Bevacizumab in Combination With Fluorouracil and Leucovorin: An Active Regimen for First-Line Metastatic Colorectal Cancer
http://www.100md.com
《临床肿瘤学》
the Duke University Medical Center, Durham, NC
Kaiser Permanente Medical Care Program, Northern California
Genentech Inc, South San Francisco
Department of Medicine, University of California at Los Angeles, Los Angeles, CA
The Sarah Cannon Cancer Center, Nashville, TN
Hematology and Cancer Center of Northeastern Pennsylvania, Dunmore, PA
Division of Medical Oncology, Vanderbilt University Medical Center, Nashville, TN
ABSTRACT
PURPOSE: In a phase III trial, combining bevacizumab (BV)—a recombinant, humanized, monoclonal antibody targeting vascular endothelial growth factor—with irinotecan, bolus fluorouracil (FU), and leucovorin (LV; IFL) increased survival compared with IFL alone in first-line treatment of patients with metastatic colorectal cancer (CRC). Results for the parent study of IFL/BV versus IFL/placebo are reported elsewhere. Here, we describe efficacy and safety results for the third patient cohort in this trial, who received BV combined with FU/LV, and compare them with results for concurrently enrolled patients who received IFL.
METHODS: Patients (N = 923) were randomly assigned to receive IFL/placebo (control), IFL/BV, or FU/LV/BV. Bevacizumab (Avastin; Genentech Inc, South San Francisco, CA) 5 mg/kg was administered intravenously every 2 weeks. Before an interim analysis confirmed acceptable safety for IFL/BV, 313 patients were concurrently randomly assigned to these three arms; after this analysis, the FU/LV/BV arm was discontinued.
RESULTS: Median overall survivals were 18.3 and 15.1 months with FU/LV/BV (n = 110) and IFL/placebo (n = 100), respectively. Median progression-free survivals were 8.8 and 6.8 months, respectively. Overall response rates were 40.0% and 37.0%, and median response durations were 8.5 and 7.2 months, respectively. Adverse events consistent with those expected from FU/leucovorin- or IFL-based regimens were seen, as were modest increases in hypertension and bleeding in the bevacizumab arm, which were generally easily managed.
CONCLUSION: The FU/LV/BV regimen seems as effective as IFL and has an acceptable safety profile. FU/LV/BV is an active alternative treatment regimen for patients with previously untreated metastatic CRC.
INTRODUCTION
Patients with metastatic colorectal cancer (CRC) have a poor prognosis. Until recently, the standard treatment for metastatic CRC in the United States was chemotherapy with fluorouracil (FU) plus leucovorin (LV), a biochemical modulator of FU.1,2 The combination of FU/LV improves response rates among patients with metastatic CRC compared with FU alone.3 Addition of irinotecan to FU/LV (combination IFL therapy) has been shown to increase survival, time to disease progression, and response rates in two randomized clinical trials, each with approximately 400 patients.4,5 However, some patients are not ideal candidates for irinotecan-containing regimens or do not tolerate these regimens, which are commonly associated with significant myelosuppression, asthenia, and GI toxicities.4-6 Therefore, alternative combination therapies that are active and better tolerated are needed.
Bevacizumab (BV; Avastin; rhuMAb vascular endothelial growth factor [VEGF]; Genentech Inc, South San Francisco, CA) is a recombinant humanized monoclonal antibody that blocks the activity of VEGF.7 VEGF is a central mediator of tumor-related angiogenesis, and its expression has been shown to be prognostic for a number of solid tumors.8-10 Results of an initial randomized phase II trial in 104 patients with metastatic CRC suggested that adding BV to FU/LV increased clinical benefit relative to FU/LV alone, with an acceptable toxicity profile.11 Subsequently, a large, phase III, multicenter, double-blind, randomized trial in patients with metastatic CRC demonstrated that the addition of BV to first-line IFL chemotherapy (n = 402) prolonged survival by 4.7 months compared with IFL alone (n = 411; P < .001).12 Significant benefit was also demonstrated for all secondary efficacy end points: progression-free survival increased from 6.2 months to 10.6 months (P < .001), overall response rate increased from 34.8% to 44.8% (P = .004), and the duration of response increased from 7.1 months to 10.4 months (P = .001).
Because no safety data on the IFL/BV combination were available at the time the phase III trial began, the study design initially included a third treatment arm in which patients received BV in combination with FU and LV (FU/LV/BV arm). A planned interim safety analysis was performed for the purpose of selecting one BV-containing arm for comparison with IFL. If there had been excess toxicity with IFL/BV, that arm would have been discontinued. However, after the planned interim safety analysis revealed an acceptable safety profile for the IFL/BV combination, enrollment onto the FU/LV/BV treatment arm stopped to allow full enrollment in the two primary study arms (ie, IFL/placebo and IFL/BV).
Because some patients do not tolerate more aggressive chemotherapy, the combination of BV and FU/LV remains a potentially useful therapeutic alternative. In this article, we present the results of an analysis of data from the FU/LV/BV arm of the phase III trial and for concurrently enrolled patients in the primary study arm that was treated with irinotecan/FU/LV and placebo. This is the only randomized study to date directly comparing the addition of a traditional cytotoxic agent (irinotecan) with the addition of a targeted antiangiogenic agent (BV) to a FU/LV-based chemotherapy regimen.
METHODS
Patients
Patients eligible to participate in the phase III study had histologically confirmed, bidimensionally measurable, metastatic CRC. Patients had to be at least 18 years old, have an Eastern Cooperative Oncology Group (ECOG) performance status13 of 0 or 1, have a life expectancy greater than 3 months, and have adequate hematologic, hepatic, and renal function (including 500 mg of proteinuria over 24 hours). Patients were excluded from enrollment if they had received prior chemotherapy for metastatic cancer (with the exception of adjuvant or radiosensitizing use of fluoropyrimidines combined with LV and/or levamisole that stopped at least 12 months before study enrollment), radiotherapy within the preceding 14 days, or any prior CRC biotherapy. Other exclusion criteria included known CNS metastases, major surgery within the preceding 28 days, clinically significant cardiovascular disease, clinically detectable ascites, pregnancy or lactation, chronic use of aspirin (> 325 mg/d) or other nonsteroidal anti-inflammatory drugs, preexisting bleeding diatheses, coagulopathy, or full-dose anticoagulation.
