This article was presented at the 41st Annual Meeting of the American Society for Therapeutic Radiology and Oncology, San Antonio, Texas, October 31–November 4, 1999.
Unresectable hepatocellular carcinoma treated with radiotherapy and/or chemoembolization
Version of Record online: 21 JUN 2001
Copyright © 2001 Wiley-Liss, Inc.
International Journal of Cancer
Volume 96, Issue 4, pages 243–252, 20 August 2001
How to Cite
Chia-Hsien Cheng, J., Chuang, V. P., Cheng, S. H., Lin, Y.-M., Cheng, T.-I., Yang, P.-S., Jian, J. J.-M., You, D.-L., Horng, C.-F. and Huang, A. T. (2001), Unresectable hepatocellular carcinoma treated with radiotherapy and/or chemoembolization. Int. J. Cancer, 96: 243–252. doi: 10.1002/ijc.1022
- Issue online: 19 JUL 2001
- Version of Record online: 21 JUN 2001
- Manuscript Accepted: 20 APR 2001
- Manuscript Revised: 19 APR 2001
- Manuscript Received: 17 JAN 2001
- National Science Council, Executive Yuan, Taiwan, Republic of China. Grant Numbers: NSC 89-2320-B-368-003-M08, NSC 89-2314-B-368-003-M08
- hepatocellular carcinoma;
- pattern of failure
The purpose of our study was to evaluate the outcome, patterns of failure, and toxicity for patients with unresectable hepatocellular carcinoma (HCC) treated with radiotherapy, transcatheter arterial chemoembolization (TACE), or combined TACE and radiotherapy. Forty-two patients with unresectable HCC were treated with combined radiotherapy and TACE (TACE+RT group, 17 patients), radiotherapy alone (RT group, 9 patients), or with TACE alone (TACE group, 16 patients). Mean dose of radiation was 46.9 ± 5.8 Gy in a daily fraction of 1.8 to 2 Gy, directed only to the cancer-involved areas of the liver. TACE was performed with a combination of Lipiodol, doxorubicin, cisplatin, and mitomycin C, followed by Gelfoam or Ivalon embolization. Tumor size was smaller in the TACE group (mean: 5.4 cm) compared with the TACE+RT group (8.6 cm) and the RT group (13.1 cm) (P = 0.0003). The median follow-up was 24 months in the TACE+RT group, 28 months in the RT group, and 23 months in the TACE group. Survival was significantly worse for patients treated with radiotherapy alone due to the selection bias of patients with more advanced disease and compromised condition in this group. In contrast, the TACE+RT and TACE groups had comparable survival (two-year rates: TACE+RT 58%, TACE 56%, P = 0.69). The local control rate for the treated tumors was similar in the TACE+RT and TACE groups (P = 0.11). The intrahepatic recurrence outside the treated tumors was common and similar between these two groups (P = 0.48). The extrahepatic progression-free survival was significantly shorter for patients in the TACE+RT group than in the TACE group (two-year rates: TACE+RT 36%, TACE 100%, P = 0.002). Seven patients died from complications of treatment. Local radiotherapy may be added to treat patients with unresectable HCC, and the control of progression of the treated tumors was promising even in patients with large hepatic tumors. Survival of patients with combined TACE and radiotherapy was similar to that with TACE as the only treatment, while a significant portion of the patients treated with radiotherapy developed extrahepatic metastasis. ©2001 Wiley-Liss, Inc.
Hepatocellular carcinoma (HCC) is one of the most common malignancies in Asian countries . Surgical resection is potentially the only form of curative treatment. However, less than 20% of patients are surgical candidates at diagnosis [2–4]. Nonsurgical treatment options include transcatheter arterial chemoembolization (TACE) and percutaneous acetic acid or ethanol injection therapy. They are sometimes unsatisfactory, especially for patients with portal vein thrombosis or large infiltrative hepatic tumors [5, 6]. Repeated treatments are often necessary [7, 8]. Several institutions have reported promising results in patients with unresectable HCC treated with radiotherapy to a portion of the liver [9–12]. With the advances of three-dimensional conformal planning technique, local radiation treatment to the liver has become safer [13, 14]. Although the effect is more prominent within the site to which radiation is given, most patients have recurrence elsewhere within the liver or develop extrahepatic metastasis [9, 10, 15]. We retrospectively reviewed 42 patients with unresectable HCC at a single institution, treated with TACE alone or with radiotherapy as part or all of their treatments. The outcome, patterns of failure, and toxicities are analyzed in this article.
