of the papilla. A study performed in 2001 showed that in patients with extrahepatic lesions lying higher than 2 cm from the papilla and with a relative poor prognosis (<3 months), due to more advanced disease or to a worse general condition, the sphincter of Oddi should be also stent­ed in order to reduce the post-procedural morbidity (Fig. 1b-c) [20].

8. Plastic vs. metallic stents

Bare self-expandable metallic stents have a mean pa­tency of approximately 6-8 months and this is superior to what the plastic endoprostheses offer (Fig. 1c) [21, 46]. According to Lammer et al. [21], the use of self-ex­panding metal stents appears to show substantial bene­fits for patients with MBOs and also appears to be more cost-effective than the use of plastic ones, even though the individual plastic stent is significantly cheaper than the metallic one. The cost-effectiveness lays on the fact that the patency is higher and the number of repeated interventions significantly lower [46]. A recently pub­lished meta-analysis includes results from 785 patients in 10 randomised trials half of which received a metallic stent and half a plastic one [47]. The results showed that metal stents were associated with a significantly longer patency, smaller number of re-interventions and longer survival of the patient, indicating once again that plas­tic stents do not offer any significant benefit for MBOs.

9. Bare vs. Covered stents

What is of interest is the comparison between uncov­ered and covered metallic stents. ePTFE/FEP covered stents were used in two randomised trials published in the literature. In the first trial [38], covered stents were directly compared with uncovered in patients with Bismuth type I cholangiocarcinoma. Sixty patients (36 men and 24 women, with age range 46-78 years) were randomised with the use of a sealed envelope for the placement of a covered or a bare stent. In 21 cases the tumour also infiltrated the cystic duct. Patients were fol­lowed-up with telephone interviews and on an outpa­tient basis. Technical success was 100% for both groups. Minor early complications were noticed in 13.3% of the bare stent group and 10% of the patients of the covered stent group. The mean follow-up period was 212 days (45-675 days) and all patients had passed away at the end of the study. Thirty-day mortality was zero for both groups. Median survival time was 180.5 days for the bare stents and 243.5 days for the covered stents, withp<0.05. Stent’s mean patency rate was 166 days for the mesh stent and 227.3 days for the covered stent, withp<0.05. Stent dysfunction occurred in 9 patients with bare stent after a mean period of 133.1 days and forceps biopsy re­vealed ingrowth in 88.8%. Dysfunction occurred also in 4 cases of the covered stent group after a mean period of 179.5 days and it was due to tumour overgrowth in 2 and due to sludge incrustation in another 2. Tumour in­growth occurred exclusively in the mesh stent group. There was also no difference in the overall cost of the two groups after a cost analysis.

The second prospective randomised trial that com­pared ePTFE/FEP covered stents with bare stents was performed in patients with pancreatic adenocarcino­ma and revealed similar results [37]. The study was per­formed in 80 patients that were also randomised with the use of a sealed envelope into a bare stent and a cov­ered stent group. The patients were 53 men and 27 wom­en with an age range from 41 to 79 years (mean 62.7 years). Technical success was 100% in both groups. Ear­ly complications were observed in 10% of the bare and 12.5% of the covered stent group. Median follow-up time was 192 days (range of 104-603 days); all patients passed away by the end of the study. The 30-day mortality was zero for both groups. Median survival time was 203.2 days for the bare stent group and 247 days for the cov­ered stent group, and this difference was not statistical­ly significant. Mean primary patency was 166 days for the uncovered and 234 days for the covered stents, with p< 0.05. Dysfunction occurred in 12 bare stents after a mean period of 82.9 days and it was due to tumour ingrowth in 91.6% of the cases. Dysfunction occurred in 4 covered stents after a mean period of 126.5 days and it was due to tumour overgrowth in 2 and due to sludge in 2. Cost anal­ysis revealed that there was no difference in the overall cost of the two groups. The two randomised trials showed that the ePTFE/FEP covered metal stents appear to reduce significantly the rate of stent’s dysfunction. The micro porous membrane appears to limit completely the risk of tumour ingrowth, which is the main issue of the use of metallic endopros­thesis in malignant biliary obstruction. However, in or­der to benefit from the results of covered stents patient’s survival needs to be long enough for ingrowth to occur. If survival of more than six months can be predicted -by the lack of metastatic disease and the performance sta­tus of the patients- then use of an ePTFE/FEP-covered stent is completely justified. The mentioned device ap­pears to limit also another of the major problems of the covered stents which is stents’ migration, by having the lateral barbs (anchoring fins).

There are still some investigations going on about prevention of “overgrowth”. An improvement might be achieved if covered stent placement could be com­bined with bare stent extension proximal and distal to the coverage. In this manner, the papilla could be stent­ed with bare metal, so that potential episode of acute pancreatitis is avoided and also tumour overgrowth can be delayed due to the bare stent above the tumour area (Fig. 4a-d).

