Hepatocellular carcinoma (HCC) is the sixth most common cancer worldwide (626 000 new cancer cases per year). Given the poor prognosis, the number of deaths is almost the same (598 000 cases per year).1 Especially, the liver is a common place for metastases — primary from cancers of the colon, breast, lung and stomach. Liver metastasis is a major cause of cancer-related mortality.2 Although surgery is the first-line treatment for liver tumors, only 8–37% patients are suitable for surgery resection, with 5-year survival rate of 25–39%.2–5 Other effective treatment modalities are needed in patients with unresectable liver tumors.
Radiofrequency ablation (RFA), as a minimally invasive local treatment, has advantages such as relatively low risk, feasibility in non-surgical candidates, and cost-effective. Therefore, RFA has become a widely accepted treatment modality for small and unresectable liver tumors.2,4 However, even though it is a minimally invasive local treatment, awareness and management of complications are important.6–9 Among the complications, biliary injury is one of the most devastating complications that can cause cholangitis, liver failure and even death.7–13 To the best of our knowledge, few studies have focused on risk factors for developing biliary injury associated with RFA and its long-term effect on the overall survival. Therefore, the purpose of this study was to clarify the incidence, severity, prognosis and risk factors of biliary injury associated with RFA in patients with hepatic malignancies through long-term follow-up.
This clinical study was approved by the Institutional Ethics Committee, School of Oncology of Peking University (Beijing, China). Written informed consent was obtained from all patients enrolled in the study.
Between June 2001 and January 2009, a total of 698 consecutive patients underwent ultrasound-guided percutaneous RFA during this period. All patients were prospectively registered on an electronic database. The records were reviewed retrospectively until the end of January 2010.
The final diagnosis of HCC was made by ultrasound-guided fine needle aspiration biopsy on at least one tumor. Histopathological grading of tumor differentiation was performed according to Edmondson et al.14 Histological grading was obtained in all HCC patients. For liver metastasis, most diagnoses of liver tumors were based on histopathological findings of liver specimens (150 patients, 64.4%). The remaining tumors (83 patients, 35.6%) were diagnosed based on typical computed tomography (CT) and (or) magnetic resonance imaging (MRI) scans, with consideration of elevated tumor markers (carcino-embryonic antigen (CEA), carbohydrate antigen 19–9 (CA 19–9)).
All the patients met the following criteria for RFA treatment: (1) nodular tumor≤7 cm in maximum diameter, multi-nodular (up to five in number) tumors ≤3 cm in maximum diameter each, (2) tumors visible on ultrasound, (3) the absence of portal venous thrombosis, (4) a prothrombin time ratio greater than 50% (prothrombin time with international normalized ratio <1.7), a platelet count >50 000/μl, and (5) extrahepatic metastases that had been surgically resected or locally controlled.
After reviewing the database, patients with insufficient clinical data (n=16), or who did not meet the diagnosis of bile duct injury (n=14) or with a follow-up less than 12 months (n=30) were excluded. Finally, 638 patients (405 HCCs and 233 liver metastases) who underwent 955 sessions of RFA were included. The demographic and clinico-pathologic data of 638 patients with hepatic malignancies treated with RFA were summarized in Table 1. Among the 405 HCC patients, 243 patients had stage I/II disease, 162 patients had III/IV disease according to the AJCC-TNM system. Forty-one patients were identified with at least one tumor located centrally (tumor-bile duct distance of 5–10 mm). Among the 233 liver metastasis, the original tumors were located in the colon and rectum (n=121), stomach (n=28), breast (n=31), lung (n=20), pancreas (n=8), and other organs (n=25).
Twelve patients with liver metastasis were identified with at least one tumor located centrally.
RFA procedure and treatment strategies
The RFA system used in this study was a 460-kHz generator (Model 1500; Rita Medical Systems, Mountain View, CA, USA). Expandable electrodes consisted of an outer 14-gauge insulated needle 15 cm in length. Nine prongs were deployed and retracted by a movable hub, and deployment diameters ranged from 3 cm to 5 cm. The time required to produce a 5-cm ablation sphere was approximately 20 minutes. An Aloka SSD-4000 or α 10 Ultrasonography System (Aloka Co., Ltd., Tokyo, Japan) with a 3.5–5.0 MHz probe was used to guide the percutaneous RFA.
