This article is a US Government work and, as such, is in the public domain in the United States of America.
Many patients with colorectal carcinoma develop unresectable metastases confined to the liver that remain the life-limiting component of disease despite best available systemic or regional chemotherapy. In the current study, the authors present their results using vascular isolation and perfusion of the liver for individuals with progressive, unresectable liver metastases from colorectal carcinoma that were refractory to both previous systemic and regional chemotherapy.
Seven patients with refractory, progressive, unresectable colorectal carcinoma metastases confined to the liver underwent a 60-minute hyperthermic (39–40 °C) isolated hepatic perfusion (IHP) and were followed for toxicity, response, and survival.
There was no surgical- or treatment-related mortality; all patients experienced transient Grade 3–4 (according to National Cancer Institute common toxicity criteria) hepatic toxicity. At a median potential follow-up of 16 months, the overall objective radiographic response rate (all partial responses) was 71% (5 of 7 assessable patients). It is interesting to note that two patients who were treated with tumor necrosis factor (TNF) alone demonstrated no response to therapy compared with all five patients who were treated with melphalan and TNF (three patients) or melphalan alone (two patients). For the 5 patients who responded to treatment, the median duration of response was 10 months (range, 10–13 months) and in all 7 patients the mean overall survival was 19.7 months (range, 2–33 months), including 5 months and 7.5 months, respectively, for the 2 patients treated with TNF alone.
The results of the current study demonstrate that IHP using melphalan with or without TNF has significant antitumor activity in this patient population. IHP deserves continued clinical evaluation as a therapeutic modality for patients with unresectable colorectal carcinoma metastases to the liver. Cancer 2002;95:730–6. Published 2002 American Cancer Society.
Combination systemic chemotherapy using 5-fluorouracil (5-FU) and leucovorin (LV) administered with camptothecin currently is considered first-line therapy for metastatic colorectal carcinoma. However, the overall confirmed response rates in 2 large random assignment trials was only 39% and the progression-free survival was only 7 months.1, 2 Hepatic artery infusion (HAI) of chemotherapy with floxuridine (FUDR)- or 5-FU-based regimens has been reported to result in response rates of 50–78% in previously untreated patients.3, 4 Although a number of prospective random assignment trials consistently have shown significantly higher response rates with HAI using FUDR compared with systemic chemotherapy,5–8 to our knowledge individual studies have not consistently demonstrated an improvement in overall survival. Two meta-analyses have demonstrated a modest but significant improvement in survival with HAI therapy; 1 study reported a median survival of 16 months versus 12.2 months in patients treated with HAI compared with those treated with systemic chemotherapy.9, 10 Approximately 50% of all patients who respond to HAI therapy will fail with progression of hepatic disease.10 For patients who have received previous systemic chemotherapy or who have a tumor burden in the liver > 25% the response rates and duration of response with HAI are decreased.11–13 Fordy et al reported an overall response rate of only 14% with a median response duration of 7 months using HAI with FUDR in 35 patients with 5-FU-resistant colorectal carcinoma liver metastases.12 Clearly, additional treatment options are needed for patients with refractory hepatic metastases.
Isolated hepatic perfusion (IHP) is a regional therapy in which the vascular supply of the liver is isolated, all collateral circulation is controlled, and the organ is perfused using hyperthermia with chemotherapy or biologic agents (i.e., tumor necrosis factor [TNF]) via a recirculating oxygenated perfusion circuit.14 Several centers have reported results using IHP with hyperthermia alone, melphalan or other chemotherapeutics, and TNF.15–18 We previously reported that IHP with TNF and melphalan is associated with a 74% radiographic objective response rate, which is observed in a variety of histologies.19 More recently, we reported that IHP using melphalan alone followed by HAI using an FUDR-based regimen also is associated with a > 70% response rate and a median duration of response of 14.5 months in patients with colorectal carcinoma.20 Moreover, IHP was reported to have significant efficacy, even in those patients who had failed previous systemic chemotherapy or who had a hepatic replacement by tumor of > 25%.
Because many patients who have isolated, unresectable hepatic metastases from colorectal carcinoma will fail systemic or HAI chemotherapy with progression of disease in the liver, we analyzed a cohort of seven colorectal carcinoma patients who underwent IHP after having failed both previous systemic and HAI therapy to determine the potential utility of IHP in this setting.