MATERIALS AND METHODS
From March 1994 through December 1997, 42 patients with unresectable HCC had their treatments at Koo Foundation Sun Yat-Sen Cancer Center, Taipei, Taiwan. Unresectability was determined by the participating surgeons who used currently accepted surgical criteria . Thirty-seven patients (88%) had histological confirmation of HCC. The remaining five patients had no tissue confirmation due to refusal of biopsy, but the imaging findings were compatible with HCC and their alpha-fetoprotein (AFP) levels were markedly elevated (>200 ng/ml). Thirty-four patients were male and eight female. Mean age was 58 years, ranging from 26 to 82. Performance status by the Karnofsky scale was above or equal to 70% in all patients. According to Child-Pugh classification for cirrhosis of the liver, 40 patients were in class A and two in class B. Seventeen patients had abnormally elevated level of AFP.
All patients had no evidence of extrahepatic metastasis before treatment. Mean size of the treated hepatic tumor, defined as the mean of three largest diameters in longitudinal, horizontal, and vertical directions on CT scan, was 8.4 ± 5.0 cm. Portal vein thrombosis before treatment was present in eight patients, confirmed either by CT scan or by the angiogram. According to the American Joint Committee on Cancer (AJCC) staging system, 14 patients were in stage II, nine in stage IIIA, and 19 in stage IVA.
Twenty-six patients underwent local radiotherapy as part or all of their treatments. Seventeen patients (TACE+RT group) had combined TACE and radiotherapy, whereas nine patients had radiotherapy alone (RT group). The entry requirements for radiation treatment included: (i) disease confined to the liver but surgically unresectable due to main portal vein thrombosis, inferior vena cava invasion, main hepatic vein invasion, obstructive jaundice, or tumor involving one or both lobes of liver with inadequate hepatic function reserve for surgery; (ii) Karnofsky performance scale ≥ 70%; (iii) indocyanine green (ICG) retention rate ≤ 30% at 15 min; (iv) absence of uncontrollable ascites; (v) no prior radiotherapy to the liver.
The other 16 patients (TACE group) were treated with TACE alone. Patient characteristics are compared and shown in Table 1. Differences were significant in tumor size, Child-Pugh classification, serum type of chronic hepatitis, and borderline significant in performance status. Patients in TACE+RT group and the RT group had significantly larger hepatic tumors than those in the TACE group. The difference is statistically significant (8.6 cm and 13.1 cm vs. 5.4 cm, P = 0.003). The only two patients in Child B classification of cirrhosis of liver were both in the RT group (P = 0.02). No patient in the RT group had a Karnofsky performance scale > 80. The differences in tumor size, Child-Pugh classification, and performance status between the three groups reflected the selection bias during the study period. Patients with larger hepatic tumors, more advanced disease, and relatively compromised condition frequently were not candidates for TACE and tended to have radiotherapy alone. For this reason, the statistical comparisons in patterns of failure and survival outcome were performed only between the TACE+RT and the TACE groups.
|All patients (no.)||17||9||16|
|Median follow-up (months)||24||28||23|
|Karnofsky scale (no.)|
|Child-Pugh classification (no.)|
|Serum hepatitis antigen marker (no.)|
|None or unknown||0||4||1|
|Tumor size (cm)|
|Mean||8.6 ± 4.1||13.1 ± 3.3||5.4 ± 4.5||0.003|
|AJCC stage (no.)|
|Alfa-fetoprotein level (ng/ml)|
|Portal vein thrombosis (no.)|
The radiation treatment technique has been previously reported . Radiation therapy was given to the hepatic tumor, either as the only treatment or as an adjunct to TACE/ethanol injection. Radiation portals were designed to include the gross hepatic tumor on the CT scan with 1.5 to 2 cm margins. No patient was given radiation to the whole liver. Three-dimensional CT-based computerized treatment planning was frequently used to determine the best combination of coplanar and noncoplanar portals. The dose-volume criteria of radiation treatment are described in Table 2. Radiation therapy was delivered with a 6-MV or 18-MV linear accelerator. Mean dose of radiation was 46.9 ± 5.8 Gy (range: 36.0–61.6 Gy), in a daily fraction of 1.8 to 2 Gy. Weekly verification films were obtained.