Ethics
All patients provided written informed consent for their study participation. The protocol was approved by the institutional review board of each investigational site and was conducted in accordance with the Declaration of Helsinki, current US Food and Drug Administration Good Clinical Practices, and local ethical and legal requirements.
Random Assignment and Treatment
Patients were randomly assigned 1:1:1 to three first-line treatment regimens (Table 1) using a dynamic randomization algorithm designed to attain balance overall, within each study center, and within each of the following classifications: baseline ECOG performance status (0 v 1), number of metastatic sites (1 v > 1), and primary disease site (colon v rectum). Study treatment continued until disease progression or unacceptable toxicity developed or for a maximum of 96 weeks. At the time of disease progression, the patient's treatment assignment was unblinded and second-line treatment could be offered. If the patient had been treated in one of the BV-containing arms, continuing treatment with BV was an option during second-line treatment. Patients assigned to IFL/placebo for first-line treatment were not allowed to cross-over to BV for second-line treatment.
The dose and schedule of IFL in the irinotecan-containing arms were based on the regimen described by Saltz et al.14 The dose and schedule of FU/LV used in the FU/LV/BV arm were based on the Roswell Park regimen described by Petrelli et al.15 Treatment assignment was double-blinded for the IFL/placebo (control) and IFL/BV arms, but was not blinded for the FU/LV/BV arm. The study protocol included an interim analysis by an unblinded, independent data monitoring committee (DMC), to be conducted after the random assignment of approximately 100 patients into each treatment arm, to determine whether the IFL/BV regimen was safe for continued study. If the DMC found no significant safety issues with the addition of BV to IFL, this study arm was to be fully enrolled and enrollment of patients into the FU/LV/BV arm was to stop. Reciprocally, if there had been excess toxicity in the IFL/BV arm, then enrollment to the IFL/BV arm would have been stopped and the FU/LV/BV arm would have been fully enrolled for comparison with IFL/placebo.
BV (or placebo) was administered concomitantly with chemotherapy. Patients received standard premedication for chemotherapy. If a patient's weight changed by 10% during the study, BV and chemotherapy doses were recalculated. Standard intra- and intercycle dose modifications of irinotecan and FU (as described in the package insert for irinotecan16) were permitted if grade 2 to grade 4 toxicities occurred (based on National Cancer Institute Common Toxicity Criteria, version 2). LV and BV doses were not altered.
Assessments
After initial screening, patients' tumor status was assessed by the investigator every 6 weeks for the first 24 weeks, then every 12 weeks for the remainder of therapy. Investigators assessed tumor responses and progression using the Response Evaluation Criteria in Solid Tumors.17 All complete and partial responses required confirmation at least 4 weeks later. All patients were followed for survival and subsequent disease treatment until death, loss to follow-up, or study termination.
To evaluate safety of the study treatments, adverse events, laboratory test results, and changes in vital signs were recorded and were graded using the National Cancer Institute Common Toxicity Criteria, version 2. All adverse events were collected preceding the interim safety analysis that determined which of the BV-containing chemotherapy arms would continue to enroll patients.
Statistical Analyses
The efficacy and safety analyses included all patients who were enrolled in the discontinued FU/LV/BV arm and all patients in the IFL/placebo arm who were enrolled concurrently with the patients in the FU/LV/BV arm.
The efficacy and safety analyses presented in this article were described as exploratory analyses in the study protocol. For patients alive at the time of analysis, duration of survival was measured from randomization to the last time of contact. Progression-free survival was defined as time from randomization until disease progression or death on study. Patients who did not have disease progression were censored at last tumor assessment, or day 0 if no postbaseline assessment had been performed. For the analysis of response, patients who had no adequate follow-up were categorized as nonresponders. The safety assessment data were not adjusted for the differential time on treatment.
RESULTS
At the time of the interim safety analysis, 313 patients had been randomly assigned: 100 patients into the IFL/placebo arm, 103 patients into the IFL/BV arm, and 110 patients into the FU/LV/BV arm. The DMC concluded that the IFL/BV combination had an acceptable safety profile and recommended that enrollment in the IFL/placebo and IFL/BV arms continue until approximately 400 patients per treatment arm were enrolled and, therefore, the FU/LV/BV arm was closed to further enrollment. Demographic and baseline disease characteristics for the concurrently enrolled patients in the IFL/placebo and FU/LV/BV arms, which comprise the study population for efficacy analyses, are summarized in Table 2. Major or minor eligibility deviations were similar across arms: 18.0% in the IFL/placebo arm and 19.1% in the FU/LV/BV arm. Of these randomly assigned patients, 98 in the IFL/placebo arm and 109 in the FU/LV/BV arm received at least one study treatment administration (population for safety analyses).
Results for the primary and secondary efficacy outcome measures for patients receiving either irinotecan or BV in combination with FU/LV are listed in Table 3. Patients who received BV combined with FU/LV had a median survival duration of 18.3 months compared with 15.1 months for IFL/placebo. The stratified hazard ratio for death in the FU/LV/BV arm compared with IFL/placebo was 0.82 (95% CI, 0.59 to 1.15; P = .25). The Kaplan-Meier curves for duration of survival are shown in Figure 1. Patients in the FU/LV/BV arm had a median duration of progression-free survival of 8.8 months, compared with 6.8 months in the IFL/placebo arm. The stratified hazard ratio for progression or death in the FU/LV/BV arm compared with the IFL/placebo arm was 0.86 (95% CI, 0.60 to 1.24; P = .42). Kaplan-Meier curves for progression-free survival are shown in Figure 2.
