Skip to main content
Device Global Header

Subscribe
or Renew

Advanced Search

Original Article

A Comparison of Lipiodol Chemoembolization and Conservative Treatment for Unresectable Hepatocellular Carcinoma

List of authors.
  • Groupe d'Etude et de Traitement du Carcinome Hépatocellulaire*

Abstract

Background

Chemoembolization with Lipiodol (iodized oil) is widely used to treat patients with unresectable hepatocellular carcinoma. Severe side effects have been reported, and improved survival has not been clearly demonstrated.

Methods

Patients with unresectable hepatocellular carcinoma who did not have severe liver disease and who met additional entry criteria were randomly assigned to receive either Lipiodol chemoembolization (70 mg of cisplatin, 10 ml of Lipiodol, and gelatin-sponge [Gelfoam] particles delivered through the hepatic artery) or conservative management involving treatment of complications and pain. Courses of treatment were to be given every two months for a maximum of four courses. The main end point was survival.

Results

The study was stopped in December 1992, after a sequential analysis showed the lack of the expected benefit from chemoembolization. As of October 1, 1994, 39 of the 50 patients assigned to chemoembolization and 40 of the 46 patients assigned to conservative management had died. Twenty-six patients assigned to chemoembolization received all four courses of treatment. There was no significant difference in survival between the two groups, although there was a trend favoring the chemoembolization group (estimated relative risk of death in the control group, 1.4; 95 percent confidence interval, 0.9 to 2.2; P = 0.13). The comparison of survival between the two groups was not substantially changed by adjustments for differences in base-line and prognostic characteristics (adjusted relative risk, 1.3; 95 percent confidence interval, 0.8 to 2.1; P = 0.31). At one year, the estimated survival rates were 62 percent in the chemoembolization group (95 percent confidence interval, 48.6 to 75.4 percent) and 43.5 percent in the conservative-management group (95 percent confidence interval, 29.2 to 57.8 percent). In the chemoembolization group, tumor growth, as assessed by tumor size and serum alpha-fetoprotein concentration, was reduced and the incidence of portal obstruction was lower than in the conservative-management group. Liver failure occurred after 47 courses of treatment in 30 patients assigned to chemoembolization.

Conclusions

In a group of patients with unresectable hepatocellular carcinoma but without severe liver disease, Lipiodol chemoembolization reduced tumor growth, often caused acute liver failure, and did not significantly improve survival.

Introduction

The efficacy of therapies for hepatocellular carcinoma is poor.1,2 Hepatocellular carcinoma is a hypervascular tumor.3 Hepatic intraarterial injection of antitumor agents, performed to increase the local concentration of drugs and reduce systemic side effects,4 and intraarterial embolization, which causes ischemic necrosis of the tumor,5 have been used as palliative treatments, either alone or in combination (chemoembolization). Mixing anticancer drugs with Lipiodol, an iodized oily agent that remains selectively in tumors for long periods,6 may enhance the antitumor effect.7,8 Treatment combining intraarterial chemotherapy with Lipiodol and embolization, or Lipiodol chemoembolization, is considered the most effective of these methods on the basis of reports of a decrease in tumor size in nonrandomized trials.9-13 Although this method is widely used, severe side effects have been reported9,14,15 and improved survival, as suggested by case–control studies, has not been clearly demonstrated.

To test the hypothesis that Lipiodol chemoembolization would result in a 75 percent rate of survival after eight months, as compared with the 50 percent rate associated with conservative management, we compared four courses of Lipiodol chemoembolization with conservative treatment in patients who had unresectable hepatocellular carcinoma but not severe liver disease.

Methods

Consecutive patients who met the entry criteria and who agreed to participate were included in this multicenter, sequential, randomized, open-label trial. The diagnosis of hepatocellular carcinoma was based on histologic or cytologic findings or on findings of cirrhosis, liver tumor, and a serum alpha-fetoprotein value exceeding 250 ng per milliliter. Patients were excluded if they had an indication for surgery, had been treated previously for hepatocellular carcinoma, had severe hepatic disease (defined as the presence of one of the following: encephalopathy, gastrointestinal hemorrhage in the past month, clinical ascites, a serum bilirubin concentration of at least 2.9 mg per deciliter [50 μmol per liter], or a serum albumin concentration below 30 g per liter), had vascular contraindications to Lipiodol chemoembolization (such as portal obstruction of at least three segmental branches), had contraindications to treatment with cisplatin, had a serum creatinine concentration of 120 μmol per liter (1.36 mg per deciliter) or higher, or had extrahepatic metastasis.