|Nonirradiated liver||Indocyanine green retention rate at 15 min|
|<1/3||40 Gy||No RT||No RT|
|1/3–1/2||50 Gy||40 Gy||No RT|
|> 1/2||60–66 Gy||50 Gy||40 Gy|
TACE was designed for every patient, unless contraindicated. The relative contraindications of TACE included main portal vein thrombosis, serum bilirubin level > 3 mg/dl, and hepatofugal portal flow shown on angiogram. In the TACE+RT group, six patients had TACE before radiotherapy, three patients after radiotherapy, and eight patients both before and after radiotherapy. Eleven patients had TACE after radiotherapy due to evidence of intrahepatic disease progression during follow-up. In the TACE+RT group, the number of courses of TACE ranged from one to three (median two courses) before radiotherapy, and from one to three (median two courses) after radiotherapy. In the TACE group, the number of courses of TACE ranged from one to four (median two courses). The mixture of iodized oil contrast medium (Lipiodol) and chemotherapeutic agents consisting of doxorubicin, cisplatin and mitomycin C were used for infusion simultaneously with embolization during TACE. Ivalon particles (Polyvinyl alcohol foam; Unipoint Industries, High Point, NC) or gelatin sponge particles were used as the embolization material. An interval of at least 30 days between TACE and radiation treatment was necessary irrespective of which treatment was given first.
Patients were monitored weekly with complete blood counts and liver function tests during TACE or radiation treatment. CT scans were obtained before TACE/radiotherapy, 3 to 6 weeks after completion of TACE/radiotherapy, and then every 3 to 6 months, if indicated. Local progression was defined as progression of the primary tumor with size increase of more than 25%. Regional progression was defined as CT evidence of new development or progression of the hepatic tumor(s) elsewhere within the liver, not extending from the target volume of radiation. Extrahepatic progression was defined as any evidence of disease recurrence outside the liver. Survival was estimated from the date of initiation of the first treatment, whether radiotherapy or TACE. The Kaplan-Meier method and Wilcoxon test were used in the analyses of survival outcome and prognostic factors.
After a median follow-up of 24 months, 20 patients remained alive and 22 dead. The median survival for all patients was 19 months, and the two-year survival rate was 49%. Fourteen patients were progression-free at the time of the last follow-up, four in TACE+RT group, five in RT group, and five in TACE group. Only two patients in the TACE+RT group and one patient in the RT group compared with five in the TACE group had local progression of the hepatic tumor treated with radiotherapy. Ten patients in the TACE+RT group, three patients in RT group, and nine patients in the TACE group had evidence of regional progression elsewhere within the liver. Ten patients in the TACE+RT group and three patients in RT group developed extrahepatic progression, but no patient in the TACE group developed distant metastasis (Table 3).