The objective response rate (complete and partial responders) during first-line therapy was 40.0% in the FU/LV/BV arm, compared with 37.0% in the IFL/placebo arm (P = .66). Complete responses were seen in 3.6% and 2.0% of patients in the FU/LV/BV and IFL/placebo arms, respectively. Median duration of objective response during first-line therapy was 8.5 months (95% CI, 6.2 to 10.8 months) for the FU/LV/BV arm and 7.2 months (95% CI, 5.4 to 11.0 months) for the IFL/placebo arm.
Use of second-line therapies in the FU/LV/BV group was 70% compared with 51% in the IFL/placebo group, primarily as a result of the greater use of irinotecan after progression on FU/LV/BV (Table 4). In the FU/LV/BV arm, 57.3% of patients received second-line irinotecan. Oxaliplatin was used in 14.5% of patients in the FU/LV/BV arm and 24.0% of patients in the IFL/placebo arm. The overall safety evaluation for patients in the FU/LV/BV arm and the concurrently enrolled patients in the IFL/placebo arm is summarized in Tables 5 and 6. All grade adverse events with rates 10% higher or lower in FU/LV/BV treatment arm compared with the IFL/placebo treatment arm are listed in Table 5, and additional targeted adverse events identified as potential BV-related toxicities in previous clinical trials (thromboembolic events, hemorrhage/bleeding, and proteinuria)11,18 are listed in Table 6. These rates are not adjusted for time on therapy, which was longer in the FU/LV/BV arm. GI perforation, an uncommon but previously unexpected BV-related toxicity, was not noted in the cohort of patients receiving FU/LV/BV at the time the safety analysis was performed.
The percentages of patients with any grade 3 to 4 event were 77.1% for FU/LV/BV and 81.6% for IFL/placebo. Several known BV-related toxicities were increased in the FU/LV/BV arm compared with the IFL/placebo arm, including grade 3 hemorrhage (1.0% for IFL/placebo v 6.4% for FU/LV/BV), grade 3 hypertension (3.1% for IFL/placebo v 18.3% for FU/LV/BV), and grade 3 proteinuria (0% for IFL/placebo v 1.8% for FU/LV/BV). Skin, taste, and lacrimation toxicity was reported more frequently in the FU/LV/BV group. Neutropenia, including grade 3 or 4, occurred less frequently in the FU/LV/BV group (54% for IFL/placebo v 11% for FU/LV/BV). Febrile neutropenia was reported in four patients in the IFL/placebo group and in three patients in the FU/LV/BV group. Relative to the IFL/placebo arm, a higher percentage of patients in the FU/LV/BV arm had an adverse event causing withdrawal from the study (6.1% v 10.1%), but not an adverse event leading to hospitalization (40.8% v 40.4%). More FU/LV/BV patients died from any cause within 60 days after random assignment (3.0% v 5.5%). Causes of 60-day mortality were as follows: death from progressive disease (three patients, FU/LV/BV; one patient, IFL/placebo); septicemia (one patient, IFL/placebo); pulmonary edema (one patient, FU/LV/BV); probable pulmonary embolism (one patient, IFL/placebo); multisystem organ failure (one patient, FU/LV/BV); and death of unknown cause (one patient, FU/LV/BV).
DISCUSSION
The addition of BV, a novel antiangiogenic therapeutic, to the IFL regimen in first-line treatment of metastatic CRC has been reported to improve overall survival, progression-free survival, objective response rate, and duration of response compared with IFL alone.12 We now report further data from this study demonstrating that the combination of FU/LV plus BV without irinotecan is also an active regimen and seemed to be at least as effective as irinotecan plus FU/LV for treatment of metastatic CRC. These data represent the only randomized comparison of FU/LV/BV versus IFL alone. Although this treatment comparison was not the primary end point of the pivotal study, it represents a planned analysis of concurrently randomly assigned patients with prospectively collected efficacy and safety data. The trends for improved efficacy for FU/LV/BV compared with IFL were not statistically significant; however, the activity of FU/LV/BV in this moderately sized study is noteworthy for several reasons. First, the data are consistent with those of a prior randomized phase II study, which demonstrated improved outcomes with the addition of BV to FU/LV.11 These data are also consistent with a more recent, 200-patient randomized phase II study of FU/LV with or without BV in patients with CRC who are not candidates for irinotecan, where improved response rates and time to progression were also seen with the addition of BV.19
The notable tolerability and activity of FU/LV/BV suggest that this regimen may have value in the treatment of previously untreated patients with CRC, including those who cannot tolerate more aggressive therapies. The increase in taste and lacrimation toxicity reported in this trial was not seen in the randomized phase II study with FU/LV and BV.19 The reason for the imbalance in 60-day mortality rate in the FU/LV/BV arm (5.5%) versus the IFL/placebo arm (3.0%) is not readily apparent but is likely due to chance because of small patient numbers. In the pivotal study, 60-day mortality rates were not different: 4.9% for IFL/placebo (n = 397) versus 3.0% for IFL/BV (n = 393).12 Similarly, in a study focused on patients who were not candidates for IFL, mostly because of older age or lesser performance status, the 60-day all cause mortality was 13.5% for FU/LV/placebo (n = 104) and 5.0% for FU/LV/BV (n = 100). These data do, however, emphasize the toxicity of FU/LV, particularly when given in bolus rather than infusional regimens.20
The treatment of metastatic CRC is evolving quickly. After decades with only one active chemotherapy agent (FU), five new active drugs (irinotecan, capecitabine, oxaliplatin, cetuximab, and BV) have recently become available. At the time our trial was designed and initiated, the IFL regimen had just been approved by the US Food and Drug Administration and was accepted as standard of care for first-line treatment of metastatic CRC in the United States. However, toxicities associated with irinotecan-based therapy can be significant compared with those with FU/LV and include increased rates of diarrhea, nausea and vomiting, neutropenia, and asthenia.4,5 Moreover, although the addition of irinotecan to FU/LV improved survival by a median of 2.2 months, this benefit was not seen for several common subgroups, including patients older than 65 years and those with an ECOG performance status greater than zero, a history of prior adjuvant therapy, or an elevated lactate dehydrogenase level.5 Importantly, BV therapy conferred clinical benefit for these subgroups when added to IFL.12