Randomization

After providing written informed consent, prospective participants underwent arteriography. If no contraindications were identified, the patients were enrolled and underwent randomization. Randomization was stratified according to center and was carried out by telephone. The patients were assigned to one of two regimens: treatment, with one course of Lipiodol chemoembolization every two months for a total of four courses, with the first course given immediately after randomization; or conservative management, with standardized pain medications (acetaminophen or morphine, given in doses appropriate for the level of pain) and treatment of complications.

Chemoembolization

All participating radiologists had prior experience with Lipiodol chemoembolization. First, the superior mesenteric artery was injected to assess portal flow. When portal flow was adequate, the hepatic artery was catheterized. The tip of the catheter was placed distal to the gastroduodenal artery or in either the left or right branches of the hepatic artery. The emulsion consisted of 70 mg of cisplatin and 10 ml of Lipiodol (Lipiodol Ultrafluide, Laboratoire Guerbet, Aulnay-sous-Bois, France) and was prepared just before injection by mixing through a three-way stopcock from one syringe to another. With fluoroscopic guidance, the radiologists injected this mixture into the arteries that fed the tumors. The injection could be slowed or discontinued if retrograde flow occurred. Embolization was subsequently performed with gelatin-sponge (Gelfoam) particles. The goal was to decrease arterial flow without causing total obstruction. Particular care was taken to avoid reflux. After embolization, angiography was performed to determine the extent of vascular occlusion and to assess blood flow in other arterial vessels. Amoxicillin–clavulanic acid (3 g per day) and metronidazole (1.5 g per day) were administered intravenously for 24 hours before the procedure and continued for eight days either intravenously or, when possible, orally.

Follow-up

Serum alpha-fetoprotein was measured and an abdominal computed tomographic scan was obtained at the beginning of the study, every two months during the first year, and thereafter every four months. All computed tomographic scans were reviewed by two radiologists who were unaware of the patient's clinical data or treatment assignment. The average diameter of the largest nodule recorded during computed tomographic scanning before the study began was noted, and changes in the size of the tumor on subsequent scans were expressed as a percentage of the initial size of the nodule. Changes in the serum alpha-fetoprotein concentration were expressed as a percentage of the base-line levels. The extent of initial segmental portal obstruction was also assessed on hepatic arteriography.

Statistical Analysis

The main end point was survival. The study was conducted as a sequential trial to minimize the number of patients needed and obtain more rapid results. A triangular test design was used.16 The accumulated data were examined after approximately every 10 deaths. At each examination, the z and V statistics were calculated. A positive z value indicated that chemoembolization was superior to conservative management, and a negative value that chemoembolization was inferior. The V statistic approximately represents the sample size. Once the sequential values of z and V were calculated, they were plotted against one another and the plotted point was compared with the stopping boundary for the trial. The study was designed to have a type I error of 5 percent with a power of 90 percent to detect an increased survival benefit for chemoembolization after eight months as compared with conservative management. Since the survival rate after conservative management was assumed to be 50 percent, to be deemed a superior treatment, chemoembolization would have to yield a rate of 75 percent. With this approach, if the plotted point lies above the upper boundary of the triangle or below the lower boundary, the trial is stopped (upper boundary: chemoembolization 50 percent more effective than conservative management; lower boundary: less than a 50 percent difference in survival between the two treatments).

Comparisons between groups were on an intention-to-treat basis. Comparisons were made with the Wilcoxon test for continuous variables, Fisher's exact test for binary variables, and the chi-square test for categorical variables. Survival curves were estimated according to the Kaplan–Meier method17 and compared with use of the log-rank test.18 To determine factors predictive of death and to adjust the treatment comparison for base-line differences between the groups and for prognostic factors, the Cox model was used.19 The selection was based on the likelihood-ratio test. All variables having a P value in the univariate analysis of 5 percent or less were included in the model. SAS (SAS Institute, Cary, N.C.) and BMDP (BMDP Statistical Software, Los Angeles) software packages were used.

