Efficacy and safety of endoscopic ultrasound-guided hepaticogastrostomy for biliary drainage in hypervascular hepatocellular carcinoma: a retrospective study from Japan

Article information

Clin Endosc. 2024;.ce.2024.079

Publication date (electronic) : 2024 November 11

doi : https://doi.org/10.5946/ce.2024.079

Kenneth Tachi, Kazuo Hara, Nozomi Okuno, Shin Haba, Takamichi Kuwahara, Toshitaka Fukui, Ahmed Mohammed Sadek, Hossam El-Din Shaaban Mahmoud Ibrahim, Minako Urata, Takashi Kondo, Yoshitaro Yamamoto

Department of Gastroenterology, Aichi Cancer Center Hospital, Nagoya, Japan

Correspondence: Kazuo Hara Department of Gastroenterology, Aichi Cancer Center Hospital, 1-1 Kanokoden, Chikusa-ku, Nagoya 464-8681, Japan Tel: +81-52-762-6111 E-mail: khara@aichi-cc.jp

Received 2024 April 2; Revised 2024 July 28; Accepted 2024 July 29.

Abstract

Graphical abstract

INTRODUCTION

Hepatocellular carcinoma (HCC) is the seventh most commonly diagnosed cancer and the third leading cause of cancer-related deaths worldwide.1 Jaundice is common and reported in 5% to 44% of patients at the time of diagnosis. It often occurs due to underlying liver cirrhosis and/or extensive hepatic parenchymal destruction by the tumor. Biliary obstruction as a cause of jaundice is uncommon, but does occur in the late stages of the disease in approximately 0.5% to 13% of cases.2,3 Biliary obstruction may result from tumor thrombi, hemobilia, direct tumor infiltration, extrinsic compression by tumor encasement, or metastatic lymph nodes of the major bile ducts.4 The presence of obstructive jaundice is associated with debilitating symptoms, such as pruritus, denial of access to chemotherapy, and poor prognosis.3 Therefore, biliary drainage offers palliation of jaundice and significantly improves patient survival.2,5-8

Endoscopic retrograde cholangiopancreatography (ERCP) biliary drainage (EBD) via transpapillary biliary stent placement has traditionally been the standard of care for unresectable malignant biliary obstructions.9,10 Percutaneous transhepatic biliary drainage (PTBD) or surgery are alternatives when EBD fails. However, with the advent of endoscopic ultrasound (EUS)-based therapies, EUS-guided biliary drainage procedures, including transmural drainage, such as EUS-guided hepaticogastrostomy (EUS-HGS) and EUS-guided choledochoduodenostomy (EUS-CDS), have evolved into salvage procedures when ERCP fails or is not possible.11 They are even used as first-line options for biliary drainage in patients with unresectable malignant biliary obstructions at some centers.12 These have been shown to have equivalent or superior technical and clinical success rates to PTBD and surgery, but are less invasive, have fewer side effects, and offer a better quality of life.13,14

EBD for obstructive jaundice in patients with HCC poses a unique challenge because of the highly vascularized nature of the tumor and the associated increased risk of bleeding from traversing the tumor. Subsequent stent exchange is associated with a high risk of bleeding. However, EUS-HGS is performed via a route that does not traverse the tumor (Fig. 1). Although we acknowledge that bleeding is the most common complication of EUS-HGS and that cirrhosis, if present, can pose an additional bleeding risk, we hypothesized that EUS-HGS could be a safe and effective biliary drainage option for HCC patients with obstructive jaundice and retrospectively investigated the use of EUS-HGS in these patients.

Fig. 1.

Schematic representation showing the position of a stent placed via (A) endoscopic retrograde cholangiopancreatography biliary drainage and (B) endoscopic ultrasound-guided hepaticogastrostomy for drainage of malignant biliary obstruction.

METHODS

The medical records of all patients with biliary obstruction due to unresectable hypervascular HCC who underwent EUS-HSG drainage at the Aichi Cancer Center, Japan, between February 2017 and August 2023, were retrospectively reviewed (Fig. 2A). At Aichi Cancer Center, EUS-HGS is the primary drainage procedure for malignant biliary obstructions. This procedure was performed to avoid ERCP-associated pancreatitis.

Fig. 2.

(A) Computed tomography scan image of a hypervascular hepatocellular carcinoma causing biliary obstruction (arrowhead). (B) Endoscopic ultrasound image of dilated intrahepatic bile duct, B3. (C) Fluoroscopy image highlighting the dilated bile duct and appropriately placed guidewire after contrast injection. (D) Endoscopic image of the implantation site and intragastric portion of a successfully deployed fully covered self-expandable metal stent.