All RFA was performed under real-time sonographic guidance by two radiologists, who had more than 10-year experience of interventional radiology. The ablative margin enveloped the entire tumor, as well as a 5–10 mm margin of surrounding normal tissue. In lesions larger than 3.5 cm, overlapping ablations were used. Individualized treatment strategies for tumors located at problematic locations were used whenever possible.15,16
We placed no restriction on RFA except for the location of tumors within 5 mm of the central bile duct.17 When we treated the lesions adjacent to the bile duct, we inserted the expandable electrodes in the direction parallel to the wall of the bile ducts to avoid prongs being accidentally inserted into the bile duct. This procedure was performed by carefully monitoring the relationship between the electrode tip and bile duct under real-time ultrasound scanning (Figure 1). The RF generator was not turned on until we were confident that the prongs had not penetrated adjacent structures.
All patients underwent percutaneous RFA under moderate sedation anesthesia, induced with intravenous administration of 2.5–5.0 mg of midazolam (Roche; Basel, Switzerland) and 50–100 μg of fentanyl (Fentaini; Renfu, Yichang, China). Local infiltration anesthesia was induced by 5–15 ml of 1% lidocaine (Liduokayin; Yimin, Beijing, China). When tumors were adjacent to the diaphragm or the liver surface, an intravenous bolus of propofol (Diprivan; Zeneca, Macclesfield, UK, 1–2 mg/kg) and fentanyl (50–100 μg) were given to enhance anesthesia in combination with local anesthesia.
After treatment, these patients were usually monitored for 2–4 hours to confirm the absence of intra-abdominal active bleeding signs before returning to their inpatient room. If there were no acute major complications, in-patients were discharged from the hospital 2 days later.
Clinical evaluation and follow-up
One month after RFA, a repeated blood test and tumor markers’ (α-fetoprotein (AFP), CEA, CA 19–9) detection were conducted. CT or MRI was performed on all patients 1 month after RFA, together with ultrasound. Imaging and laboratory follow-ups were conducted every 3 months during the first year, and every 4–6 months during the following years. Follow-up CT/MRI imaging studies were read by two radiologists with more than 10 years of experience in reading liver scans.
Early tumor necrosis was identified if no enhancement was observed with well-defined margins. Residual unablated tumors were defined as irregular peripheral enhancing foci in the ablation zone on either contrast-enhanced CT or MRI. Additional RFA was required for residual unablated tumors.
Local recurrence was defined as enhancement in the periphery of the RFA-treated area, and remote HCC recurrence was defined when a new liver lesion other than the RFA-treated area was detected. When recurrences were confirmed, the tumors were usually treated with additional RFA if possible.
Diagnosis of biliary injury
Patients did not receive any treatment that would induce biliary injury before RFA, such as percutaneous transhepatic cholangiography, endoscopic retrograde cholangiopancreatography, and biliary surgery. For patients who had received transcatheter arterial chemoembolization (TACE) treatment, the time interval between the most recent RFA and TACE should be more than 3 months.18 The most common type of biliary injury was dilatation of the intrahepatic bile ducts, which included a dilated bile duct diameter greater than or equal to the adjacent portal vein branch on ultrasound (as well as color Doppler flow imaging) or CT. The conditions described above were considered as biliary injury associated with RFA. The interval between the occurrence of biliary injury and the most recent RFA time was defined as the onset time of biliary injury. Dilatation of a bile duct that had existed before RFA, or that was because of biliary invasion of tumor, was not considered as an event of biliary injury.19
In our database, we focused on unique complications, and therefore, we graded biliary injuries into three levels according to severity as follows.20 Mild injury involved only imaging findings without increased serum bilirubin levels or slightly abnormal bilirubin levels (<17.1 U/L). Moderate injury involved imaging findings and abnormal bilirubin levels (17.1–34.2 U/L), but interventional therapy or prescription medicine was required. Severe injury involved persistent complications, and a trial of interventional radiology or surgery was required, with increased bilirubin levels (>34.2 U/L).
Evaluation of risk factors and overall survival
To assess the impact of biliary injury, overall survival was described both patients with biliary injury and those without biliary injury associated with RFA.