MATERIALS AND METHODS
Between April 1993 and February 2001, 105 patients with metastatic colorectal carcinoma that was confined to the liver underwent IHP on protocols approved by the National Cancer Institute (NCI) Institutional Review Board and, when indicated, NCI Cancer Therapy Evaluation Program. Seven patients were identified who previously had received systemic and HAI chemotherapy for the treatment of established hepatic metastases from colorectal carcinoma and who were determined to have progressive disease based on imaging studies or rising serum carcinoembryonic antigen levels. All patients had measurable, biopsy proven colorectal carcinoma metastases and underwent computed tomography (CT) scans of the chest, abdomen, and pelvis to rule out extrahepatic sites of metastases. Eligibility criteria included an Eastern Cooperative Oncology Group (ECOG) performance status of 0 or 1, a serum bilirubin level < 2 mg/dL, a platelet count > 150,000/mL, a serum creatinine level ≤ 1.5 mg/dL, and a partial thromboplastin time within 1 second of the upper of normal. Patients were excluded if there was evidence of cirrhosis or portal hypertension.
All patients underwent preoperative arteriography or magnetic resonance imaging (MRI) and angiography to assess the adequacy of vessels for cannulation before laparotomy. All patients were evaluated at 6 weeks after treatment and then at 3-month intervals for the first 12 months and 4-month intervals for the second 12 months after IHP. Signed informed consent was obtained from each patient before participation on the trial.
Our technique of IHP has been described in detail previously.21 In brief, the liver is mobilized completely from all diaphragmatic attachments. Limited malignant lymphadenopathy in the porta hepatis is not considered a contraindication to IHP and is resected in the course of preparing the porta hepatis structures for IHP. Venous outflow from the liver is collected using a cannula positioned in an isolated segment of the retrohepatic inferior vena cava (IVC). All collateral vessels to the retrohepatic IVC, including the right adrenal vein, are divided in preparation for treatment. Inflow to the liver is via a cannula positioned in the gastroduodenal artery and during treatment arterial inflow to the liver is temporarily controlled with a vascular occluding clamp across the common hepatic artery just distal to the celiac axis. Portal venous flow and IVC flow below the liver are shunted to the systemic circulation during IHP using a bypass circuit in which blood is shunted to the axillary vein using a centrifugal pump.
The perfusion circuit is comprised of a roller pump, oxygenator, and heat exchanger containing a 700 mL perfusate of saline, 300 mL of packed erythrocytes, sodium bicarbonate, and heparin. The temperature of the perfusate is controlled by a Hemotherm water cooler/heater (Model #400; Cincinnati Sub Zero Products, Cincinnati, OH) so that hepatic tissue temperatures are maintained near 40 °C during treatment. Arterial and venous perfusate blood gas analyses are performed during IHP and sodium bicarbonate is added to the circuit to maintain an arterial perfusate pH level between 7.2 and 7.3. Perfusion flow rates are adjusted for maximum flow while maintaining a line pressure of < 200 mL of Hg and a stable reservoir volume. For patients treated with TNF, a continuous intraoperative leak monitoring system with I-131-labeled human serum albumin as previously described was used.22 We subsequently have shown that by using a standard surgical mobilization of the liver and IHP technique, complete vascular isolation of the liver can be achieved routinely and the leak monitoring system is no longer used.21
Two patients were treated with TNF alone (at doses of 0.6 mg and 1 mg, respectively) as part of a Phase I dose escalation trial and 5 patients were treated with a combination of 1.5 mg/kg of melphalan with or without 1 mg of TNF as prescribed on 2 subsequent Phase II IHP trials. The perfusion interval was 60 minutes after the addition of the agents and at the end of the treatment the liver was flushed through the perfusion circuit with 1.5 L of crystalloid and 1.5 L of colloid; the portal vein was flushed with 1 L of normal saline.
