Update on Hepatic Intra-Arterial Chemotherapy

July 1, 1997

Colorectal carcinoma is a common problem in the United States, and the liver is the most frequent site of metastatic disease. Because there is a good pharmacologic rationale for the use of hepatic intra-arterial (HIA) therapy, and because of the disappointing survival observed with systemic chemotherapy, studies of hepatic arterial infusion have been conducted.

Colorectal carcinoma is a common problem in the United States, and theliver is the most frequent site of metastatic disease. Because there isa good pharmacologic rationale for the use of hepatic intra-arterial (HIA)therapy, and because of the disappointing survival observed with systemicchemotherapy, studies of hepatic arterial infusion have been conducted.

Reviewing the Randomized Trials

In reviewing the randomized studies, Dr. Venook cites American trialsseparately and asserts that none of them showed an increase in survivalwith regional therapy. However, two large American studies permitted acrossover from systemic to hepatic arterial therapy.[1,2] This could haveincreased survival in the systemic arm and served to diminish a potentiallysignificant difference between the systemic and HIA therapy groups. Inthe Memorial Sloan-Kettering Cancer Center (MSKCC) study,[1] patients onthe systemic arm had infusion ports placed at surgery so that they couldreceive HIA therapy at a later date. The most common reason for not crossingover to HIA therapy was occlusion of the infusion ports.

In this study, 60% of patients on the systemic arm crossed over to HIAtherapy. Of these, 26% had a partial response, while 32% had a minor responseor stable disease with HIA FUDR after progressing on systemic FUDR. Thepatients who crossed over to receive hepatic arterial infusion experienceda doubling of survival compared to those who never received HIA therapy:18 vs 8 months in the MSKCC study and 24 vs 12 months in the Northern OncologyGroup (NCOG) study.[2]

The two other American studies were too small to assess survival differencesbetween the arms.[3,4] In one of these studies, 49% of patients in theHIA therapy group received inadequate treatment.

Thus, the six completed randomized studies mentioned had three majorproblems: crossover design in two American studies, inadequate systemicchemotherapy in two European studies, and small numbers of patients intwo American studies.

Dr. Venook then touches on the meta-analyses of HIA therapy vs systemictherapy that have been performed.[5,6] An increase in survival in the hepaticarterial group was observed if all studies were included. However, no increasein survival was seen if the two European studies in which inadequate systemictherapy was administered were excluded. In a meta-analysis, it is desirableto include only prospective randomized studies utilizing adequate therapy.However, if this were done in the case of systemic vs HIA therapy, onewould be left with small studies, and it would be difficult to make conclusionsabout survival.

Dr. Venook offers several reasons why an improvement in response ratedoes not translate into improved survival. He makes the point that colorectalcancer is a systemic disease, and therefore, regional therapy should notaffect survival. In autopsy data, however, Weiss noted that 46% of patientsdying of colorectal cancer had metastases limited to the liver.[7]

It is also clear that a significant proportion of patients with colorectalcarcinoma die of progressive hepatic disease despite the presence of diseasein other organs. Toxicity and surgical complications may affect survivalwith HIA, but until we have a study without a crossover design that hasan adequate number of patients, we will not be able to answer the questionabout whether there is a survival advantage for HIA.

Potential Areas of Improvement

A strong part of this review focuses on potential areas of improvement,including the standardization of surgical techniques and the optimizationof delivery of chemotherapy to the liver in an attempt to reduce toxicityand increase response rates.

The review of surgical issues is excellent. Certainly, this type oftreatment requires a team approach. First, the radiologist has to carefullydefine the arterial supply, so that the surgeon knows where the vesselsto the duodenum and stomach are located and also whether there are anyreplaced or accessory arteries.

The surgeon must devascularize the stomach and duodenum and ensure adequateperfusion of the liver. There is a learning curve, with fewer complicationsarising after surgery performed by more experienced surgeons. In a reviewby Campbell et al, inexperienced surgeons had a technical complicationrate of 37% vs 7% for experienced surgeons.[8]

The nuclear medicine physician must read the perfusion scans accuratelyand make sure that there are no small areas of extrahepatic perfusion;obviously, the large areas of missed perfusion are easier to identify.

The medical oncologist must carefully follow the patient's liver functiontests and must be cognizant of the need for dose modifications dictatedby changes in liver function. The types of modifications that we use atMSKCC are summarized in Table 1.

The multidisciplinary approach may be difficult to implement in somecenters, but it is important for this type of therapy in order to maximizebenefit and reduce both the financial costs and medical risks of hepaticinfusional chemotherapy.