We defined multiple conditions as potential risk factors for biliary injury associated with RFA. The clinical data analysis consisted of liver background, the final pathological diagnosis of the tumor, Child-Pugh grading, TNM staging, pathological classification, tumor size, tumor number, treatment session, and tumor location. We also noted the coexistence of each predisposing factor for every RFA procedure.
Continuous variables were expressed as median (range). Comparison of categorical variables was performed using the chi-square test (or Fisher's exact test where appropriate). Multivariate analysis of potential risk factors was performed using the stepwise Logistic regression model. Survival time was calculated from the time of the first RFA to the date of death or last follow-up. The cumulative survival rate was calculated by the use of Kaplan-Meier curves. P <0.05 was considered statistically significant. All data were analyzed with SPSS 13.0 software (SPSS Inc., Chicago, USA).
Biliary injuries developed in 17 patients after 17 RFA procedures (17/955 (1.8%), 10 sessions with HCC and 7 sessions with liver metastasis patients). There was no significant difference between the HCC group and the liver metastasis group with respect to the degree of biliary complications (chi-square test, χ2=0.029, P=0.864) (Table 2). The incidence of major (moderate and severe) complications was 0.8% per session. The average onset time was 12 weeks (2–36 weeks).
Focal peripheral intrahepatic bile duct dilation was detected in 9 cases of the mild group. One patient showed progression during follow-up, from focal to multiple intrahepatic dilation. The patient is still alive and with ongoing follow-ups (Figure 2). The other 8 patients showed stable changes in the bile ducts with no progression on follow-up.
In 6 patients of the moderate group, a biloma occurred in a male HCC patient with liver cirrhosis and diabetes. Five months after RFA, CT and ultrasound both showed biloma formation, and multiple percutaneous fine-needle aspiration was performed and the jaundice gradually eased (Figure 3). There was also a case with biliary tract infection associated with local hepatic necrosis and focal biliary dilatation. The patient successfully recovered after treatment with intravenous antibiotics. The remaining 4 patients with bile duct dilation and jaundice (direct bilirubin / total bilirubin >50%) were given percutaneous transhepatic biliary drainage or percutaneous fine-needle aspiration. After the above treatments, the jaundice in the moderate group gradually eased or disappeared.
Two patients in the severe group were metastasis with bile leak. They were given external drainage and antibiotic treatment. However, the 2 fistulas did not heal until the time of death. The 2 patients survived for 7 and 11 months, and died of systemic metastasis.
Risk factors analysis
Tables 3 and 4 show the univariate analysis results of risk factor analysis by Fisher's exact test. In HCC patients, statistically significant risk factors for biliary injury associated with RFA included tumor size (P=0.027), the distance between the tumor border and the nearest bile duct (P <0.001), multiple treatments (≥4) and infiltration of adjacent vessels (P <0.001), whereas TNM stage, tumor pathological grading and the number of lesions had no significant association on risk of biliary injuries associated with RFA. In metastasis patients, risk factors of biliary injury included the site of the primary tumors (P=0.036) and tumor located centrally (P=0.010).
Multivariate analysis identified that vessel infiltration (OR=4.816, P=0.034), and treatment sessions ≥4 sessions (OR=34.62, P=0.026) were independent risk factors for biliary injury associatied RFA of HCC, while tumor located centrally (OR=11.564, P=0.043) was the only independent risk factor in the metastasis group.
Using the Kaplan-Meier method for survival analysis, the median survival time of the HCC patients with/without biliary injuries was 39 months (95% CI: 31.5–46.5) and 47 months (95% CI: 22.2–71.8), respectively. The median survival time of liver metastasis patients with/without biliary injury was 23 months (95% CI: 18.9–27.1) and 18 months (95% CI: 12.9–23.1) respectively. Although statistical analysis was not applicable because of the small number of patients with biliary injuries, there seemed no notable difference in overall survival between patients with and without biliary injuries.
