SEARCH

SEARCH BY CITATION

Keywords:

  • carcinoma;
  • hepatocellular;
  • portal vein;
  • drug therapy;
  • interferons;
  • fluorouracil

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

BACKGROUND

Hepatocellular carcinoma (HCC) with portal venous invasion (PVI) has a very poor prognosis, with a median survival of 3 months and virtually no survival at 1 year. The combination of intraarterial 5-fluorouracil (FU) and systemic interferon-α (IFNα) was recently reported to be effective against HCC with PVI, but these were small pilot studies.

METHODS

One hundred and sixteen patients with HCC with PVI received IFNα (5,000,000 U intramuscularly on Days 1, 3, and 5 of each week of treatment) and 5-FU (500 mg into hepatic artery on Days 1-5 of the first and second week of each 4-week cycle). The therapy was either terminated at the end of the first cycle in cases with progressive disease, or continued for at least 3 cycles, when responses to treatment were evaluated by Eastern Cooperative Oncology Group criteria. The survival rate was compared with that of historical controls (n = 40).

RESULTS

Nineteen (16%) patients showed complete response and another 42 (36%) showed partial response. Adverse events were limited to nausea and appetite loss. The survival rates at 12 and 24 months among overall patients were 34% and 18%, respectively, in contrast to 15% and 5% among the historical controls. Survival rates at 12 and 24 months were 81% and 59% among complete responders, respectively, and 43% and 18% among partial responders.

CONCLUSION

The combination therapy with 5-FU and IFN was safe, and substantially improved the survival rate among the complete responders. These results provide a rationale for future randomized controlled trials. Cancer 2006. © 2006 American Cancer Society.

Hepatocellular carcinoma (HCC) is one of the most common cancers worldwide, causing approximately 250,000 annual deaths.1, 2 The incidence of HCC has been increasing in Japan in the last 30 years3 and also in the US more recently.4 Advances in imaging techniques have facilitated the detection of HCC at early stages,5–9 and those in therapeutic modalities, such as hepatic resection, percutaneous ethanol injection, radiofrequency ablation, and transplantation, have substantially improved the prognosis of HCC patients.10–14 However, the long-term prognosis of HCC remains far from satisfactory, mainly because of the frequent recurrence of HCC.15, 16

In particular, portal venous invasion (PVI), which is reported to develop in 16% to 65% of HCC cases,17–20 is a common and serious sequela. First, tumor cells may spread out through the portal tract, resulting in extensive intrahepatic metastases. Second, portal vein occlusion may further deteriorate liver function, causing liver failure. Third, portal hypertension may be aggravated and lead to intractable ascites or variceal rupture.21–26 These sequelae are often lethal by themselves, and will contraindicate further treatment of HCC, including liver transplantation.

Surgical resection of PVI has been reported but the results were disappointing.27–29 Transcatheter chemoembolization to necrotize PVI have also been reported to necrotize PVI, but the results were unsatisfactory.30, 31 Arterial or systematic infusion of chemotherapeutic agents are rarely effective.32–35 Radiation is sometimes effective against PVI, but the indication is often limited by the extent of the lesion or impaired liver function.36

The combination of interferon-α (IFNα) and 5-fluorouracil (5-FU) was first reported by Wadler et al.37 in 1989 using advanced colorectal cancer. In 1993, Patt et al.38 reported the effectiveness of combination of intramuscular IFNα and intravenous 5-FU in HCC patients. Subsequently, Urabe et al.39 reported the combination of IFNα, 5-FU, cisplatin, methotrexate, and folinic acid in 1998. In 2002, Sakon et al.40 reported the efficacy and safety of the combination of systemic IFNα and 5-FU administered to the hepatic artery in 8 HCC patients with PVI. Encouraged by these preliminary reports, we started to treat HCC patients with PVI with the combination therapy. In this study, we report the results of 116 consecutive HCC patients with PVI treated with the combination of IFNα and 5-FU, and compared the outcomes of 40 patients who were previously treated otherwise.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patients

HCC and PVI were diagnosed with contrast-enhanced computed tomography (CT) or magnetic resonance imaging. The staging of PVI was also aided by Doppler ultrasonography. Ultrasound-guided tumor biopsy was also performed when differentiation from nontumorous portal thrombosis was required. Eligibility criteria were as follows: HCC with PVI to at least 1 of the main branches of the portal vein; no indication for radiation therapy; an Eastern Cooperative Oncology Group (ECOG) performance status41 of 0-2 (bed rest less than 50% of the daytime or better); no uncontrollable ascites; leukocyte count > 3000/μL; platelet count > 50,000/μL; total bilirubin < 3.0 mg/dL; serum creatinine < 1.5 mg/dL, and no contraindication for the implantation of the intraarterial catheter and the drug delivery system. During the study, we amended the criteria to exclude patients with extrahepatic metastases. We chose 40 other HCC patients with PVI we had treated previously with other modalities than the combination therapy as historical controls. They received radiation therapy (n = 8), intraarterial chemotherapy with epirubicin hydrochloride (n = 6), or supportive care only (n = 26) between August 1994 and August 2000.