The activity seen with FU/LV/BV in this study cohort was notable across all efficacy end points, namely survival, progression-free survival, and response rate, and these results are consistent with those from other studies of BV plus FU/LV in CRC. Use of active second-line chemotherapy was greater in the FU/LV/BV arm compared with IFL/placebo, which may confound the survival analysis, as has been shown in other survival studies in CRC.21 Patients enrolled on this study, including those on the FU/LV/BV arm, were all candidates for irinotecan therapy. When patients in the FU/LV/BV arm experienced disease progression, they were more likely to receive irinotecan. Use of oxaliplatin was similar in the two arms. The FU/LV/BV regimen was generally well tolerated, and the adverse effect profile was consistent with the expected toxicities for bolus FU/LV alone and for BV, as reported in other trials.11,12,19
In conclusion, the efficacy and toxicity data from the FU/LV/BV group of the pivotal study of BV in CRC confirm that FU/LV/BV is an active and generally well tolerated first-line regimen for the treatment of metastatic CRC. These data confirm the activity of BV in CRC beyond combination with only IFL, suggesting that anti-VEGF therapy may be a broadly applicable approach for the treatment of metastatic CRC. This regimen represents an important therapeutic option for these patients.
Authors' Disclosures of Potential Conflicts of Interest
The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Employment: Eric Holmgren, Genentech; Julie Hambleton, Genentech; William F. Novotny, Genentech. Consultant/Advisory Role: Herbert I. Hurwitz, Genentech, Pfizer; William Heim, Genentech; Jordan Berlin, Genentech, Pfizer. Stock Ownership: William Heim, Genentech; Eric Holmgren, Genentech; Julie Hambleton, Genentech; William F. Novotny, Genentech. Honoraria: John D. Hainsworth, Genentech; William Heim, Genentech; Jordan Berlin, Genentech; Fairooz Kabbinavar, Genentech. Research Funding: Herbert I. Hurwitz, AstraZeneca, Bristol-Myers Squibb, Cephalon, Genentech, GlaxoSmithKline, Pfizer, Sanofi-Synthelabo. Other Remuneration: William Heim, Genentech. For a detailed description of these categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and Disclosures of Potential Conflicts of Interest found in Information for Contributors in the front of each issue.
Acknowledgment
We thank the patients who participated in this study and their families and support teams, the health care teams at each center, the Data Safety Monitoring Board, and the Data Coordinating Center (Statistics Collaborative) for their invaluable assistance in the conduct of this study. We thank Linda Phillips for her assistance in the preparation of this manuscript and Somnath Sarkar and Robert Mass for their critical review.
NOTES
H.I.H. was supported in part by career development grant No. K23 CA085582–04 from the National Cancer Institute, National Institutes of Health, Bethesda, MD.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Thirion P, Michiels S, Pignon JP, et al: Meta-Analysis Group in Cancer: Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer: An updated meta-analysis. J Clin Oncol 22:3766-3775, 2004 [erratum in J Clin Oncol 23:1337-1338, 2005]
Moertel CG: Chemotherapy for colorectal cancer. N Engl J Med 330:1136-1142, 1994
Thirion P, Michiels S, Pignon JP, et al: Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer: An updated meta-analysis. J Clin Oncol 22:3766-3775, 2004
Douillard JY, Cunningham D, Roth AD, et al: Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: A multicentre randomised trial. Lancet 355:1041-1047, 2000
Saltz LB, Cox JV, Blanke C, et al: Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 343:905-914, 2000
Freyer G, Rougier P, Bugat R, et al: Prognostic factors for tumour response, progression-free survival and toxicity in metastatic colorectal cancer patients given irinotecan (CPT-11) as second-line chemotherapy after 5FU failure: CPT-11 F205, F220, F221 and V222 study groups. Br J Cancer 83:431-437, 2000
Presta LG, Chen H, O'Connor SJ, et al: Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer Res 57:4593-4599, 1997
Folkman J: Tumor angiogenesis: Therapeutic implications. N Engl J Med 285:1182-1186, 1971
Ferrara N, Davis-Smyth T: The biology of vascular endothelial growth factor. Endocr Rev 18:4-25, 1997
Folkman J: Antiangiogenic therapy, in DeVita VTJ, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology. Philadelphia, PA, Lippincott-Raven, 1997, pp3075-3085
Kabbinavar F, Hurwitz HI, Fehrenbacher L, et al: Phase II, randomized trial comparing bevacizumab plus fluorouracil (FU)/leucovorin (LV) with FU/LV alone in patients with metastatic colorectal cancer. J Clin Oncol 21:60-65, 2003
Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
Oken MM, Creech RH, Tormey DC, et al: Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 5:649-655, 1982
Saltz LB, Kanowitz J, Kemeny NE, et al: Phase I clinical and pharmacokinetic study of irinotecan, fluorouracil, and leucovorin in patients with advanced solid tumors. J Clin Oncol 14:2959-2967, 1996
Petrelli N, Douglass HO Jr, Herrera L, et al: The modulation of fluorouracil with leucovorin in metastatic colorectal carcinoma: A prospective randomized phase III trial—Gastrointestinal Tumor Study Group. J Clin Oncol 7:1419-1426, 1989
Camptosar [package insert]. Kalamazoo, MI, Pharmacia & Upjohn Company, 2002
Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205-216, 2000
Yang JC, Haworth L, Sherry RM, et al: A randomized trial of bevacizumab, an anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N Engl J Med 349:427-434, 2003