RFA has gained increasing clinical attention in modern management of malignant liver tumors nowadays.2,4 As an increasing amount of centers use RFA, the incidence of serious complications can reach 2.2–10.0%, with the death rate up to 0.5%.8–13 Of these, biliary injury is one of the most severe complications that may result in stenosis, bile leak, biloma formation, bile fistula formation, cholangitis and even death.9,10,12,13 However, the complication of biliary injury associated with RFA has not been fully investigated.
Incidence of biliary injury
According to this retrospective study with a long-term follow-up, the overall incidence of biliary injury was 1.8% per session, with a major complication rate of 0.8%, which was lower than those in previous studies.12,19 Distal bile duct dilation is the most common type of mild and moderate biliary injury complications. In our study, during follow-up, most of the patients showed no change or had slight progression. After long-term follow-up, we found that patients with biliary injuries had a similar survival time as patients without biliary injury.
In fact, not all biliary injuries are biliary complications. Kim et al19 reported that bile duct changes after RFA of HCC patients were frequent (12% per session), that most were of no clinical significance, and that major complications requiring additional treatment were rare. Mild biliary injuries after RFA are clinically negligible and need not to be regarded as a complication. For moderate and severe biliary injury, with timely percutaneous drainage or proper drug treatments as well as close follow-up, the outcome of the patients was acceptable.
In previous studies, only Machi et al12 reported that biliary injuries was the main type of complications after RFA (46 patients with 204 tumors). Their RFA inclusion criteria were loose, compared with other similar report. Their cases included large tumors, major vessel / main bile duct invasion, extrahepatic metastasis and decompensated liver function. RFA was performed under percutaneous, laparoscopic, or intraoperative ultrasound guidance. Except for the fact that one patient died of liver failure (1.7%, 1/60 sessions), three cases of major complications were biliary injuries (5.0% per session), and all were from metastatic carcinomas. One case of bile leakage required prolonged percutaneous drainage and placement of an endoscopic biliary stent, and 2 cases of biliary stricture were the result of thermal injury.
Mechanism of biliary injury
The possible mechanism of biliary injury after RFA is considered as thermal damage to the bile duct. Bile flow is slower, and has little cooling effect compared with blood flow. Therefore, bile ducts are vulnerable to heat damage. It is considered that heat from RF energy produces nonspecific thermal injury to tumor tissue as well as to adjacent normal structures. Heat damages bile ducts and results in either ductal stenosis with dilation of the proximal bile duct, or abscess formation.19
Onset time of biliary injuries
In clinical trials, biliary injury usually develops slowly and it can be as late as 9 months after RFA.19 According to standardized terminology and reporting criteria,20 biliary injury (2–36 weeks) can be a periprocedural complication (within 30 days) and a delayed complication (more than 30 days after ablation). Therefore, detection of biliary injury requires close follow-up during the first year.
Wang et al18 reported that TACE-induced biliary injury usually occurred from 3 weeks to 3 months among 1240 patients with hepatic malignancies who had undergone a total of 2680 TACE procedures. So, in order to eliminate the interference of TACE at the maximum extent, we set the inclusion criteria that the time interval between the most recent RFA and TACE should be more than 3 months. However, for patients who had receive comprehensive treatments, further strictly controlled trials may provide more definitive evidence about the real onset time.
Risk factors analysis
To minimize the complication of biliary injuries, it is important to understand the risk factors associated with biliary injuries. Central location of tumors was found to be the most important factor in previous studies.9,21 Ohnishi et al17 reported a 46% incidence of bile duct injury following RFA of HCC within 5 mm of the central bile duct. Unfortunately, it was not the case in the present study. Our analysis showed that a central location of tumors was a potential risk factor by univariate analysis but it was not identified as an independent risk factor by multivariate analysis in HCC patients because of other coexistent factors; for example, tumor size and treatment session. It is not surprising that the risk of biliary injury increases along with a larger tumor size and more treatment sessions. In addition, intrahepatic bile ducts are lined by cuboidal epithelium near the lobules with the supportive connective tissue. The epithelium gradually becomes columnar closer to porta hepatis, while smooth muscle fibers emerged in the extrahepatic bile duct wall.22 Smooth muscle is less vulnerable to thermal injuries than connective tissue. Therefore, centrally located tumor and peripheral tumors might not affect biliary injury in some patients.