Response Assessment

Before treatment, patients were evaluated by physical examination including the evaluation of performance status, laboratory tests including tumor markers, and CT, which was performed within 4 weeks before the commencement of therapy.

CT was repeated at the end of each therapeutic cycle to assess the response to treatment. The response was classified according to the ECOG criteria as: complete response when all measurable lesions disappeared including signs, symptoms, and biochemical changes related to the tumor, which must have lasted for more than 4 weeks, and no appearance of new lesions; partial response when the sum of the products of the greatest perpendicular dimension of each lesion was reduced by more than 50% and no appearance of new lesions; stable disease when the reduction was smaller than 50% or there was a smaller than 25% increase and no appearance of new lesions; and progressive disease when the increase was greater than 25% or appearance of new lesions.

Implantation of Arterial Catheter

An indwelling intraarterial catheter (Anthron P-U Catheter, TORAY, Tokyo, Japan) was inserted by direct femoral arterial puncture and its tip was put in the proper or common hepatic artery, embolizing the right gastric and the gastroduodenal arteries to avoid efflux of chemotherapeutic agents into the stomach and duodenum. The other end of the catheter was connected to a drug delivery system (P-U Celsite Port, TORAY) subcutaneously implanted in the lower abdomen.

Treatment Protocol

Patients who provided fully informed written consent were treated according to the protocol below. One cycle of treatment consisted of 4 weeks, where 5,000,000 U (5 MU) IFNα (OIF; Otsuka Pharmaceutical, Tokyo, Japan) was administered intramuscularly on Days 1, 3, and 5 of each week, resulting in a total dose of 60 MU in a cycle. 5-FU (500 mg/day, Kyowa Hakko, Tokyo, Japan) was administered into the hepatic artery over 5 hours using a portable infusion pump on Days 1-5 of the first and second weeks through the intraarterial catheter (5 g in a cycle).

At the end of each cycle the response to the therapy was assessed as described above. The combination therapy was discontinued in patients with progressive disease; otherwise, the treatment was repeated for at least 2 more cycles. The therapy was terminated in the event of a level 3 adverse effect according to the ECOG classification. Exceptions were changes in platelet count, leukocyte count, and total bilirubin that were attributable to cirrhosis and fever that was attributable to IFN.

Statistical Analyses

The difference of mean was analyzed with Student t-test. The difference in frequency distribution was analyzed with Fisher exact test or the chi-square test. Predisposing factors for the complete response to the combination therapy, as evaluated by the above criteria, were analyzed with logistic regression, where each continuous variable was transformed into a binary variable divided by the median value. In survival time analysis, data collection ended on December 10, 2004. Cumulative survival was calculated using the Kaplan-Meier method and the difference among the groups was analyzed with the log rank test. Independent factors for survival were assessed with the Cox proportional hazard regression model, including the comparison with historical controls. All statistic calculations were performed with SAS v. 8 (Cary, NC).

RESULTS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

Patient Profile

A total of 116 patients, 95 men and 21 women, with an average age of 64 years (range, 39-79) received the combination therapy between September 2000 and May 2004 (Table 1). The etiology of the background liver disease was hepatitis C virus (HCV) in 78 patients, hepatitis B virus (HBV) in 24 (one positive also for HCV), and non-B non-C hepatitis in 15. Eighty-eight patients had PVI at a major branch (Vp 3) and 28 in the main trunk (Vp 4). The mean size of intrahepatic HCC tumor was 8 cm in diameter (range, 2-15). The baseline characteristics of 40 historical controls are also shown in the Table 1. There were no differences in tumor characteristics or liver function between the 2 groups.

Table 1. Baseline Clinical Characteristics
Characteristics*IFN + 5-FUControlP
  • AST: aspartate aminotransferase; ALT: alanine aminotransferase; HBV: hepatitis B virus; HCV: hepatitis C virus.