Systemic and regional toxicities were graded according to the NCI common toxicity criteria. Regional (hepatic) toxicities were graded as elevations in liver function tests that persisted for > 7 days after treatment; nonhepatic toxicities were graded as those that did not reverse within 24 hours after the surgical procedure. All responses were determined based on radiographic findings using MRI or CT scans. Complete response was defined as the complete disappearance of all established tumor without evidence of new lesions for at least 30 days and a partial response was defined as a ≥ 50% reduction in the sum of the product of the perpendicular dimensions of all measurable lesions in the liver for at least 30 days without the progression of any single lesion or the development of new sites of hepatic disease. A minor response was defined as a 25–49% reduction in size by these criteria. No patients were lost to follow-up and survival data were complete through November 2001. Survival analysis was performed using Kaplan–Meier statistics. All data are presented as the mean ± the standard deviation or as the median followed by the range.
Table 1 presents the clinical profile of the seven colorectal carcinoma patients who had failed previous systemic and HAI therapy for established unresectable liver metastases prior to IHP. All patients received 5-FU and LV and four patients subsequently received second-line systemic therapy. All patients received HAI therapy, four of whom received an FUDR-based regimen, 2 of whom received 5-FU, and 1 who received cisplatin. There was an advanced burden of disease in the liver in all patients as indicated by a median of 11 hepatic metastases, a median size of 8 cm for the largest metastasis, and an average of 50% hepatic replacement by tumor.
Time from diagnosis of primary tumor (mos) (range)
ECOG performance status
Prior therapeutic treatment
Cisplatin and interferon
No. of previous therapies per patient
No. of hepatic metastases Median (range)
Size of largest lesion (cm) Median (range)
Percent hepatic replacement Median (range)
All patients were treated using an identical perfusion technique and parameters with respect to the duration of treatment and degree of hyperthermia (Table 2). Two patients received TNF alone (0.6 mg and 1 mg, respectively) and 5 patients received either 1.5 mg/kg of melphalan alone (n = 2 patients) or combined with 1 mg of TNF (n = 3 patients). The high perfusion pressures reflect the measured pressure within the arterial line of the circuit and represent the pressure generated by flow through a 3-mm arterial cannula in the gastroduodenal artery. The proper hepatic artery routinely is palpated at regular intervals during IHP and the estimated pressure in the hepatic arterial tree is consistently lower than that in the perfusion circuit. There was no leak of perfusate based on either the use of a continuous intraoperative leak monitoring system in five patients or the presence of stable perfusion parameters and reservoir volume in the circuit during treatment in two patients who received melphalan alone.
Table 2. Perfusion and Surgical Data
TNF: tumor necrosis factor.
Surgical time (hrs) (range)
TNF + melphalan
Melphalan dose (mg) (range)
Perfusion flow rate (mL/min) (range)
Perfusion pressure (mm Hg) (range)
Bypass flow rate (mL/min) (range)
Central liver temperature (°C) (range)
Estimated blood loss (L) (range)
% perfusate leak
The toxicities associated with treatment were consistent with our previously published results and were transient in nature.19, 23 All patients had some degree of hepatic toxicity as reflected in elevations in transaminase levels and serum bilirubin levels that were entirely reversible in nature (Table 3). Three patients who received TNF in the perfusate had transient post-IHP hypotension that resolved within 24 hours of the perfusion and was treated successfully with volume resuscitation only.
Table 3. Treatment-Related Toxicity (According to National Cancer Institute Common Toxicity Criteria)
There were seven patients evaluable for response. Both patients who received TNF alone in the perfusate had no meaningful response to therapy (one had stable disease and one patient achieved a marginal response of 3 months' duration), which is consistent with the experience of others demonstrating no antitumor activity with TNF alone in isolated limb perfusion for in-transit melanoma or unresectable extremity sarcoma.24, 25 There were 5 bona fide radiographic partial responses (71%) with a mean duration of response of 10 months (range, 10–13 months) (Table 4) (Figs. 1 and 2). The relatively short time to systemic recurrence reflects the high risk of harboring an occult systemic metastases in this patient group. Despite early systemic recurrence, the mean overall survival was 19.7 months (range, 3–33 months) (Fig. 3).
Table 4. Results of Treatment
CEA: carcinoembryonic antigen.