The last recorded clinical and laboratory data before the EUS-HGS procedure were used as baseline characteristics. Baseline ascites, if present, was graded as mild when detectable only by ultrasound examination, moderate when it caused moderately symmetrical distension of the abdomen, and severe or large when it caused marked abdominal distension.15

All EUS-HGS procedures were performed under midazolam sedation with the patients lying prone. A gastroscope was initially used to identify the gastroesophageal junction, which was marked with a clip to direct the endosonographer away from erroneous esophageal puncture. All EUS-HGSs were performed or supervised by an experienced endosonographer using a convex echoendoscope (oblique-viewing EUS GU-UCT260; Olympus Medical Systems or EG-740UT; FUJIFILM) or a forward-viewing EUS (GU-UCT260J; Olympus Medical Systems). The echoendoscope was inserted into the stomach and manipulated under endoscopic, ultrasonographic, and fluoroscopic guidance to identify the relevant structures.

After identification of the dilated left intrahepatic bile ducts (IHBD), either B2 or B3 was punctured after excluding the presence of interposing vessels with color Doppler (Fig. 2B). Depending on the diameter of the targeted IHBD, either a 22-G (Expect Slimline; Boston Scientific Co.) or 19-G EUS-FNA needle (EZ shot 3 plus; Olympus) preloaded with a guidewire through an attached Y-connector was chosen for puncturing the gastric wall and across the intervening liver tissue into the bile duct. The 19-G needle was selected if the IHBD diameter was greater than 5 mm and a 22-G for the rest. A 0.025-inch guidewire (VisiGlide2; Olympus Medical Systems), M-Through (Medicos Hirata), or a 0.018 inch (Fielder 18; Olympus Medical Systems) was used depending on whether a 19-G or 22-G needle respectively was used for the puncture. The guidewire was subsequently manipulated into the bile duct under fluoroscopic and echoendoscopic guidance, and a small amount of contrast medium was injected through the Y-connector to confirm the appropriate placement of the needle and guidewire (Fig. 2C). Once confirmed, the guidewire was inserted towards the hilum and the needle was withdrawn while maintaining the position of the guidewire. An uneven double-lumen cannula (PIOLAX Medical Devices) was passed through the guidewire into the bile duct. Bile content aspiration was performed to confirm the locations of the intraductal catheter tips. The aspirated bile was sent for culture and sensitivity testing. If a 22-G needle and 0.018-inch wire were used, fistula dilation was necessary. The fistula tract was dilated using a non-cautery device. The choice of device for dilation, a 6-Fr tapered cannula mechanical dilator (ES dilator soft type; Zeon Medical Co., Ltd.), uneven double-lumen cannula (PIOLAX Medical Devices), or a drill dilator (7-Fr, Tornus ES; Asahi Intec), was at the discretion of the endosonographer.

Cholangiography was performed by using an uneven double-lumen cannula. Finally, a fully covered self-expandable metal stent (FCSEMS) (6 mm×12 cm, HANAROSTENT Benefit; M.I.Tech Co.,Ltd.) was placed and the implantation site was confirmed both fluoroscopically and endoscopically (Fig. 2D).

All patients were administered intravenous antibiotics before and after the procedure. Other post-procedure management steps included monitoring symptoms (fever, abdominal pain, vomiting, hematemesis, and melena), vital signs (blood pressure, pulse rate, and body temperature), physical examination, blood tests (full blood count, C-reactive protein, and bilirubin), and contrast-enhanced computed tomography within 24 hours after the procedure to identify early complications.

Patients were discharged when they had no complaints and were clinically stable and a downward trend in bilirubin levels was observed. After discharge, clinical reviews were arranged, and blood tests were performed to assess cholestasis approximately 2 and 4 weeks after the procedure.

The total duration of the procedure was measured from the time of gastroscope insertion to the completion of all procedures. The following outcomes were evaluated: (1) technical success was defined as successful placement of the stent via the stomach into the left IHBD, as planned and confirmed by endoscopy and computed tomography; (2) clinical success was defined as postprocedural improvement in cholangitis or reduction in the total serum bilirubin level by 50% or more from preprocedural values within 4 weeks14; (3) complications: clinical, laboratory and/or radiological evaluation consistent with any of the following: bleeding (reduction in the serum hemoglobin level to <2.0 g/L from the pretreatment value), bile leakage, biloma, stent migration, stent misplacement, intrahepatic hematoma, sepsis, peritonitis, pseudoaneurysm of the hepatic artery, and death. The severity of adverse events was classified into grades I–V according to the classification of adverse events for GastRointEstinal Endoscopy (AGREE).16

RESULTS

Ten patients with HCC who underwent 14 EUS-HGS attempts were included in this study. One patient underwent two procedures on the same day, whereas three patients underwent repeat procedures on different days ranging from 4 to 120 days apart.