Kabbinavar FF, Schulz J, McCleod M, et al: Addition of bevacizumab to bolus fluorouracil and leucovorin in first-line metastatic colorectal cancer: Results of a randomized phase II trial. J Clin Oncol 23 (in press)
Meta-Analysis Group in Cancer: Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol 16:301-308, 1998
Goldberg RM, Sargent DJ, Morton RF, et al: A randomized controlled trial of fluorouracil plus leucovorin, irinotecan, and oxaliplatin combinations in patients with previously untreated metastatic colorectal cancer. J Clin Oncol 22:23-30, 2004(Herbert I. Hurwitz, Louis)
Kaiser Permanente Medical Care Program, Northern California
Genentech Inc, South San Francisco
Department of Medicine, University of California at Los Angeles, Los Angeles, CA
The Sarah Cannon Cancer Center, Nashville, TN
Hematology and Cancer Center of Northeastern Pennsylvania, Dunmore, PA
Division of Medical Oncology, Vanderbilt University Medical Center, Nashville, TN
ABSTRACT
PURPOSE: In a phase III trial, combining bevacizumab (BV)—a recombinant, humanized, monoclonal antibody targeting vascular endothelial growth factor—with irinotecan, bolus fluorouracil (FU), and leucovorin (LV; IFL) increased survival compared with IFL alone in first-line treatment of patients with metastatic colorectal cancer (CRC). Results for the parent study of IFL/BV versus IFL/placebo are reported elsewhere. Here, we describe efficacy and safety results for the third patient cohort in this trial, who received BV combined with FU/LV, and compare them with results for concurrently enrolled patients who received IFL.
METHODS: Patients (N = 923) were randomly assigned to receive IFL/placebo (control), IFL/BV, or FU/LV/BV. Bevacizumab (Avastin; Genentech Inc, South San Francisco, CA) 5 mg/kg was administered intravenously every 2 weeks. Before an interim analysis confirmed acceptable safety for IFL/BV, 313 patients were concurrently randomly assigned to these three arms; after this analysis, the FU/LV/BV arm was discontinued.
RESULTS: Median overall survivals were 18.3 and 15.1 months with FU/LV/BV (n = 110) and IFL/placebo (n = 100), respectively. Median progression-free survivals were 8.8 and 6.8 months, respectively. Overall response rates were 40.0% and 37.0%, and median response durations were 8.5 and 7.2 months, respectively. Adverse events consistent with those expected from FU/leucovorin- or IFL-based regimens were seen, as were modest increases in hypertension and bleeding in the bevacizumab arm, which were generally easily managed.
CONCLUSION: The FU/LV/BV regimen seems as effective as IFL and has an acceptable safety profile. FU/LV/BV is an active alternative treatment regimen for patients with previously untreated metastatic CRC.
INTRODUCTION
Patients with metastatic colorectal cancer (CRC) have a poor prognosis. Until recently, the standard treatment for metastatic CRC in the United States was chemotherapy with fluorouracil (FU) plus leucovorin (LV), a biochemical modulator of FU.1,2 The combination of FU/LV improves response rates among patients with metastatic CRC compared with FU alone.3 Addition of irinotecan to FU/LV (combination IFL therapy) has been shown to increase survival, time to disease progression, and response rates in two randomized clinical trials, each with approximately 400 patients.4,5 However, some patients are not ideal candidates for irinotecan-containing regimens or do not tolerate these regimens, which are commonly associated with significant myelosuppression, asthenia, and GI toxicities.4-6 Therefore, alternative combination therapies that are active and better tolerated are needed.
Bevacizumab (BV; Avastin; rhuMAb vascular endothelial growth factor [VEGF]; Genentech Inc, South San Francisco, CA) is a recombinant humanized monoclonal antibody that blocks the activity of VEGF.7 VEGF is a central mediator of tumor-related angiogenesis, and its expression has been shown to be prognostic for a number of solid tumors.8-10 Results of an initial randomized phase II trial in 104 patients with metastatic CRC suggested that adding BV to FU/LV increased clinical benefit relative to FU/LV alone, with an acceptable toxicity profile.11 Subsequently, a large, phase III, multicenter, double-blind, randomized trial in patients with metastatic CRC demonstrated that the addition of BV to first-line IFL chemotherapy (n = 402) prolonged survival by 4.7 months compared with IFL alone (n = 411; P < .001).12 Significant benefit was also demonstrated for all secondary efficacy end points: progression-free survival increased from 6.2 months to 10.6 months (P < .001), overall response rate increased from 34.8% to 44.8% (P = .004), and the duration of response increased from 7.1 months to 10.4 months (P = .001).
Because no safety data on the IFL/BV combination were available at the time the phase III trial began, the study design initially included a third treatment arm in which patients received BV in combination with FU and LV (FU/LV/BV arm). A planned interim safety analysis was performed for the purpose of selecting one BV-containing arm for comparison with IFL. If there had been excess toxicity with IFL/BV, that arm would have been discontinued. However, after the planned interim safety analysis revealed an acceptable safety profile for the IFL/BV combination, enrollment onto the FU/LV/BV treatment arm stopped to allow full enrollment in the two primary study arms (ie, IFL/placebo and IFL/BV).
Because some patients do not tolerate more aggressive chemotherapy, the combination of BV and FU/LV remains a potentially useful therapeutic alternative. In this article, we present the results of an analysis of data from the FU/LV/BV arm of the phase III trial and for concurrently enrolled patients in the primary study arm that was treated with irinotecan/FU/LV and placebo. This is the only randomized study to date directly comparing the addition of a traditional cytotoxic agent (irinotecan) with the addition of a targeted antiangiogenic agent (BV) to a FU/LV-based chemotherapy regimen.