|Pattern of failure|
|Local progression only||0||0||2|
|Regional progression only||3||0||6|
|Extrahepatic progression only||1||1||0|
|Local and regional progression||0||1||3|
|Local and extrahepatic progression||2||0||0|
|Regional and extrahepatic progression||7||2||0|
The survival rates at two years were 58%, 11%, and 56% for patients in the TACE+RT group, RT group, and the TACE group, respectively (Fig. 1). All patients in the RT group died of their malignant disease. The median survival was 4.8 months in the RT group, whereas it had not been reached in both the TACE+RT and the TACE groups. The survival between the TACE+RT and the TACE groups was comparable (P = 0.69), but both were significantly better than the RT group (TACE+RT vs. RT, P = 0.0007; TACE vs. the RT, P = 0.01). No statistically significant difference was identified in local and regional control between the TACE+RT and the TACE groups. The two-year local progression-free survival rates for the treated hepatic tumor were 83% and 71% for the TACE+RT and the TACE groups, respectively (P = 0.11, Fig. 2). The two-year regional progression-free survival rates were 43% and 26% (P = 0.48, Fig. 3). The two-year extrahepatic progression-free survival rates were 36% and 100% (Fig. 4). The difference in extrahepatic progression-free survival between TACE+RT and the TACE groups was significant (P = 0.002). The two-year rates of local, regional, and extrahepatic progression-free survival in RT group were 50%, 52%, and 67%, respectively.
Treatment-related toxicity could be assessed for all patients. In the TACE+RT and the RT groups, six patients developed radiation-induced liver disease (RILD), manifested by the development of an anicteric elevation of alkaline phosphatase level of at least twofold and nonmalignant ascites in the absence of documented progressive disease, or elevated transaminases of at lease fivefold the upper limit of normal or of pretreatment level. All RILDs occurred within three months after completion of radiation treatment. Four of them died of hepatic failure and the other two recovered. Four patients had radiation-related gastrointestinal bleeding. Three patients had CT evidence of radiation pneumonitis, but only one of them was symptomatic. Most patients had transient elevation of transaminases during radiation treatment and TACE. In the TACE group, one patient died of TACE-related hepatic failure within two weeks after the treatment. All other TACE-related side effects, including nausea, fever, transient elevation of serum transaminases, and pain at upper abdomen, usually occurred within a week from the treatment. All these side effects were tolerable and reversible.
In the analysis of prognostic factors, age, gender, performance status, and Child-Pugh class did not influence survival statistically. No factor had a statistically significant impact on local control in all three groups.
In the TACE+RT group, the two-year survival rates for patients with T4 and non-T4 disease were 21% and 80%, respectively (P = 0.01). T4 disease was associated with significantly worse regional progression-free survival (P = 0.007). A significant difference was observed for TACE+RT group patients in both regional progression-free survival (P = 0.002) and extrahepatic progression-free survival (P = 0.02) for patients with portal vein thrombosis before treatment (Table 4).
|Prognostic factor||TACE+RT (%)||P value||RT (%)||P-value||TACE (%)||P-value|
|Two-year overall survival|
|Tumor size > 10cm||60||0.80||14||0.80||33||0.05|
|Tumor size ≤ 10cm||58||0||67|
|AFP > 200 ng/ml||50||0.30||25||0.08||20||0.01|
|AFP ≤ 200 ng/ml||67||0||73|
|Two-year regional progression-free survival|
|Portal vein thrombosis (+)||0||0.002||0||0.008||50||0.63|
|Portal vein thrombosis (−)||49||100||32|
|Two-year extrahepatic progression-free survival|
|Portal vein thrombosis (+)||0||0.02||33||0.07|
|Portal vein thrombosis (−)||39||83|
In the RT group, patients with portal vein thrombosis had significantly worse regional progression-free survival (P = 0.008) and extrahepatic progression-free survival (P = 0.07) (Table 4).
In the TACE group, the two-year survival rates were 33% and 70% for patients with T4 and non-T4 disease, respectively (P = 0.04). Tumor size was also a prognostic factor in survival. Patients with hepatic tumor of more than 10cm in size had a significantly shorter two-year survival compared with those with smaller tumor (33% vs. 67%, P = 0.05). Patients with pretreatment AFP of more than 200 ng/ml had a significantly negative impact on survival (20% vs. 73%, P = 0.01). No known factors that influence the regional control within the liver was identified by statistical testing (Table 4).