The baseline patient characteristics are presented in Table 1. The mean age of the patients was 71.5 years with more males (80.0%) than females. Six patients presented with comorbidities: six (60.0%) had hypertension, and four (40.0%) had diabetes. Five (50.0%) had background cirrhosis and infection with viral hepatitis. Ascites was present in 5 of the 11 separate procedures. The mean total bilirubin was 130 (44.5–244.5) μmol/L, hemoglobin was 11.5 (7.3–14.5) g/L and C-reactive protein was 7.1 (0.6–17.9 mg/L). The mean baseline platelet count of 130.13 (60–270)×10⁹/L and international normalized ratio of 1.22 (1.0–1.52) did not warrant any pre-procedure correction of coagulopathy. None of the patients had cholangitis as an indication for EUS-HGS. Majority (69.2%) of the biliary obstructions occurred at the level of the hepatic hilum. The tumor, either by direct invasion or extrinsic compression of the bile duct, was the cause of the obstruction in 92.3% of cases. In the remaining 7.7%, the obstruction was caused by tumor thrombi. Three (60.0%) of the five patients with background cirrhosis were categorized in Child-Pugh class B. All cases were deemed inoperable, and four (36.4%) of the 11 distinct cases were classified as stage II according to the American Joint Commission on Cancer Staging System.

Table 1.

Baseline patient characteristics

The technical details of the procedure are presented in Table 2. Six (42.9%) of the attempted 14 EUS-HGS were performed as primary procedures without prior ERCP. In another six (42.9%), EBD had been attempted at another facility before referral to the Aichi Cancer Center. One patient underwent two EUS-HGSs (simultaneous B2 and B3) in the same session, and another patient underwent separate EUS-HGSs of B2 and B3 4 days apart. These procedures were performed to improve the quality of drainage or the percentage of drained liver segments. A third patient also underwent a second EUS-HGS (B3) 3 months after the initial diagnosis (B2) because of the recurrence of biliary obstruction and increased bilirubin levels. In one patient, the initial attempt at EUS-HGS failed because of the inability to correctly identify and puncture the bile duct. However, the second attempt 18 days later was successful. Therefore, an overall technical success rate of 92.9% was achieved. Technical success was 100% for the primary procedures.

Table 2.

Technical details of EUS-HGS procedures

The mean diameter of the IHBD in the 13 successful cases was 3.7 (2.0 – 6.4) mm. In 7 (53.8%) cases, B3 was punctured, and the 22-G needles and 0.018-inch guidewires were used. In three of these seven cases, the initial 0.018-inch guidewire was changed to a 0.025 guidewire to complete the procedure. A 6×120 mm FCSEMS was deployed in all cases. The mean procedure time was 32.7 (15–50) minutes.

Details of the postprocedural events are shown in Table 3. 50% or more reduction in baseline bilirubin levels was recorded in 8 (72.7%) by the end of week 2. Clinical success was reported in 10 (90.9%) of the 11 distinct procedures. Clinical success and other post-procedure details could not be assessed in two cases because they were too close to other procedures to be evaluated separately. All patients resumed eating within 24 hours. No adverse bleeding events were observed. However, other mild (AGREE grades I and II) adverse events were reported in 3 (27.3%) patients, all of which resolved with conservative management. Two patients with AGREE grade I required antipyretics, whereas a single patient with grade II required additional antibiotics to treat cholecystitis. The patient in whom the initial EUS-HGS was unsuccessful had cholecystitis that persisted at the time of the second attempt and continued thereafter. The mean duration of hospital stay was 11 (4–37) days.

Table 3.

Post-EUS events for 11 procedures^a)

DISCUSSION

Advancements in therapeutic options for patients with late-stage HCC warrant the identification of safer and more effective biliary drainage methods so that these patients can benefit from such therapies, enjoy jaundice palliation, and improve survival. In this study, we demonstrated that EUS-HGS is a technically feasible and highly effective procedure for patients with HCC that can be performed by experienced surgeons without causing bleeding. To the best of our knowledge, this is the first report of EUS-HGS as a drainage procedure performed solely in patients with HCC.

The technical success rate of 92.9% reported in this study was comparable to the 93.1% and 93.7 reported by Li et al.17 and Uemura et al.,18 respectively in meta-analyses of EUS-HGS for drainage of various unresectable malignant biliary obstructions. It was also not inferior to EBD procedures previously reported for HCC.5 In patients in whom EUS-BD failed initially, the procedure was successful on a second attempt. This high technical success rate was achieved despite the presence of ascites in three patients before the performance of EUS-HGS. Although moderate ascites is generally considered a contraindication, Yasuda et al.19 demonstrated that such a procedure is technically feasible and safe. The experience of the group undoubtedly played a role in high success rates.