  • *

    The demographics of the 2 study groups. Control group was not significantly different from IFN + 5-FU group.

  • Student t-test.

  • Fisher exact test.

  • §

    Mean ± SD.

  • Data are expressed as numbers (%).

No. of patients11640 
Age, y§62.7 ± 9.164.5 ± 8.4.29
Gender male/female95/2132/8.79
Extent of portal vein invasion: trunk/first branch28/8815/25.23
Tumor location: unilobular/bilobular28/8815/22.10
Tumor size (cm)§9.3 ± 3.88.4 ± 4.0.09
Child-Pugh status: A/B/C7/89/106/29/5.14
Total bilirubin (mg/dL)§1.2 ± 0.91.4 ± 1.0.35
Albumin (g/dL)§3.4 ± 0.53.3 ± 0.5.24
AST (IU/L)§99.7 ± 63.486.7 ± 29.4.21
ALT (IU/L)§64.5 ± 77.565.6 ± 17.3.93
Prothromin time (%)§81.5 ± 14.179.3 ± 13.1.38
Platelet count (×109/L)§143,000 ± 82,000116000 ± 65000.06
α-fetoprotein (Positivity (≥ 20 ng/mL) %, median ng/mL)83%, 96995%, 546.84
Des-γ-carboxy prothrombin (Positivity (≥ 40 AU/mL)%, median AU/mL)86%, 1585.087%, 1307.0.06
Etiology : HBV/HCV/HBV + HCV/others23/77/1/157/31/0/2.44

All of the patients received at least 1 cycle of the combination therapy (range, 1-7 cycles; average, 2.1 cycles). The combination therapy was discontinued when the response was revealed to be progressive disease. Otherwise, the treatment was repeated, up to 4 or 5 times, so far as general conditions permitted. All of the 116 patients received the combination therapy: 1 cycle: 48 (41%) patients, 2 cycles: 32 (28%) patients, 3 cycles: 20 (17%) patients, 4 cycles or more: 16 (14%) patients.

Response to Treatment

Nineteen (16.4%) patients had a “complete response,” 42 (36.2%) had “partial response,” 2 (1.7%) had “stable disease,” and 53 (45.7%) “progressive disease,” resulting in a response rate (complete and partial responses) of 52.6%. The average duration of complete and partial responses was 13.6 and 4.8 months, respectively. The complete response rate was higher among patients with HCV infection (22%) than among others (5%) (Tables 2, 3). The final response to treatment was predictable, with the early response of tumor biomarker levels at Week 2 among patients whose markers were positive before treatment (α-fetoprotein (AFP) was positive (> 20 ng/mL) in 88%; L3 fraction of AFP (>15%) in 73%; and des-γ-carboxy prothrombin (DCP) (> 40 AU) in 87%, and at least 1 marker was positive in 97%). Patients with decreased tumor marker levels at Week 2 were very likely to achieve a final partial or complete response (90% sensitivity and 80% specificity).

Table 2. Univariate Analysis of Predictors for Complete Response
Variables*Odds Ratio95% CIP
  • AST: aspartate aminotransferase; ALT: alanine aminotransferase; HCV: hepatitis C virus; CI: confidence interval.

  • *

    Univariate analysis of clinical, biochemical, and pathologic risk factors for complete response.

  • The median value of each parameter is indicated in parentheses.

  • Statistical significance.

Age (> 64 y)1.0210.965–1.079.4729
Gender (male)0.7970.235–2.702.7155
Portal vein invasion (main trunk)0.8110.245–2.681.7314
Platelet count (>120,000 ×109/L)0.9430.871–1.022.1541
Total bilirubin (>1.0 mg/dl)0.8920.496–1.602.7014
Albumin (> 3.3 g/dL)0.6620.239–1.837.4283
AST (> 85 IU/L)1.0030.996–1.010.3538
ALT (> 51 IU/L)1.0000.993–1.007.9370
Prothrombin time (> 81%)0.9660.930–1.003.0736
α-fetoprotein (> 970 ng/mL)0.7420.275–2.006.5570
Des-γ-carboxy prothrombin (> 1585 AU/mL)0.8480.316–2.275.7440
HCV antibody (positive)5.2421.145–24.002.0328
Table 3. Multivariate Analysis of Predictors for Complete Response
Variables*Odds Ratio95% CIP
  • HCV: hepatitis C virus; CI: confidence interval.

  • *

    Multivariate logistic regression analysis of complete response. Those factors with P < .2 in the univariate analysis (Table 2) were included.