Time to in field recurrence (mos)
Time to systemic recurrence (mos)
% decrease of CEA level
Survival time (mos)
Patients with progressive unresectable metastases from colorectal carcinoma confined to the liver represent a significant clinical problem. The duration of response after best available combination chemotherapy or HAI using an FUDR-based regimen has been reported to be 7 months and 14 months, respectively.1, 2, 10 Moreover, patients who previously failed 5-FU-based systemic chemotherapy for the treatment of established hepatic metastases are reported to have a lower response rate to HAI therapy when HAI is used as second-line treatment. Response rates ranging from 14–33% have been reported in individuals with 5-FU-resistant hepatic metastases from colorectal carcinoma, with a reported median response duration of only 7 months.12, 13
Vascular isolation and perfusion of the liver, also known as IHP, has been in clinical evaluation for nearly 50 years but because it is a complex treatment to administer, is associated with significant surgical morbidity, and has not had its efficacy critically evaluated in initial clinical trials, has not gained widespread or consistent clinical application. Over the past 10 years we and others have evaluated IHP using melphalan alone or melphalan and TNF for unresectable hepatic metastases from a variety of histologies.16, 25 Using a consistent surgical technique to ensure complete vascular isolation and using doses of agents determined from appropriately designed Phase I trials, IHP using melphalan alone or melphalan and TNF can be administered safely. The data presented in the current study demonstrate the efficacy of IHP in patients with unresectable, progressive hepatic metastases from colorectal carcinoma that is refractory to both previous systemic and HAI treatment. The cohort presented in the current study not only had refractory disease but also a relatively advanced disease burden in the liver as reflected by the number and size of the metastases and the significant percent of hepatic replacement by tumor. Despite these adverse factors, IHP with melphalan or melphalan and TNF was found to result in significant antitumor activity with an overall radiographic partial response rate of 71% and a median response duration of 10 months.
Although the cohort of patients treated in the current study was small, these data provide some justification for the continued clinical evaluation of IHP for the treatment of unresectable metastases confined to the liver and, in particular, for those patients who have failed previous therapies. It is interesting to note that the lack of efficacy in the 2 colorectal carcinoma patients treated with TNF alone is consistent with the results obtained in the larger cohort of 16 patients treated in a Phase I TNF dose escalation trial. Of the 16 individuals treated on study, the 3 patients who had some evidence of antitumor activity actually were treated at the lowest doses of TNF and were the initial patients entered on the study. When it was noted that the venous pH in the perfusate was < 7 it was corrected with sodium bicarbonate to maintain a pH level between 7.2–7.3. Coincident with this, all measurable antitumor effects of IHP with escalating doses of TNF alone were eliminated, suggesting that the initial results in the first patients treated at the lowest doses of TNF may have been nonspecific effects from ischemia or acidosis within the tumor bed.25 To our knowledge, TNF alone has never been associated with meaningful antitumor activity when used in the isolation perfusion of any site against any tumor type.24, 26, 27
The data relating to the five individuals treated using melphalan with or without TNF also are consistent with a previous report examining IHP in patients with unresectable hepatic metastases from colorectal carcinoma.20 In a cohort of 50 evaluable patients, the overall response rate was 74% and was not found to be different between individuals treated with IHP using TNF and melphalan and those treated with IHP using melphalan alone followed by maintenance therapy with FUDR administered via an implantable pump.
The use of IHP in this select patient group highlights several interesting points. First the cohort represents a fairly favorable patient group in that disease progression was somewhat indolent, which afforded an opportunity for patients to receive several sequential types of therapy without experiencing rapid tumor progression with resultant hepatic decompensation. That somewhat favorable biology may be reflected in the fact that the mean survival after IHP was 19.7 months and in 1 circumstance extended to nearly 3 years. Nevertheless, it is interesting to note that the 2 patients treated with TNF alone survived only 5 months and 7.5 months, respectively, after IHP. Despite the heavy pretreatment history in this cohort, the profile of hepatic toxicities observed after the high-dose infusion of melphalan to the liver is consistent with our previous experience in patients who have been less heavily pretreated or who were untreated prior to therapy.
The results of the current study of a select group of patients with refractory, progressive, unresectable colorectal carcinoma metastases confined to the liver demonstrate that a single 60-minute IHP using melphalan with or without TNF can result in significant antitumor activity and a favorable survival. These results indicate that further clinical evaluation of IHP for unresectable hepatic metastases is warranted.