Fistula dilatation before stent deployment is an important step in EUS-HGS. Mechanical dilatation using the ES dilator soft type or the novel tornus device was preferred over the technically easier to use electric cautery devices because of their superior safety.14,20,21 Generally, a balloon dilator is a good rescue device because it creates a larger tract; however, some expertise is required to avoid bile leakage. Electric cautery devices have been shown to have a higher risk of procedure-related bleeding; therefore, avoiding these devices in these cases was good.14 Thankfully, the cautery dilator was not needed as a rescue in these cases.

Although a 19-G needle and 0.035 or 0.025 guidewires are favored during EUS-HGS, in this study, the 22-G needle and 0.018-inch wires were successfully used to start and finish the procedure in 4 of 13 procedures. Although more technically difficult to use, a smaller needle offers less traumatic puncture and, therefore, probably contributes to a lower bleeding risk, especially in the setting of cirrhosis. Experience with smaller needles and guidewires has been previously reported.22,23

The clinical success rate of 90.9% reported in this study is superior to clinical success rate of 35% to 81% reported for EBD drainage procedures.5,6,24 This was also higher than the 84.5% reported by Uemura et al.18 in a meta-analysis of EUS-BD for all malignant biliary obstructions. Successful drainage can provide patients with HCC relief from the symptoms of obstructive jaundice, particularly intense pruritus, and the opportunity to receive additional chemotherapy. Therefore, successful biliary drainage improves both patient survival and quality of life. In 8/11 (72.7%) of cases in this study, significant (≥50%) reduction in baseline bilirubin levels were achieved within 2 weeks; therefore, these patients could potentially start or restart chemotherapy after this period.

Although EBD is the gold standard for unresectable malignant biliary obstructions, several studies have demonstrated the utility of EUS-HGS as the primary drainage procedure for malignant biliary obstructions.12,25,26 Our study adds to the body of evidence that EUS-HGS is technically feasible and safe as a primary procedure, even for HCC. Technical success rate was 100% for all the seven primary EUS-HGS procedures, and clinical success was achieved in five out of six patients.

There is a higher risk of stent migration and clogging by food materials after successful stent placement with EUS-HGS than with EUS-CDS or PTBD. However, stent patency has been reported to be longer with EUS-HGS than with PTBD, ranging from 62 to 402 days.27 Longer stent patency is associated with factors such as the stent type and length. In this study, an FCSEMS with a length of 12 cm was used to conform to the recommendations. FCSEMS have a wider diameter and are less susceptible to tumor ingrowth. A long stent also allows an intragastric stent length of >3 cm, which reduces the risk of inward stent migration.

The reported total complication rate of 27.3% was higher than the pooled early adverse event rate of 18.2% reported by Isayama et al. for EUS-HGS of other malignant biliary strictures.14 This was better though than the adverse event rate of 35% and mortality of 5% reported by Winkler et al. on the feasibility of EUS-HGS for similarly inoperable malignant hilar strictures.26 In this study, all three complications were mild and resolved with conservative management. No bleeding was observed. This is significant because most biliary obstructions were caused by tumor compression of the bile duct at the hilum. Drainage through the EBD route would have meant traversing the tumor and significantly increasing bleeding risk. In a previous study that used EBD in HCC patients, a bleeding rate of 2.9% and 18.6% incidence rate of post ERCP pancreatitis were reported.5 One patient, in whom the procedure technically failed on the first attempt, developed cholecystitis. This delayed reintervention by approximately 18 days, after which a repeat EUS-HGS was successfully performed. The course of cholecystitis prolonged the hospital stay to 37 days. The mean duration of the hospital stay after the procedure was 11 days. This is longer than the mean durations of 7.5 and 8 days of hospital stay reported by Winkler et al.26 and Sportes et al.,13 respectively. However, this confirms the findings of Sportes et al.13 that EUS-HGS is associated with a shorter hospital stay than PTBD (15 days).

In addition, all patients were able to resume feeding within 24 hours after the procedure. This early resumption of feeding and the observed shorter duration of hospital stay certainly contributed to a better quality of life, which has been severally reported with EUS-HGS over other EBD salvage procedures.28

The relatively small number of patients from a single center in this study means that extrapolation of the results must be performed with caution. However, the baseline characteristics of the patients in this study mirror those previously reported for patients with HCC in Japan: more males, a mean age of more than 70 years, and a predominant HCV association.29

Clinical success could not be determined when two separate procedures were performed on two different patients. In one patient, a repeat procedure was performed less than 4 weeks after the initial procedure, while in another patient, two procedures were performed at the same time. Therefore, it is impossible to distinguish the clinical success of these procedures separately.

In conclusion, EUS-HGS is a technically feasible and clinically effective procedure that can be used to drain biliary obstructions in patients with hypervascular HCC. This can be performed without causing bleeding by experienced hands. It can be considered the primary drainage procedure or a salvage procedure when others fail. Further research is recommended to explore the survival benefits of this procedure for different stages of HCC and liver disease severities.

Article information Continued