METHODS
Patients
Patients eligible to participate in the phase III study had histologically confirmed, bidimensionally measurable, metastatic CRC. Patients had to be at least 18 years old, have an Eastern Cooperative Oncology Group (ECOG) performance status13 of 0 or 1, have a life expectancy greater than 3 months, and have adequate hematologic, hepatic, and renal function (including 500 mg of proteinuria over 24 hours). Patients were excluded from enrollment if they had received prior chemotherapy for metastatic cancer (with the exception of adjuvant or radiosensitizing use of fluoropyrimidines combined with LV and/or levamisole that stopped at least 12 months before study enrollment), radiotherapy within the preceding 14 days, or any prior CRC biotherapy. Other exclusion criteria included known CNS metastases, major surgery within the preceding 28 days, clinically significant cardiovascular disease, clinically detectable ascites, pregnancy or lactation, chronic use of aspirin (> 325 mg/d) or other nonsteroidal anti-inflammatory drugs, preexisting bleeding diatheses, coagulopathy, or full-dose anticoagulation.
Ethics
All patients provided written informed consent for their study participation. The protocol was approved by the institutional review board of each investigational site and was conducted in accordance with the Declaration of Helsinki, current US Food and Drug Administration Good Clinical Practices, and local ethical and legal requirements.
Random Assignment and Treatment
Patients were randomly assigned 1:1:1 to three first-line treatment regimens (Table 1) using a dynamic randomization algorithm designed to attain balance overall, within each study center, and within each of the following classifications: baseline ECOG performance status (0 v 1), number of metastatic sites (1 v > 1), and primary disease site (colon v rectum). Study treatment continued until disease progression or unacceptable toxicity developed or for a maximum of 96 weeks. At the time of disease progression, the patient's treatment assignment was unblinded and second-line treatment could be offered. If the patient had been treated in one of the BV-containing arms, continuing treatment with BV was an option during second-line treatment. Patients assigned to IFL/placebo for first-line treatment were not allowed to cross-over to BV for second-line treatment.
The dose and schedule of IFL in the irinotecan-containing arms were based on the regimen described by Saltz et al.14 The dose and schedule of FU/LV used in the FU/LV/BV arm were based on the Roswell Park regimen described by Petrelli et al.15 Treatment assignment was double-blinded for the IFL/placebo (control) and IFL/BV arms, but was not blinded for the FU/LV/BV arm. The study protocol included an interim analysis by an unblinded, independent data monitoring committee (DMC), to be conducted after the random assignment of approximately 100 patients into each treatment arm, to determine whether the IFL/BV regimen was safe for continued study. If the DMC found no significant safety issues with the addition of BV to IFL, this study arm was to be fully enrolled and enrollment of patients into the FU/LV/BV arm was to stop. Reciprocally, if there had been excess toxicity in the IFL/BV arm, then enrollment to the IFL/BV arm would have been stopped and the FU/LV/BV arm would have been fully enrolled for comparison with IFL/placebo.
BV (or placebo) was administered concomitantly with chemotherapy. Patients received standard premedication for chemotherapy. If a patient's weight changed by 10% during the study, BV and chemotherapy doses were recalculated. Standard intra- and intercycle dose modifications of irinotecan and FU (as described in the package insert for irinotecan16) were permitted if grade 2 to grade 4 toxicities occurred (based on National Cancer Institute Common Toxicity Criteria, version 2). LV and BV doses were not altered.
Assessments
After initial screening, patients' tumor status was assessed by the investigator every 6 weeks for the first 24 weeks, then every 12 weeks for the remainder of therapy. Investigators assessed tumor responses and progression using the Response Evaluation Criteria in Solid Tumors.17 All complete and partial responses required confirmation at least 4 weeks later. All patients were followed for survival and subsequent disease treatment until death, loss to follow-up, or study termination.
To evaluate safety of the study treatments, adverse events, laboratory test results, and changes in vital signs were recorded and were graded using the National Cancer Institute Common Toxicity Criteria, version 2. All adverse events were collected preceding the interim safety analysis that determined which of the BV-containing chemotherapy arms would continue to enroll patients.
Statistical Analyses
The efficacy and safety analyses included all patients who were enrolled in the discontinued FU/LV/BV arm and all patients in the IFL/placebo arm who were enrolled concurrently with the patients in the FU/LV/BV arm.
The efficacy and safety analyses presented in this article were described as exploratory analyses in the study protocol. For patients alive at the time of analysis, duration of survival was measured from randomization to the last time of contact. Progression-free survival was defined as time from randomization until disease progression or death on study. Patients who did not have disease progression were censored at last tumor assessment, or day 0 if no postbaseline assessment had been performed. For the analysis of response, patients who had no adequate follow-up were categorized as nonresponders. The safety assessment data were not adjusted for the differential time on treatment.
RESULTS
At the time of the interim safety analysis, 313 patients had been randomly assigned: 100 patients into the IFL/placebo arm, 103 patients into the IFL/BV arm, and 110 patients into the FU/LV/BV arm. The DMC concluded that the IFL/BV combination had an acceptable safety profile and recommended that enrollment in the IFL/placebo and IFL/BV arms continue until approximately 400 patients per treatment arm were enrolled and, therefore, the FU/LV/BV arm was closed to further enrollment. Demographic and baseline disease characteristics for the concurrently enrolled patients in the IFL/placebo and FU/LV/BV arms, which comprise the study population for efficacy analyses, are summarized in Table 2. Major or minor eligibility deviations were similar across arms: 18.0% in the IFL/placebo arm and 19.1% in the FU/LV/BV arm. Of these randomly assigned patients, 98 in the IFL/placebo arm and 109 in the FU/LV/BV arm received at least one study treatment administration (population for safety analyses).