Informed Consent

The study was conducted in accordance with the ethical standards set forth in the Declaration of Helsinki. Written informed consent was obtained from each patient. As required by French regulations, the protocol was approved by one ethics committee in France (Hôpital Avicenne, Bobigny, France), as well as by ethics committees in Brussels, Belgium, and Montreal.

Results

Sequential Analysis and Decision to Stop the Trial

Figure 1. Figure 1. Design of the Triangular Test and the Results of the Fifth Sequential Analysis.

The five interim data sets resulted in five pairs of values for the z statistic, which approximately represents the effect of lipiodol chemoembolization on overall survival as compared with that of conservative management, and the V statistic, which approximately represents the sample size. The corresponding points were plotted on the graph as they became available and compared with the stopping boundary set out by the triangular test (see the text). At the fifth analysis, the lower triangular boundary was crossed, causing the trial to be stopped.

The trial was begun on July 15, 1990. Twenty-four centers in three countries participated. Sequential analyses were performed after approximately every 10 deaths. Figure 1 shows the results of the fifth sequential analysis conducted on November 1, 1992. At that time, 92 patients were enrolled: 47 in the chemoembolization group and 45 in the conservative-management group. Forty-seven deaths had been reported: 20 in the chemoembolization group and 27 in the conservative-management group. Since the lower triangular boundary had been crossed (Figure 1), the trial was stopped, and we concluded that there was less than a 50 percent increase in survival after eight months with chemoembolization as compared with conservative management (P<0.05).

When entry was closed on December 1, 1992, 96 patients had been enrolled. During this period, 778 patients presented with hepatocellular carcinoma at the 24 centers, but 682 (88 percent) were excluded from the trial. The main reasons for exclusion were liver failure (205 patients, 30 percent), vascular contraindications (149 patients, 22 percent), indications for surgery (115 patients, 17 percent) or percutaneous alcohol injection (11 patients, 2 percent), extrahepatic metastasis (39 patients, 6 percent), and refusal by the patient to participate (93 patients, 14 percent). Seventy patients were excluded for various other reasons. Results are presented for the 96 patients who underwent randomization — 50 assigned to chemoembolization and 46 assigned to conservative management — and include follow-up through October 1, 1994. No patient was lost to follow-up, and 79 (82 percent) died.

Base-Line Characteristics

Table 1. Table 1. Base-Line Characteristics of the 96 Patients Studied, According to Treatment Group.

The diagnosis of hepatocellular carcinoma was based on histologic or cytologic findings in 79 patients and on findings of cirrhosis, liver tumor, and a serum alpha-fetoprotein concentration in excess of 250 ng per milliliter in 17 patients. The two groups were well balanced with respect to the main characteristics (Table 1). Cirrhosis was present in 87 patients (91 percent), all of whom had well-compensated liver function (Child–Pugh class A). The classification of Okuda et al.20 was used to assess liver function and tumor volume; 90 percent of the patients had stage I disease (47 in the chemoembolization group and 39 in the conservative-management group), and 10 percent had stage II disease (3 in the chemoembolization group and 7 in the conservative-management group). More patients in the conservative-management group had multinodular and large tumors and segmental portal obstruction.

Survival

Figure 2. Figure 2. Kaplan–Meier Estimates of Survival in 50 Patients Assigned to Lipiodol Chemoembolization and 46 Patients Assigned to Conservative Management (P = 0.13 by the Log-Rank Test).

As of October 1, 1994, 39 patients assigned to chemoembolization had died and 40 patients assigned to conservative management had died. The causes of death — mainly, liver failure, gastrointestinal hemorrhage, and spontaneous bacterial peritonitis — did not differ between the two groups. There was no significant improvement in survival in the chemoembolization group as compared with the conservative-management group (relative risk of death in the latter group, 1.4; 95 percent confidence interval, 0.9 to 2.2; P = 0.13 by the log-rank test) (Figure 2). At eight months, the estimated survival rate was 70 percent (95 percent confidence interval, 57.3 to 82.7 percent) in the chemoembolization group and 50 percent (95 percent confidence interval, 35.6 to 64.4 percent) in the conservative-management group; at one year, the estimated survival rates were 62 percent (95 percent confidence interval, 48.6 to 75.4 percent) and 43.5 percent (95 percent confidence interval, 29.2 to 57.8 percent), respectively; and at two years they were 37.8 percent (95 percent confidence interval, 24.3 to 51.3 percent) and 26 percent (95 percent confidence interval, 13.3 to 38.7 percent), respectively. The difference in survival between the groups was also adjusted for the following base-line variables, which were either unbalanced or prognostic as assessed by the log-rank test: Karnofsky score (P = 0.004), ascites demonstrated by ultrasonography (P<0.001), serum bilirubin concentration (P<0.001), serum albumin concentration (P = 0.004), tumor type (P = 0.02), tumor mass (P<0.001), segmental portal obstruction (P<0.001), and serum alpha-fetoprotein concentration (P = 0.009). With the use of the Cox model, the adjusted estimated risk of death remained unchanged, with a relative risk of 1.3 (95 percent confidence interval, 0.8 to 2.1) in the conservative-management group as compared with the chemoembolization group (P = 0.31 by the likelihood-ratio test).