CALGB Study Will Address Survival Issues

During the last 20 years, there has been no change in the survival formetastatic colorectal carcinoma. More than 2,000 patients have been randomizedto 5-FU plus leucovorin vs 5-FU alone. A meta-analysis of these studiesdemonstrated a median survival of 11 months and a 2-year survival of lessthan 20% for both treatment groups.[9]

Recently, new drugs have been developed for the treatment of colorectalcarcinoma: irinotecan (Camptosar), a camptothecan derivative; tomudex,a new thymidylate synthase inhibitor; and oxalaplatin, a new platinum.All of these drugs produce response rates similar to those obtained with5-FU and leucovorin. In many studies of these new agents, survival is similar,with only 20% of patients alive at 2 years. Whether combinations of theseagents will increase survival is yet to be tested.

In three recent studies of HIA using FUDR plus leucovorin, FUDR plusdexamethasone, and FUDR plus leucovorin and dexamethasone, the median survivalswere 23, 23, and 27 months, respectively.[10-12] The 2-year survival ratesin these studies were 61%, 44%, and 47%, respectively.

Because of this apparent survival advantage of HIA chemotherapy comparedto systemic chemotherapy, a new randomized study was initiated by the CALGBto ascertain whether these results can be reproduced. The CALGB study willaddress many issues that were not answered in the earlier trials:

  • Does HIA therapy improve survival, as compared with adequate systemictherapy?
  • Is there a difference in quality of life between the two treatments?
  • Is there a difference in financial cost over the entire course of treatment?
  • Is hepatic arterial infusional chemotherapy a more effective treatmentthan systemic chemotherapy for patients with metastatic colorectal carcinomalimited to the liver?

The CALGB study will not allow a crossover, will have a large numberof patients, and will utilize the best available chemotherapy in each arm.Hopefully, with the cooperation of many centers, this trial will provideus with the answers to these questions.


1. Kemeny N, Daly J, Reichman B, et al: Intrahepatic or systemic infusionof fluorodeoxyuridine in patients with liver metastases from colorectalcarcinoma. Ann Intern Med 107(4):459-465, 1987.

2. Hohn DC, Stagg RJ, Friedman MA, et al: A randomized trial of continuousintravenous vs hepatic intraarterial floxuridine in patients with colorectalcancer metastatic to the liver: The Northern Oncology Group Trial. J ClinOncol 7:1646-1654, 1989.

3. Chang AE, Schneider PD, Sugarbaker PH, et al: A prospective randomizedtrial of regional vs systemic continuous 5-fluorodeoxyuridine chemotherapyin the treatment of colorectal liver metastases. Ann Surg 206:685-693,1987.

4. Martin JK Jr, O'Connell MJ, Wieand HS, et al: Intra-arterial floxuridinevs systemic 5-FU for hepatic metastases from colorectal cancer. Arch Surg125:1022-1027, 1990.

5. Harmantas A, Rotstein LE, Langer B: Regional vs systemic chemotherapyin the treatment of colorectal carcinoma metastatic to the liver. Cancer78:1639-1645, 1996.

6. Stagg RJ, Venook AP, Chase JL, et al: Alternating hepatic intra-arterialfloxuridine and 5-FU: A less toxic regimen for treatment of liver metastasesfrom colorectal cancer. J Natl Cancer Inst 83:423-428, 1991.

7. Weiss L, Grundmann E, Torhorst J, et al: Hematogenous metastaticpatterns in colonic carcinoma: An analysis of 1541 necropsies. J Pathol150:195-203, 1986.

8. Campbell CA, Burns RC, Sitzmann JV, et al: Regional chemotherapydevices: Effect of experience and anatomy on complications. J Clin Oncol11:822-826, 1993.

9. Advanced colorectal cancer meta-analysis project. Modulation of fluorouracilby leucovorin in patients with advanced colorectal cancer: Evidence interms of response rate. J Clin Oncol 10:896-903, 1992.

10. Kemeny N, Seiter K, Conti JA, et al: Hepatic arterial floxuridineand leucovorin for unresectable liver metastases from colorectal carcinoma.Cancer 73:1134-1142, 1994.

11. Kemeny N, Seiter K, Niedzwiecki D, et al: A randomized trial ofintrahepatic infusion of fluorodeoxyuridine with dexamethasone versus fluorodeoxyuridinealone in the treatment of metastatic cancer. Cancer 69:327-334, 1992.

12. Kemeny N, Conti JA, Cohen A, et al: Phase II study of hepatic arterialfloxuridine, leucovorin, and dexamethasone for unresectable liver metastasesfrom colorectal carcinoma. J Clin Oncol 12:2288-2295, 1994