Results for the primary and secondary efficacy outcome measures for patients receiving either irinotecan or BV in combination with FU/LV are listed in Table 3. Patients who received BV combined with FU/LV had a median survival duration of 18.3 months compared with 15.1 months for IFL/placebo. The stratified hazard ratio for death in the FU/LV/BV arm compared with IFL/placebo was 0.82 (95% CI, 0.59 to 1.15; P = .25). The Kaplan-Meier curves for duration of survival are shown in Figure 1. Patients in the FU/LV/BV arm had a median duration of progression-free survival of 8.8 months, compared with 6.8 months in the IFL/placebo arm. The stratified hazard ratio for progression or death in the FU/LV/BV arm compared with the IFL/placebo arm was 0.86 (95% CI, 0.60 to 1.24; P = .42). Kaplan-Meier curves for progression-free survival are shown in Figure 2.
The objective response rate (complete and partial responders) during first-line therapy was 40.0% in the FU/LV/BV arm, compared with 37.0% in the IFL/placebo arm (P = .66). Complete responses were seen in 3.6% and 2.0% of patients in the FU/LV/BV and IFL/placebo arms, respectively. Median duration of objective response during first-line therapy was 8.5 months (95% CI, 6.2 to 10.8 months) for the FU/LV/BV arm and 7.2 months (95% CI, 5.4 to 11.0 months) for the IFL/placebo arm.
Use of second-line therapies in the FU/LV/BV group was 70% compared with 51% in the IFL/placebo group, primarily as a result of the greater use of irinotecan after progression on FU/LV/BV (Table 4). In the FU/LV/BV arm, 57.3% of patients received second-line irinotecan. Oxaliplatin was used in 14.5% of patients in the FU/LV/BV arm and 24.0% of patients in the IFL/placebo arm. The overall safety evaluation for patients in the FU/LV/BV arm and the concurrently enrolled patients in the IFL/placebo arm is summarized in Tables 5 and 6. All grade adverse events with rates 10% higher or lower in FU/LV/BV treatment arm compared with the IFL/placebo treatment arm are listed in Table 5, and additional targeted adverse events identified as potential BV-related toxicities in previous clinical trials (thromboembolic events, hemorrhage/bleeding, and proteinuria)11,18 are listed in Table 6. These rates are not adjusted for time on therapy, which was longer in the FU/LV/BV arm. GI perforation, an uncommon but previously unexpected BV-related toxicity, was not noted in the cohort of patients receiving FU/LV/BV at the time the safety analysis was performed.
The percentages of patients with any grade 3 to 4 event were 77.1% for FU/LV/BV and 81.6% for IFL/placebo. Several known BV-related toxicities were increased in the FU/LV/BV arm compared with the IFL/placebo arm, including grade 3 hemorrhage (1.0% for IFL/placebo v 6.4% for FU/LV/BV), grade 3 hypertension (3.1% for IFL/placebo v 18.3% for FU/LV/BV), and grade 3 proteinuria (0% for IFL/placebo v 1.8% for FU/LV/BV). Skin, taste, and lacrimation toxicity was reported more frequently in the FU/LV/BV group. Neutropenia, including grade 3 or 4, occurred less frequently in the FU/LV/BV group (54% for IFL/placebo v 11% for FU/LV/BV). Febrile neutropenia was reported in four patients in the IFL/placebo group and in three patients in the FU/LV/BV group. Relative to the IFL/placebo arm, a higher percentage of patients in the FU/LV/BV arm had an adverse event causing withdrawal from the study (6.1% v 10.1%), but not an adverse event leading to hospitalization (40.8% v 40.4%). More FU/LV/BV patients died from any cause within 60 days after random assignment (3.0% v 5.5%). Causes of 60-day mortality were as follows: death from progressive disease (three patients, FU/LV/BV; one patient, IFL/placebo); septicemia (one patient, IFL/placebo); pulmonary edema (one patient, FU/LV/BV); probable pulmonary embolism (one patient, IFL/placebo); multisystem organ failure (one patient, FU/LV/BV); and death of unknown cause (one patient, FU/LV/BV).
DISCUSSION
The addition of BV, a novel antiangiogenic therapeutic, to the IFL regimen in first-line treatment of metastatic CRC has been reported to improve overall survival, progression-free survival, objective response rate, and duration of response compared with IFL alone.12 We now report further data from this study demonstrating that the combination of FU/LV plus BV without irinotecan is also an active regimen and seemed to be at least as effective as irinotecan plus FU/LV for treatment of metastatic CRC. These data represent the only randomized comparison of FU/LV/BV versus IFL alone. Although this treatment comparison was not the primary end point of the pivotal study, it represents a planned analysis of concurrently randomly assigned patients with prospectively collected efficacy and safety data. The trends for improved efficacy for FU/LV/BV compared with IFL were not statistically significant; however, the activity of FU/LV/BV in this moderately sized study is noteworthy for several reasons. First, the data are consistent with those of a prior randomized phase II study, which demonstrated improved outcomes with the addition of BV to FU/LV.11 These data are also consistent with a more recent, 200-patient randomized phase II study of FU/LV with or without BV in patients with CRC who are not candidates for irinotecan, where improved response rates and time to progression were also seen with the addition of BV.19
The notable tolerability and activity of FU/LV/BV suggest that this regimen may have value in the treatment of previously untreated patients with CRC, including those who cannot tolerate more aggressive therapies. The increase in taste and lacrimation toxicity reported in this trial was not seen in the randomized phase II study with FU/LV and BV.19 The reason for the imbalance in 60-day mortality rate in the FU/LV/BV arm (5.5%) versus the IFL/placebo arm (3.0%) is not readily apparent but is likely due to chance because of small patient numbers. In the pivotal study, 60-day mortality rates were not different: 4.9% for IFL/placebo (n = 397) versus 3.0% for IFL/BV (n = 393).12 Similarly, in a study focused on patients who were not candidates for IFL, mostly because of older age or lesser performance status, the 60-day all cause mortality was 13.5% for FU/LV/placebo (n = 104) and 5.0% for FU/LV/BV (n = 100). These data do, however, emphasize the toxicity of FU/LV, particularly when given in bolus rather than infusional regimens.20