Tumor Growth

Table 2. Table 2. Maximal Changes in Tumor Size and Serum Alpha-fetoprotein Concentrations and Occurrence of Portal Obstruction in the 81 Patients Who Survived More than Two Months after Randomization.

Fifteen patients died within two months of randomization (seven in the chemoembolization group and eight in the conservative-management group). In the other 81 patients, tumor growth, as assessed by tumor size and the serum alpha-fetoprotein concentration, was reduced in the chemoembolization group as compared with the conservative-management group (Table 2). The incidence of portal obstruction was lower in the chemoembolization group (one segmental and two truncal obstructions occurring more than eight months after enrollment in three patients) than in the conservative-management group (seven segmental and three truncal obstructions occurring less than eight months after enrollment in four patients and after more than eight months in six patients). There was no difference between groups in the occurrence of extrahepatic metastasis.

Outcome and Complications of Chemoembolization

Among the 50 patients assigned to chemoembolization, 49 received at least one course of chemoembolization and 26 received all four of the planned courses. The patient who was not treated had a hepatic-artery dissection during arteriography before treatment was begun. The reasons for stopping treatment early in 23 patients were death (13 patients), arterial obstruction (5), proximal arterioportal fistula (1), extrahepatic metastasis (1), and refusal by the patient to continue (3).

Table 3. Table 3. Complications of 148 Courses of Lipiodol Chemoembolization in the 50 Patients Assigned to Such Treatment.

In the chemoembolization group, a temperature of 38°C or above, abdominal pain, or vomiting was noted in 86 percent of the courses in 47 patients (Table 3). These symptoms lasted a mean (±SD) of 2.5±2.1 days (range, 1 to 10). Liver failure, defined as the presence of encephalopathy, ascites, or an increase in the serum bilirubin concentration of 0.9 mg per deciliter (15 μmol per liter) or more, occurred after 47 courses of treatment in 30 patients. Gastrointestinal hemorrhage was caused by esophagitis (two patients), duodenal ulcer (one patient), and antral ulcerations (one patient). One patient with diffuse hepatocellular carcinoma and cirrhosis graded as Child–Pugh class A died of liver failure 10 days after the second course of treatment. Death was considered a complication of treatment because this patient's liver function was judged to be well compensated before treatment and acute liver failure occurred immediately after treatment. In the six other patients who were treated and who died within two months after randomization, death was not as clearly related to the treatment.

Other complications of chemoembolization included dysuria, hiccups, and inguinal hematoma in two patients each and worsened diabetes, a decrease in the platelet count to below 70,000 per cubic millimeter, and a decrease in the serum sodium concentration to below 125 mmol per liter in one patient each. The average period of hospitalization was 6.6 days (range, 2 to 24) for each course of treatment. The mean (±SD) number of days of hospitalization per patient during the first eight months of the study (or until death) was significantly higher in the chemoembolization group than in the conservative-management group (29±16 days vs. 13±14 days, P<0.001 by the Wilcoxon test).

Deviations from the Protocol

Two patients assigned to conservative management each received three courses of chemoembolization. One patient requested this treatment after randomization, and the other was inadvertently treated because of a mistake in transmitting the treatment assignment. These patients were analyzed as part of the conservative-management group and survived 7 and 29 months, respectively. Analysis of the data after the exclusion of these two patients did not change the survival results. One patient assigned to chemoembolization received only one course of treatment because of arterial obstruction and thereafter received percutaneous alcohol injection.