Hepatocellular carcinoma is one of the most common malignancies in Asia. Although surgical resection has been considered the treatment of choice for long-term control of the disease, most patients are not surgical candidates at the time of diagnosis, either due to extensiveness of the disease or inadequate hepatic reserve to permit hepatectomy to be a viable treatment option. Even for those who are candidates for resection, approximately 30% to 70% will eventually develop hepatic metachronous tumors or recurrence [17, 18]. Nonsurgical modalities, including TACE and ethanol or acetic acid injection, can be used selectively and repeatedly, if necessary, for patients with HCC . However, no survival benefits have been observed in at least two randomized trials of TACE [20, 21]. Hence, they are considered reasonable palliative treatment options for patients with unresectable HCC [22, 23]. Following several TACE applications, arterial occlusion may occur and give rise to collateral arterial supplies, which make repeated treatments difficult. On the other hand, local recurrence is frequent for patients with large hepatic tumors, treated with TACE or intralesional injections of ethanol, either due to complex blood supply or due to poor distribution of chemotherapeutic drugs or ethanol within the tumors .
Local radiotherapy can be an effective adjunct to the palliative treatment of unresectable HCC. The 26 patient results with median survival of 19 months and two-year survival rate of 42% confirm reports from other institutions [9–12]. Several combined modalities of treatment have been designed for patients with unresectable HCC. Matsuura et al. reported the experience in treating 22 patients with radiotherapy alone or with TACE and ethanol injection . Seong et al. treated 30 patients with local radiotherapy starting within 7 to 10 days following TACE . Robertson and Lawrence et al. have a series of reports indicating their experience in using conformal, twice a day radiotherapy given concurrently with hepatic artery fluorodeoxyuridine infusion for patients with hepatobiliary cancers . In contrast to the low hepatic tolerance and questionable benefit from whole liver irradiation reported earlier, these series demonstrated promises of conformal radiation therapy to a limited portion of the liver as a treatment with more limited toxicity.
Both the TACE+RT and the TACE groups in our study have comparable survival. Of note is that patients in the radiotherapy group had significantly larger hepatic tumors and relatively higher AFP level. These two factors have also been reported to be poor prognostic factors in several series [25–27]. The addition of radiotherapy to TACE/ethanol injection may offer patients with larger tumors a potential benefit, as seen in their exhibiting a similar survival to those patients with smaller hepatic tumors treated with TACE alone. The survival of patients treated with radiotherapy alone was significantly shorter than those in the other two groups. It is likely that the difference was from the selection bias. Patients in the RT group were usually not candidates for TACE. They had the poorest performance status, the highest AFP levels, the largest tumors, and the most advanced disease. These characteristics confer poorer outcome. The short survival may be associated with the underestimation of regional and distant metastasis. For this reason, we excluded this group from the comparison in patterns of failure. Ohto et al. also demonstrated a survival advantage in their patients with Child A/B grades and without portal vein thrombosis treated with local irradiation plus TACE compared with those with radiation alone . TACE can be used in multiple hepatic tumors within the liver, but is less effective in controlling large tumors due to their complicated blood supply. The blood supply in a small tumor is usually only the hepatic artery, while in a large tumor, the incidence of extrahepatic artery supply increases. Radiation therapy can be complementary to TACE for such large tumors. Yasuda et al. demonstrated an excellent three-year survival rate of 81% for their patients with HCC of 3 to 8 cm in size, treated with combined TACE, ethanol injection, and radiotherapy . Based on our data, 10 patients who had less than T4 disease and were treated with combined TACE and local radiotherapy had the longest survival time. The two-year survival was as high as 80%.
The patterns of failure for our patients are mainly intrahepatic multiple recurrences and extrahepatic metastasis. Intrahepatic metastasis or multicentric development of the disease has been the most important factor of failure in Asian patients with HCC treated by surgery  or TACE . Multiple intrahepatic recurrence can be due to intrahepatic extension or metachronous multicentricity of HCC [30, 31]. Only a small proportion of patients in both groups had local progression of the treated hepatic tumor(s), although those treated with radiotherapy had significantly larger tumors. With TACE alone, local recurrence is likely to be the first event of disease progression for patients with large hepatic tumors. It is our hope that local control may be further improved by radiotherapy when integrated into the treatment of patients with large tumors. Hence, the pattern of failure may be shifted from in-field progression to the intrahepatic regional recurrence or extrahepatic metastasis.