In our study, we found that vessel infiltration and treatment sessions ≥4 times were independent risk factors for biliary injury associated with RFA of HCC. Some investigators have proposed that if the blood flow in the portal vein was decreased because of the Pringle maneuver, portal vein thrombosis or vascular injury, the risk of bile duct stenosis may increase, however, there is no clinical evidence to support this notion.23 In this study, we found that vessel infiltration was a potential risk factor in HCC patients. The finding is not surprising since tumors close to vessels are also close to bile ducts and bile injury can easily occur after RFA.
Regarding to RFA of liver metastasis, we found an interesting result that the original site of the pancreas was a risk factor for biliary injury by univariate analysis. It should be noted, however, that some of these tumors were also located on the liver surface. Possible factors such as a superficial lesion, the number of cells dislodged, adhesiveness of the cells, enzyme release, and immunological characteristics of tumor cells may play a role.24 Considering the small number of cases and the potential hazards, including biliary fistula, further studies with more cases are required to confirm this finding.
Cooling of the biliary tract with chilled saline or prophylactic biliary stent placement can be attempted for central tumors during RFA treatment, however, the auxiliary methods are not widely used and both have certain risks.17,21,25 Other therapies, such as TACE, or percutaneous ethanol injection, can be alternative treatments in patients with identified prognostic factors and RF machines that can produce a larger volume of coagulation necrosis at one time are needed to reduce biliary injuries. Our strategy of inserting the needle parallel to the wall of bile ducts was adopted to decrease the risk of biliary injury and the outcome was acceptable.
This study has some limitations. (1) The retrospective nature of this study may be considered an important limitation, and the data should be confirmed by prospective studies. (2) Our study was based on the experience of a single center, thereby limiting patient numbers and patient selection to local practice. A multicenter study may improve the objectivity of the evaluation of biliary injury. (3) Biliary injury was not proved pathologically proven in the mild group, because these patients did not present any clinical symptoms. Biliary injury was confirmed based on imaging findings and follow-up.
In conclusion, we found that the incidence of biliary injury was infrequent (1.8%) and it seemed not have significant impact on the patients’ long-term survival. Vessel infiltration and treatment sessions ≥4 times were independent risk factors for biliary injury associatied with RFA of HCC, while tumor located centrally was an independent risk factor in the metastasis group. More attention needs to be paid to pre-RFA case selection to help predict biliary complications and control them.
1. Parkin DM, Bray F, Ferlay J, Pisani P. Global Cancer Statistics, 2002. CA Cancer J Clin 2005; 55: 74-108.
2. Ng KK, Poon RT. Radiofrequency ablation
for malignant liver tumor. Surg Oncology 2005; 14: 41-52.
3. Bradley AL, Chapman WC, Wright JK, Marsh JW, Geevarghese S, Blair KT, et al. Surgical experience with hepatic colorectal metastasis. Am Surg 1999; 65: 560-566.
4. Wang YH, Liu JF, Li F, Li A, Liu Q, Liu DB, et al. Radiofrequency ablation
combined with transarterial chemoembolization for unresectable primary liver cancer. Chin Med J 2009; 122: 889-894.
5. Tsim NC, Frampton AE, Habib NA, Jiao LR. Surgical treatment for liver cancer. World J Gastroenterol 2010; 16: 927-933.
6. Liang P, Wang Y, Yu X, Dong B. Malignant liver tumors: treatment with percutaneous microwave ablation — complications
among cohort of 1136 patients. Radiology 2009; 251: 933-940.
7. Chen MH, Dai Y, Yan K, Yang W, Gao W, Wu W, et al. Intraperitoneal hemorrhage during and after percutaneous radiofrequency ablation
of hepatic tumors: reasons and management. Chin Med J 2005; 118: 1682-1687.
8. Livraghi T, Solbiati L, Meloni F, Ierace T, Goldberg SN, Gazelle GS. Treatment of focal liver tumors with percutaneous radio-frequency ablation: complications
encounterd in a multicenter study. Radiology 2003; 226: 441-451.
9. Mulier S, Mulier P, Ni Y, Miao Y, Dupas B, Marchal G, et al. Complications
of radiofrequency coagulation of liver tumors. Br J Surg 2002; 89: 1206-1222.