Discussion

Using a sequential design, we did not demonstrate a survival benefit for repeated Lipiodol chemoembolization in patients with unresectable hepatocellular carcinoma but without severe liver disease. Nonetheless, there was a trend toward increased survival, with an estimated relative risk of death of 1.4 in the conservative-management group (95 percent confidence interval, 0.9 to 2.2) as compared with the chemoembolization group. Since imbalances in base-line and prognostic characteristics between the groups could have favored the chemoembolization group, adjustments were made with a Cox model. After this analysis, the relative risk of death in the conservative-management group as compared with the chemoembolization group was estimated to be 1.3 (95 percent confidence interval, 0.8 to 2.1).

Lipiodol chemoembolization has numerous side effects, requires hospitalization, and is expensive.9,14,15 Since it did not markedly improve survival, we do not believe its use can be recommended. Death was the main reason treatment was discontinued before four courses had been completed, emphasizing the frequency of treatment failure.

Consensus is lacking about the most suitable method of chemoembolization. The use of Lipiodol and Gelfoam particles is the most generally accepted method. As compared with Gelfoam particles, Gelfoam powder may cause more harm to areas of the liver without tumor, especially in patients with cirrhosis.21,22 The use of chemotherapy is also controversial. Cisplatin has been suggested to be more effective than doxorubicin,23,24 particularly in one randomized trial.8 We used a lower dose of cisplatin (70 mg) than have other studies8,25 in order to reduce the side effects of the medication, especially renal toxicity. Our prior experience suggested that patients with alcoholic cirrhosis are highly susceptible to the toxic effects of cisplatin.

In our study, tumor size decreased in more than half the patients who received chemoembolization and was stable in most other patients in this group. This finding is consistent with those of previous nonrandomized studies.9,12,25-27 In contrast, tumor size increased in half the patients assigned to conservative management. The pattern of findings for the serum alpha-fetoprotein concentration was similar. The incidence of portal obstruction was higher in the conservative-management group than the chemoembolization group.

Case–control studies have reported that chemoembolization may markedly increase survival in hepatocellular carcinoma.9,10,12,13,28 The type of controls selected in these studies, however, is open to question. The beneficial effect suggested in one study could be related to the very low survival rate in the control group (0 percent at one year, in contrast with 43 percent in our study).12 In a randomized trial comparing chemoembolization and conservative management, survival was not significantly improved in the group treated with doxorubicin and Gelfoam powder.29

We found that Lipiodol chemoembolization significantly inhibited tumor growth. This benefit may have been offset by worsened liver function, particularly in patients with cirrhosis. Prognostic factors for survival were related to growth of the tumor, as well as to signs of severe chronic liver disease, such as abnormal serum bilirubin and serum albumin concentrations, as previously shown.30-32 Liver failure is a frequent cause of death in patients with cirrhosis and hepatocellular carcinoma.20,33-35 Despite the exclusion of patients with overt liver failure, three fifths of the patients had at least one episode of liver decompensation within a few days of treatment, and one patient died. It is possible that Lipiodol chemoembolization may be beneficial in a more select group of patients, but factors predictive of a response to this treatment, without the development of side effects, have yet to be determined.36,37

Funding and Disclosures

Supported by grants from Assistance Publique–Hôpitaux de Paris (87/1990), Société Nationale Française de Gastroentérologie, and Caisses d'Epargne d'Ile-de-France.

We are indebted to Professor Daniel Dhumeaux for help in initiating the study and to Professor Alain Roche for technical advice.

Author Affiliations

Address reprint requests to Dr. Jean-Claude Trinchet at the Service d'Hépato-gastroentérologie, Hôpital Jean Verdier, Ave. du 14 juillet, 93143 Bondy CEDEX, France.

The members of the Groupe d'Etude et de Traitement du Carcinome Hépatocellulaire are listed in the Appendix. This article was prepared by Jean-Claude Trinchet, M.D., Ph.D., Antoine Abou Rached, M.D., and Michel Beaugrand, M.D., Hôpital Jean Verdier, Bondy, France; Didier Mathieu, M.D., Hôpital Henri Mondor, Créteil, France; and Sylvie Chevret, M.D., Ph.D., and Claude Chastang, M.D., Ph.D., Hôpital Saint Louis, Paris. Dr. Trinchet, as principal investigator, assumes full responsibility for the overall content and integrity of the manuscript.