It is noteworthy that patients treated with radiotherapy developed extrahepatic metastasis more frequently than those with TACE alone. Thirteen patients in the TACE+RT and the RT groups had systemic metastasis during follow-up compared with none in the TACE group. Even with the exclusion of subgroups of patients with poor prognostic factors, such as large tumor, T4 disease, or portal vein thrombosis, the difference is still statistically significant. Due to the small sample size in each group, it is difficult to stratify patients by other characteristics. The study from Korea also showed a significant proportion of their patients developed distant metastasis after the combined treatment of TACE and radiotherapy. In the Korean series, the mean tumor size was 9 cm and 11 of 30 patients had evidence of portal vein thrombosis before treatment . Similarly, large tumor size, advanced disease status, presence of portal vein thrombosis, incomplete tumor necrosis, and arterioportal shunt, were reported to be associated with distant metastasis . This altered pattern of failure may be real or from a selection bias. Patients who have larger tumors and more advanced disease were often given the combined modalities of TACE and radiotherapy. These patients are more likely to have preexisting micrometastasis before treatment. With the improved disease control within the liver and lengthened survival, distant metastasis becomes more visible in the follow-up period. Recent advances in molecular biology also confirmed the correlation of the presence of circulating HCC cells in peripheral venous blood with advanced disease status and subsequent distant metastasis [33, 34]. Interestingly, it was also reported that the portal venous system might play the role of an efferent vessel in metastasis . One wonders whether the tumor thrombi in the portal vein facilitate the intrahepatic or extrahepatic spread of HCC through the portal venous network due to the presence of arterio-portal shunting in the tumor.
It deserves further attention that a significant proportion of patients in the TACE+RT and the RT groups developed grade 3 or 4 toxicities from the treatments. Six patients had RILD and four of them died of this complication. In contrast, only one patient died of treatment-related hepatic failure in the TACE group. A high proportion of patients with HCC in Taiwan are associated with chronic hepatitis and subclinical or overt cirrhosis of liver. These patients may be more susceptible to hepatic irradiation due to the preexisting subclinical compromise of hepatic reserve. With the advanced treatment planning system, the dose-volume data can be analyzed and compared between patients with and without RILD. The traditional dose-volume criteria for hepatic tolerance to irradiation, such as volume fraction receiving less than 30 Gy, volume fraction with < 50% of isocenter dose , the prediction score from surgical series , or mean hepatic dose , failed to predict RILD for our patients (data not shown). The currently accepted criteria may not be appropriately applied in our patients with high incidence of chronic viral hepatitis. It is recommended that effective volume method and normal tissue complication probability (NTCP) model be used in the protocols of dose prescription and escalation [14, 38]. We similarly found that the NTCP model better predicted the occurrence of RILD in our series (data not shown). On the other hand, intensity-modulated radiation therapy (IMRT) has been increasingly reported to offer adequate target coverage and dose escalation and to spare the critical structures [39, 40]. With the evolving NTCP model and the new treatment technique, further efforts are being made to investigate the optimal use of IMRT in patients with intrahepatic malignancies.
Local radiotherapy is effective in adjunct with TACE or ethanol injection for patients with unresectable HCC. Its effect is more prominent within the site to which radiation is given. Intrahepatic and extrahepatic spread of the disease are the causes of treatment failure. There is no difference in survival and local control between patients treated with combined TACE and radiotherapy and patients with TACE alone, although those who received radiotherapy have larger tumors. Patients treated with radiotherapy alone survived shortest because of their poor condition and more advanced disease. More patients with radiotherapy as part or all of their treatment developed extrahepatic metastasis after treatment. Thus, the pattern of failure following radiotherapy and TACE shifted to regional recurrence and distant metastasis but seemed to improve survival in patients with large hepatic tumors. The presence of distant metastasis is of greater clinical significance in that therapeutic concern must shift to the new lesions after successful treatment of the primary tumor.
- 1Hepatocellular carcinoma: one of the world's most common malignancies. Am J Med 1985;233:705–708., .