10. Chen TM, Huang PT, Lin LF, Tung JN. Major complications
of ultrasound-guided percutaneous radiofrequency ablations for liver malignancies: single center experience. J Gastroenterol Hepatol 2008; 23: 445-450.
11. Kasugai H, Osaki Y, Oka H, Kudo M, Seki T; Osaka Liver Cancer Study Group. Severe complications
of radiofrequency ablation
therapy for hepatocellular carcinoma: an analysis of 3891 ablations in 2614 patients. Oncology 2007; 72 Suppl 1: 72-75.
12. Machi J, Uchida S, Sumida K, Limm WM, Hundahl SA, Oishi AJ, et al. Ultrasound-guided radiofrequency thermal ablation of liver tumors: percutaneous, laparoscopic, and open surgical approaches. J Gastrointest Surg 2001; 5: 477-489.
13. Kong WT, Zhang WW, Qiu YD, Zhou T, Qiu JL, Zhang W, et al. Major complications
after radiofrequency ablation
for liver tumors: analysis of 255 patients. World J Gastroenterol 2009; 15: 2651-2656.
14. Edmondson HA, Steiner PE. Primary carcinoma of the liver. A study of 100 cases among 48 900 necropsies. Cancer 1954; 7: 462-503.
15. Chen MH, Yang W, Yan K, Zou MW, Solbiati L, Liu JB, et al. Large liver tumors: protocol for radiofrequency ablation
and its clinical application in 110 patients. Radiology 2004; 232: 260-271.
16. Chen MH, Yang W, Yan K, Hou YB, Dai Y, Gao W, et al. Radiofrequency ablation
of problematically located hepatocellular carcinoma: tailored approach. Abdom Imaging 2008; 33: 428-436.
17. Ohnishi T, Yasuda I, Nishigaki Y, Hayashi H, Otsuji K, Mukai T, et al. Intraductal chilled saline perfusion to prevent bile duct injury
during percutaneous radiofrequency ablation
for hepatocellular carcinoma. J Gastroenterol Hepatol 2008; 23: 410-415.
18. Wang MQ, Shao RH, Ye HY, Wang ZQ, Wang ZP, Liu FY. Investigation of bile duct injury
after transcatheter arterial chemoembolization. Chin J Oncol (Chin) 2005; 27: 609-612.
19. Kim SH, Lim HK, Choi D, Lee WJ, Kim SH, Kim MJ, et al. Changes in bile ducts after radiofrequency ablation
of hepatocellular carcinoma: frequency and clinical significance. Am J Roentgenol 2004; 183: 1611-1617.
20. Goldberg SN, Grassi CJ, Cardella JF, Charboneau JW, Dodd GD 3rd, Dupuy DE, et al. Image-guided tumor ablation: standardization of terminology and reporting criteria. J Vasc Interv Radiol 2009; 20 Suppl 7: s377-s390.
21. Ogawa T, Kawamoto H, Kobayashi Y, Nakamura S, Miyatake H, Harada R, et al. Prevention of biliary complication in radiofrequency ablation
for hepatocellular carcinoma-Cooling effect by endoscopic nasobiliary drainage tube. Eur J Radiol 2010; 73: 385-390.
22. Hong SM, Kang GH, Lee HY, Ro JY. Smooth muscle distribution in the extrahepatic bile duct: histologic and immunohistochemical studies of 122 cases. Am J Surg Pathol 2000; 24: 660-667.
23. Denys AL, De Baere T, Mahe C, Sabourin JC, Sa Cunha A, Germain S, et al. Radio-frequency tissue ablation of the liver: effects of vascular occlusion on lesion diameter and biliary and portal damages in a pig model. Eur Radio 2001; 11: 2102-2108.
24. Smith EH. Complications
of percutaneous abdominal fine-needle biopsy. Radiology 1991; 178: 253-258.
25. Wood TF, Rose DM, Chung M, Allegra DP, Foshag LJ, Bilchik AJ. Radiofrequency ablation
of 231 unresectable hepatic tumors: indications, limitations, and complications
. Ann Surg Oncol 2000; 7: 593-600.
Keywords:© 2011 Chinese Medical Association
hepatic malignancy; radiofrequency ablation; biliary tract; injury; complications