10. Percutaneous vs. Endoscopic drainage

The decision on whether to obtain an endoscopic or a percutaneous access to a blocked biliary system has been based mainly on local expertise and availability. There is very limited comparison of the two approach­es in the literature. The area where the two methods have been more extensively compared is the preoper­ative biliary drainage where the endoscopic approach is considered to jeopardise the aseptic biliary environ­ment and lead to infections of patients that will be op­erated. The existing studies have been analysed in a very recently published meta- analysis with a focus on patients with resectable perihilar cholangiocarcinoma [48]. The authors included four retrospective studies on 433 patients; 275 (63.5%) underwent an endoscop­ic drainage and 158 (36.5%) a percutaneous. The per­cutaneous approach has demonstrated to offer superi­or results in terms of procedure-related morbidity and rate of conversion from one procedure to the other and cholangitis rate whereas pancreatitis occurred exclu­sively in the endoscopic group.

In a very recent publication the results of EUS are com­pared with the use of percutaneous transhepatic cholan­giographic endobiliary forceps biopsy (PTC-EFB) in termsof diagnostic samples performance in 137 patients in a retrospective cohort study showing similar sensitivity, negative predictive value and accuracy [49]. The authors concluded that endobiliary biopsy should be the treat­ment of choice in case biliary drainage is also required.

11. Complications and how to deal with them

The most common minor complications are pain, stent migration, stent insufficient expansion and fever. Pain is treated or prevented by IV or IM administration of an­algesics and/or sedatives [41]. Stent insufficient expan­sion can be corrected by post-stenting balloon dilata­tion. Stent misplacement or migration can be corrected by placement of a second stent.

Most feared major complications are sepsis, bleeding and bile leakage. As mentioned above, complication rate within the first 30 days is about 2%, while late complica­tions rate can reach 16% [23]. Comparing uncovered to covered stents, minor early complications were noticed in 10-13.3% of the bare stent group and 10-12.5% of the patients of the covered stent group, with a 30-day mor­tality of 0% for both groups [37, 38].

In order to prevent such serious complications, biliary interventions should be performed under IV antibiotic coverage [50]. Any biliary intervention is considered at the minimum a clean-contaminated procedure and there­fore the recommendation is that all patients scheduled for biliary drainage receive prophylactic antibiotics pri­or to the procedure [50-52]. Transient bacteraemia occurs in approximately 2% of patients after biliary intervention [51]. If a patient develops fever and/or chills following biliary intervention, antibiotics may be continued, fluid resuscitation should be initiated and the need for blood cultures considered. In some cases, infection does not re­spond to these measures and additional drainage may be required to address incompletely drained or isolated bile segments. In patients with sub-segmental isolation, mul­tiple drains could potentially be required and long-term antibiotic suppression may be favoured [52].

Arterial bleeding is a relatively rare complication of PTBD, appearing in 0.6-2.3% and when it does not resolve spontaneously, it should be treated by selective arteri­al embolisation [53-55]. Arterial complications might be prevented by obtaining access from second or third or­der ducts, located in the periphery of the liver and not near the hilum. Central punctures might be complicated with portal vein and/or arterial injury that will be man­ifested with haemobilia and/or pseudoaneurysm forma­tion. It appears that there is a higher incidence of haemo­bilia associated with left lobe puncture, but did not reach the threshold of statistical significance (p=0.077) in pre­vious studies [56, 57]. An emergency angiography should be considered in all patients in whom a pseudoaneurysm is suspected following hepatobiliary interventions. Tran­scatheter arterial coil embolisation is a safe and effec­tive treatment for pseudoaneurysm with a technical suc­cess rate of 95.8% [57, 58]. Minor complications can be observed after embolisation in 80.6% patients, 76.4% of whom may have hepatic ischaemia and 4.2% focal he­patic infarction [57]. Surgical intervention should be re­served for patients for whom embolisation is not possi­ble or fails [59].

Bile leakage can occur through the parenchymal tract after catheter removal(Fig. 2b-c). In order to prevent this, several embolisation techniques have been proposed [60-62]. One is by placing small pieces of expandable gel­atin foam pellet (Hunter Biopsy Sealing Device, Vascu­lar Solutions, MO, USA) [59]. The authors report that uti­lisation of this method has the potential to reduce the morbidity and mortality burden associated with post-PTC haemorrhage by preventing bleeding from the liver ac­cess tract. The same kind of embolisation can be achieved by injection of n-butyl cyanoacrylate mixed with Lipiodol (Fig. 1c-d) [62]. A relatively newer device is Hep-Plug sealing (IFU, Vascular Solutions, MO, USA), which works in a similar way as the Hunter Biopsy Sealing Device.

12. Conclusions and future perspectives

Percutaneous transhepatic biliary procedures are inte­grated in the management of patients with MBOs. In case of operable disease preoperative biliary drainage may be performed - with better results than the endoscopic ap­proach- offering decompression of the biliary tree and ac­cess for endobiliary biopsy. In case of palliative approach either woven bare stents or covered stents may be used to alleviate jaundice for the patient’s life span. Such proce­dures have to be part of the everyday armamentarium of interventional radiology centres. Future perspectives will probably be in the direction of smaller profile and func­tional (or “drug eluting”) stents and endobiliary ablation treatment that are still in a very ear

Figure Figure Figure Figure Figure Figure