Appendix

The following investigators participated in the Groupe d'Etude et de Traitement du Carcinome Hépatocellulaire:Centre Hospitalier Universitaire, Angers, France — F. Oberti, O. Ruget, P. Calès, and C. Caron-Poitreau; Hôpital Jean Verdier, Bondy, France — J.C. Trinchet, A. Abou Rached, Y. Ajavon, and M. Beaugrand; Hôpital du Haut Lévêque, Pessac, France — F. Dumas, P. Couzigou, and J. Drouillard; Hôpital Saint André, Bordeaux, France — C. Balabaud and P. Grelet; Hôpital Saint Luc et Cliniques Universitaires de Mont-Godinne, Brussels, Belgium — A. Guebel, M. Mélange, J. Pringot, and B. Van Beers; Hôpital Beaujon, Clichy, France — A. Hadengue, V. Vilgrain, and S. Erlinger; Hôpital Henri Mondor, Créteil, France — C. Duvoux, D. Mathieu, and D. Dhumeaux; Hôpital Albert Michallon, Grenoble, France — J.P. Zarski and S. Dalsoglio; Hôpital Edouard Herriot, Lyons, France — F. Mion, P.J. Valette, and P. Paliard; Centre Hospitalier Lyon Sud, Pierre Bénite, France — S. Claudel-Bonvoisin, J.C. Emery, and L. Descos; Hôpital Saint-Luc, Montreal — M. Dagenais, M.P. Dufresne, R. Lapointe, and P.M. Huet; Centre Hospitalier Universitaire, Vandoeuvre-les-Nancy, France — V. Bas, D. Regent, C. Bazin, and M.A. Bigard; Hôpital Guillaume et Renée Laënnec, Nantes, France — B. Humeau, A. Bury, M. Le Rhun, F. Lerat, H. Gibaud, R. Rymer, and L. Le Bodic; Hôtel Dieu, Nantes, France —C. Masliah, B. Pouliquen, and D. Perrin; Hôpital Saint Louis, Paris — C. Chastang; Hôpital Saint Antoine, Paris — O. Chazouillères, A. Loria, R. Poupon, A. Pauwels, N. Mostefa-Kara, V.G. Lévy, P. Balladur, J.P. Deutsch, and J.M. Tubiana; Hôpital Necker, Paris — S. Pol and P. Berthelot; Hôpital de la Salpêtrière, Paris — D. Valla, J.C. Bousquet, and P. Opolon; Centre Hospitalier, Pontoise, France — M. Bouvry; Centre Hospitalier Delafontaine, Saint Denis, France — H. Labadie; Centre Hospitalier Général, Saint-Nazaire, France — T. Martin; Centre Hospitalier Universitaire, Strasbourg, France — J.P. Bronowicki and D. Vetter; Clinique de l'Orangerie, Strasbourg, France — J.J. Wenger; and Centre Hospitalier Universitaire, Toulouse, France — L. Buscail, J. Frexinos, and F. Joffre.

References (37)

  1. 1. Colombo M. Hepatocellular carcinoma in cirrhotics. Semin Liver Dis 1993;13:374-383

  2. 2. Dusheiko GM, Hobbs KEF, Dick R, Burroughs AK. Treatment of small hepatocellular carcinomas. Lancet 1992;340:285-288

  3. 3. Matsui O, Kadoya M, Kameyama T, et al. Benign and malignant nodules in cirrhotic livers: distinction based on blood supply. Radiology 1991;178:493-497

  4. 4. Ensminger WD, Gyves JW. Regional cancer chemotherapy. Cancer Treat Rep 1984;68:101-115

  5. 5. Bruix J, Castells A, Montanya X, et al. Phase II study of transarterial embolization in European patients with hepatocellular carcinoma: need for controlled trials. Hepatology 1994;20:643-650

  6. 6. Raoul JL, Bourguet P, Bretagne JF, et al. Hepatic artery injection of I-131-labeled lipiodol. I. Biodistribution study results in patients with hepatocellular carcinoma and liver metastases. Radiology 1988;168:541-545