  • The median value of each parameter is indicated in parentheses.

Platelet count (> 120,000 × 109/L)0.9610.878–1.052.3875
Prothrombin time (> 81%)0.9840.943–1.026.4499
HCV antibody (positive)4.2620.896–20.267.0683

Safety

Fever occurred in about 90% of patients, usually after the first IFN administration and gradually improved during subsequent administrations. An elevated aspartate aminotransferase (AST) level and decreased leukocyte or platelet count were found in 60% to 80% of patients, but none resulted in termination of the therapy or required granulocyte-colony-stimulating factor (G-CSF) administration. Nausea and vomiting, mostly ECOG Grade 1, occurred in about 50% patients. There was 1 case of stomatitis and another of depression (Grade 3 adverse event). The latter patient was the only case of discontinuation of treatment during the first cycle. There were no complications resulting from the arterial catheter.

Survival

Cumulative survival among all patients treated with the combination therapy is shown in Figure 1. The survival rates at 6, 12, and 24 months were 53%, 34%, and 18%, respectively, with a median survival of 6.9 months, compared with a survival rate of 40%, 15%, and 5%, respectively, in the historical control group. The survival was significantly different between the 2 groups (P < .01 by the log-rank test).

thumbnail image

Figure 1. Comparison of the overall survival rate across the IFN + 5FU and control groups. The rate was significantly higher in the IFN + 5FU groups than in the control group (P < .01). Open circle, IFN + 5FU group; open triangle, control group.

Download figure to PowerPoint

As shown in Figure 2, the response to treatment significantly affected survival. The complete responders showed the best survival rate of 81% and 59% at 12 and 24 months, respectively. In contrast, patients with stable disease or progressive disease had poorer outcomes, comparable to the historical control. Patients with a partial response showed an intermediate prognosis, with a survival rate of 43% and 18% at 12 and 24 months, respectively. Univariate (Table 4) and multivariate analyses (Table 5) showed that the predictors for survival were complete response (P<.0001), cycle number (P = .0038), VP grade (P = .0123), and total bilirubin concentration (P = .0339). HCC invasion into the main trunk of the portal vein and total bilirubin concentration greater than 1.0 mg/dL predicted a poor prognosis independent of the treatment response.

thumbnail image

Figure 2. Comparison of the overall survival rate across the complete response, partial response, progress disease, and control groups. The rate was significantly higher in the complete response group than in the other 3 groups. Open circle, complete response group; open triangle, partial response group; open square, progressive disease group; closed circle, control group.

Download figure to PowerPoint

Table 4. Univariate Analysis of Predictors for Survival
Variables*Risk Ratio95% CIP
  • AST: aspartate aminotransferase; ALT: alanine aminotransferase; HCV: hepatitis C virus; CI: confidence interval.

  • *

    Univariate analysis of clinical, biochemical, and pathologic risk factors for survival.

  • The median value of each parameter is indicated in parentheses.

  • Statistical significance.

Age (>64 y)0.9990.977–1.021.9344
Gender (male)0.8980.516–1.565.7054
Portal vein invasion (main trunk)0.5030.302–0.837.0081
Platelet count (>120,000 x109/L)1.0060.981–1.031.6373
Total bilirubin (<1.0 mg/dl)1.4811.210–1.813.0001
Albumin (>3.3 g/dL)0.6940.445–1.083.1078
AST (>85 IU/L)1.0021.000–1.005.0832
ALT (>51 IU/L)1.0031.000–1.005.0329
Prothrombin time (>81%)1.0000.987–1.014.9941
α-fetoprotein (>970 ng/mL)0.9480.626–1.434.8012
Des-γ-carboxy prothrombin (>1585 AU/mL)0.6720.441–1.024.0645
HCV antibody (positive)1.3570.867–2.124.1812
Complete response5.4572.603–11.442<.0001
Table 5. Multivariate Analysis of Predictors for Survival
Variables*Risk ratio95% CIP
  • AST: aspartate aminotransferase; ALT: alanine aminotransferase; HCV: hepatitis C virus; CI: confidence interval.

  • *

    Stepwise multiple Cox regression analysis in the IFN + 5-FU group. Those variables with P<.2 in univariate analysis (Table 4) were included.

  • The median value of each parameter is indicated in parentheses.

  • Statistical significance.