  7. 7. Nakao N, Uchida H, Kamino K, et al. Effectiveness of lipiodol in transcatheter arterial embolization of hepatocellular carcinoma. Cancer Chemother Pharmacol 1992;31:Suppl:S72-S76

  8. 8. Kasugai H, Kojima J, Tatsuta M, et al. Treatment of hepatocellular carcinoma by transcatheter arterial embolization combined with intraarterial infusion of a mixture of cisplatin and ethiodized oil. Gastroenterology 1989;97:965-971

  9. 9. Bismuth H, Morino M, Sherlock D, et al. Primary treatment of hepatocellular carcinoma by arterial chemoembolization. Am J Surg 1992;163:387-394

  10. 10. Aoyama K, Tsukishiro T, Okada K, et al. Evaluation of transcatheter arterial embolization with epirubicin-lipiodol emulsion for hepatocellular carcinoma. Cancer Chemother Pharmacol 1992;31:Suppl:S55-S59

  11. 11. Nakamura H, Mitani T, Murakami T, et al. Five-year survival after transcatheter chemoembolization for hepatocellular carcinoma. Cancer Chemother Pharmacol 1994;33:Suppl:S89-S92

  12. 12. Vetter D, Wenger J, Bergier J, Doffoel M, Bockel R. Transcatheter oily chemoembolization in the management of advanced hepatocellular carcinoma in cirrhosis: results of a Western comparative study in 60 patients. Hepatology 1991;13:427-433

  13. 13. Yamashita Y, Takahashi M, Koga Y, et al. Prognostic factors in the treatment of hepatocellular carcinoma with transcatheter arterial embolization and arterial infusion. Cancer 1991;67:385-391

  14. 14. Venook AP, Stagg RJ, Lewis BJ, et al. Chemoembolization for hepatocellular carcinoma. J Clin Oncol 1990;8:1108-1114

  15. 15. Stuart K, Stokes K, Jenkins R, Trey C, Clouse M. Treatment of hepatocellular carcinoma using doxorubicin/ethiodized oil/gelatin powder chemoembolization. Cancer 1993;72:3202-3209

  16. 16. Whitehead J. The design and analysis of sequential clinical trials. 2nd ed. New York: Ellis Horwood, 1992.

  17. 17. Kaplan EL, Meier P. Nonparametric estimation from incomplete observations. J Am Stat Assoc 1958;53:457-481

  18. 18. Peto R, Peto J. Asymptotically efficient rank invariant test procedures. J R Stat Soc [A] 1972;135:185-206

  19. 19. Cox DR. Regression models and life-tables. J R Stat Soc [B] 1972;34:187-220

  20. 20. Okuda K, Ohtsuki T, Obata H, et al. Natural history of hepatocellular carcinoma and prognosis in relation to treatment: study of 850 patients. Cancer 1985;56:918-928

  21. 21. Nakamura H, Tanaka T, Hori S, et al. Transcatheter embolization of hepatocellular carcinoma: assessment of efficacy in cases of resection following embolization. Radiology 1983;147:401-405

  22. 22. Kan ZX, Sato M, Ivancev K, et al. Distribution and effect of iodized poppyseed oil in the liver after hepatic artery embolization: experimental study in several animal species. Radiology 1993;186:861-866

  23. 23. Yodono H, Saito Y, Saikawa Y, Midorikawa H, Yokoyama Y, Takekawa S. Combination chemoembolization therapy for hepatocellular carcinoma: mainly, using cisplatin (CDDP). Cancer Chemother Pharmacol 1989;23:Suppl:S42-S44

  24. 24. Imaoka S, Sasaki Y, Shibata T, et al. A pre-operative chemoembolization therapy using lipiodol, cisplatin and gelatin sponge for hepatocellular carcinoma. Cancer Chemother Pharmacol 1989;23:Suppl:S126-S128

  25. 25. Sasaki Y, Imaoka S, Kasugai H, et al. A new approach to chemoembolization therapy for hepatoma using ethiodized oil, cisplatin, and gelatin sponge. Cancer 1987;60:1194-1203

  26. 26. Matsui O, Kadoya M, Yoshikawa J, Gabata T, Takashima T, Demachi H. Subsegmental transcatheter arterial embolization for small hepatocellular carcinomas: local therapeutic effect and 5-year survival rate. Cancer Chemother Pharmacol 1994;33:Suppl:S84-S88

  27. 27. Yu YQ, Xu DB, Zhou XD, Lu JZ, Tang ZY, Mack P. Experience with liver resection after hepatic arterial chemoembolization for hepatocellular carcinoma. Cancer 1993;71:62-65

  28. 28. Nakamura H, Hashimoto T, Oi H, Sawada S. Transcatheter oily chemoembolization of hepatocellular carcinoma. Radiology 1989;170:783-786

  29. 29. Pelletier G, Roche A, Ink O, et al. A randomized trial of hepatic arterial chemoembolization in patients with unresectable hepatocellular carcinoma. J Hepatol 1990;11:181-184

  30. 30. Akashi Y, Koreeda C, Enomoto S, et al. Prognosis of unresectable hepatocellular carcinoma: an evaluation based on multivariate analysis of 90 cases. Hepatology 1991;14:262-268

  31. 31. Attali P, Prod'Homme S, Pelletier G, et al. Prognostic factors in patients with hepatocellular carcinoma: attempts for the selection of patients with prolonged survival. Cancer 1987;59:2108-2111

  32. 32. Calvet X, Bruix J, Gines P, et al. Prognostic factors of hepatocellular carcinoma in the west: a multivariate analysis in 206 patients. Hepatology 1990;12:753-760

  33. 33. Ohnishi K, Tanabe Y, Ryu M, et al. Prognosis of hepatocellular carcinoma smaller than 5 cm in relation to treatment: study of 100 patients. Hepatology 1987;7:1285-1290

  34. 34. Calvet X, Bruix J, Bru C, et al. Natural history of hepatocellular carcinoma in Spain: five years' experience in 249 cases. J Hepatol 1990;10:311-317

  35. 35. The Liver Cancer Study Group of Japan. Primary liver cancer in Japan: clinicopathologic features and results of surgical treatment. Ann Surg 1990;211:277-287

  36. 36. Khan KN, Nakata K, Kusumoto Y, et al. Evaluation of nontumorous tissue damage by transcatheter arterial embolization for hepatocellular carcinoma. Cancer Res 1991;51:5667-5671

  37. 37. Miyoshi S, Minami Y, Kawata S, et al. Changes in hepatic functional reserve after transcatheter embolization of hepatocellular carcinoma: assessment by maximal removal rate of indocyanine green. J Hepatol 1988;6:332-336

Citing Articles (637)

    Letters

    Figures/Media

    1. Figure 1. Design of the Triangular Test and the Results of the Fifth Sequential Analysis.
      Figure 1. Design of the Triangular Test and the Results of the Fifth Sequential Analysis.

      The five interim data sets resulted in five pairs of values for the z statistic, which approximately represents the effect of lipiodol chemoembolization on overall survival as compared with that of conservative management, and the V statistic, which approximately represents the sample size. The corresponding points were plotted on the graph as they became available and compared with the stopping boundary set out by the triangular test (see the text). At the fifth analysis, the lower triangular boundary was crossed, causing the trial to be stopped.

    2. Table 1. Base-Line Characteristics of the 96 Patients Studied, According to Treatment Group.
      Table 1. Base-Line Characteristics of the 96 Patients Studied, According to Treatment Group.
    3. Figure 2. Kaplan–Meier Estimates of Survival in 50 Patients Assigned to Lipiodol Chemoembolization and 46 Patients Assigned to Conservative Management (P = 0.13 by the Log-Rank Test).
      Figure 2. Kaplan–Meier Estimates of Survival in 50 Patients Assigned to Lipiodol Chemoembolization and 46 Patients Assigned to Conservative Management (P = 0.13 by the Log-Rank Test).
    4. Table 2. Maximal Changes in Tumor Size and Serum Alpha-fetoprotein Concentrations and Occurrence of Portal Obstruction in the 81 Patients Who Survived More than Two Months after Randomization.
      Table 2. Maximal Changes in Tumor Size and Serum Alpha-fetoprotein Concentrations and Occurrence of Portal Obstruction in the 81 Patients Who Survived More than Two Months after Randomization.
    5. Table 3. Complications of 148 Courses of Lipiodol Chemoembolization in the 50 Patients Assigned to Such Treatment.
      Table 3. Complications of 148 Courses of Lipiodol Chemoembolization in the 50 Patients Assigned to Such Treatment.

      More Research