Portal vein invasion (main trunk)0.00431.041–1.938.0043
Total bilirubin (<1.0 mg/dl)1.4201.041–1.938.0267
Albumin (>3.3 g/dL)0.6690.357–1.192.1742
AST (>85 IU/L)1.0020.997–1.006.2692
ALT (>51 IU/L)1.0000.995–1.004.8807
Des-γ-carboxy prothrombin (>1585 AU/mL)1.0001.000–1.000.1166
HCV antibody (positive)1.1910.722–1.963.4939
Complete response7.2933.359–15.837<.0001

Case Presentation

A 63-year-old male underwent a partial hepatectomy for HCC in March 1998. One year after surgery, multiple intrahepatic recurrences were found and transcatheter arterial embolization (TAE) was performed. Despite repeated TAE, recurrences continued to occur. Massive HCC in segment 4 and PVI for the right main portal branch appeared in August 2002. The patient received the combination therapy (3 cycles) between September 2002 and November 2002, which resulted in a decrease in the DCP level from 14,400 to 30 and also in the AFP level from 96,497 to 48. At the end of therapy, a CT scan showed no evidence of HCC with PVI and the portal blood flow was recovered (Fig. 3).

thumbnail image

Figure 3. Abdominal CT scan before the combination therapy. Massive HCC existed in segments 4 and 8 (1), with invasion into the right main branch of portal vein blocking portal blood flow (2). Abdominal computed tomography scan after the combination therapy. Massive HCC disappeared (3) and the portal blood flow reappeared (4).

Download figure to PowerPoint

DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES

In the present study, the combination chemotherapy with IFNα and 5-FU showed an objective response rate of 52% among HCC patients with PVI. In particular, 16% of patients achieved a complete response, namely disappearance of PVI and intrahepatic HCC nodules that lasted at least 4 weeks. The improvement in survival rate was remarkable among those patients, showing a survival rate of 81% and 59% at 1 and 2 years, respectively. These figures were in distinct contrast to the 15% and 5% among the historical controls. In the previous report by Sakon et al.,40 the combination therapy of 5-FU and IFNα demonstrated better responses than ours. The discrepancy may be due to the following. First, the current study included HCC patients with invasion not only into a major branch but also into the main trunk (28%). Second, the majority of patients in this study were recurrent cases (74%). Third, total bilirubin level was greater than 1.0 mg/dL in the majority of patients (67%), indicating poorer liver function and prognosis.

Reports using monotherapy with intraarterial 5-FU showed lower response rates, ranging from 13% to 22%, with a median survival of only 3.5 to 14 months.33–35 Monotherapy with IFNα, once thought to be an omnipotent anticancer drug, hardly demonstrated clinical effects against HCC.42 Thus, the combination of the 2 agents seems to have some synergism, but the mechanism of action is not known yet. It was recently reported that IFNα induces p53, which enhances apoptotic responses to 5-FU.43 We examined the combination of IFNα and 5-FU in 8 distinct human HCC lines and found that IFNα markedly increases susceptibility to 5-FU in 5 of the 8 cells.44 Several genes showed distinct gene expression profiles in the responsive cells and others. Further investigation of these genes may elucidate underlying molecular mechanisms, enabling us to predict the efficacy of this combination therapy.

Despite the prominent improvement in survival among complete responders, we must admit that the complete response rate, 16%, was not satisfactory. Although it may be possible to enhance the response among partial responders by modifying the protocol, we suspect that about half of the patients with HCC will remain unsusceptible to the combination therapy and we cannot predict responders beforehand. Fortunately, adverse events were rarely severe and all were manageable in this study. At present, we recommend starting the combination therapy with close monitoring of response, preferably that of tumor biomarkers, and treatment should be continued if there is a response after the first cycle of chemotherapy.

We found in the early phase of this study that the combination of IFNα and intraarterial 5-FU is not effective against extrahepatic metastases. This is understandable because 5-FU, administered into the hepatic artery, will not reach extrahepatic tissues in high concentrations. However, systemic administration of 5-FU or related agents may be effective against extrahepatic lesions in combination with IFNα. This possibility is highly interesting because the implantation of an indwelling catheter is one of the demerits of the present combination therapy. If the agent can be given in a less invasive way, we may be able to enlarge the indication so as to include less advanced HCC patients.

In conclusion, we show that the combination of systemic IFNα and intraarterial 5-FU is effective in a subset of HCC patients with PVI, which resulted in disappearance of PVI and intrahepatic HCC nodules in 16% of patients. These complete responders showed survival rates of 81% and 59% at 12 and 24 months, respectively.

REFERENCES

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS
  5. DISCUSSION
  6. REFERENCES