The treatment of metastatic CRC is evolving quickly. After decades with only one active chemotherapy agent (FU), five new active drugs (irinotecan, capecitabine, oxaliplatin, cetuximab, and BV) have recently become available. At the time our trial was designed and initiated, the IFL regimen had just been approved by the US Food and Drug Administration and was accepted as standard of care for first-line treatment of metastatic CRC in the United States. However, toxicities associated with irinotecan-based therapy can be significant compared with those with FU/LV and include increased rates of diarrhea, nausea and vomiting, neutropenia, and asthenia.4,5 Moreover, although the addition of irinotecan to FU/LV improved survival by a median of 2.2 months, this benefit was not seen for several common subgroups, including patients older than 65 years and those with an ECOG performance status greater than zero, a history of prior adjuvant therapy, or an elevated lactate dehydrogenase level.5 Importantly, BV therapy conferred clinical benefit for these subgroups when added to IFL.12
The activity seen with FU/LV/BV in this study cohort was notable across all efficacy end points, namely survival, progression-free survival, and response rate, and these results are consistent with those from other studies of BV plus FU/LV in CRC. Use of active second-line chemotherapy was greater in the FU/LV/BV arm compared with IFL/placebo, which may confound the survival analysis, as has been shown in other survival studies in CRC.21 Patients enrolled on this study, including those on the FU/LV/BV arm, were all candidates for irinotecan therapy. When patients in the FU/LV/BV arm experienced disease progression, they were more likely to receive irinotecan. Use of oxaliplatin was similar in the two arms. The FU/LV/BV regimen was generally well tolerated, and the adverse effect profile was consistent with the expected toxicities for bolus FU/LV alone and for BV, as reported in other trials.11,12,19
In conclusion, the efficacy and toxicity data from the FU/LV/BV group of the pivotal study of BV in CRC confirm that FU/LV/BV is an active and generally well tolerated first-line regimen for the treatment of metastatic CRC. These data confirm the activity of BV in CRC beyond combination with only IFL, suggesting that anti-VEGF therapy may be a broadly applicable approach for the treatment of metastatic CRC. This regimen represents an important therapeutic option for these patients.
Authors' Disclosures of Potential Conflicts of Interest
The following authors or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Employment: Eric Holmgren, Genentech; Julie Hambleton, Genentech; William F. Novotny, Genentech. Consultant/Advisory Role: Herbert I. Hurwitz, Genentech, Pfizer; William Heim, Genentech; Jordan Berlin, Genentech, Pfizer. Stock Ownership: William Heim, Genentech; Eric Holmgren, Genentech; Julie Hambleton, Genentech; William F. Novotny, Genentech. Honoraria: John D. Hainsworth, Genentech; William Heim, Genentech; Jordan Berlin, Genentech; Fairooz Kabbinavar, Genentech. Research Funding: Herbert I. Hurwitz, AstraZeneca, Bristol-Myers Squibb, Cephalon, Genentech, GlaxoSmithKline, Pfizer, Sanofi-Synthelabo. Other Remuneration: William Heim, Genentech. For a detailed description of these categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and Disclosures of Potential Conflicts of Interest found in Information for Contributors in the front of each issue.
Acknowledgment
We thank the patients who participated in this study and their families and support teams, the health care teams at each center, the Data Safety Monitoring Board, and the Data Coordinating Center (Statistics Collaborative) for their invaluable assistance in the conduct of this study. We thank Linda Phillips for her assistance in the preparation of this manuscript and Somnath Sarkar and Robert Mass for their critical review.
NOTES
H.I.H. was supported in part by career development grant No. K23 CA085582–04 from the National Cancer Institute, National Institutes of Health, Bethesda, MD.
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Thirion P, Michiels S, Pignon JP, et al: Meta-Analysis Group in Cancer: Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer: An updated meta-analysis. J Clin Oncol 22:3766-3775, 2004 [erratum in J Clin Oncol 23:1337-1338, 2005]
Moertel CG: Chemotherapy for colorectal cancer. N Engl J Med 330:1136-1142, 1994
Thirion P, Michiels S, Pignon JP, et al: Modulation of fluorouracil by leucovorin in patients with advanced colorectal cancer: An updated meta-analysis. J Clin Oncol 22:3766-3775, 2004
Douillard JY, Cunningham D, Roth AD, et al: Irinotecan combined with fluorouracil compared with fluorouracil alone as first-line treatment for metastatic colorectal cancer: A multicentre randomised trial. Lancet 355:1041-1047, 2000
Saltz LB, Cox JV, Blanke C, et al: Irinotecan plus fluorouracil and leucovorin for metastatic colorectal cancer. Irinotecan Study Group. N Engl J Med 343:905-914, 2000
Freyer G, Rougier P, Bugat R, et al: Prognostic factors for tumour response, progression-free survival and toxicity in metastatic colorectal cancer patients given irinotecan (CPT-11) as second-line chemotherapy after 5FU failure: CPT-11 F205, F220, F221 and V222 study groups. Br J Cancer 83:431-437, 2000
Presta LG, Chen H, O'Connor SJ, et al: Humanization of an anti-vascular endothelial growth factor monoclonal antibody for the therapy of solid tumors and other disorders. Cancer Res 57:4593-4599, 1997
Folkman J: Tumor angiogenesis: Therapeutic implications. N Engl J Med 285:1182-1186, 1971
Ferrara N, Davis-Smyth T: The biology of vascular endothelial growth factor. Endocr Rev 18:4-25, 1997
Folkman J: Antiangiogenic therapy, in DeVita VTJ, Hellman S, Rosenberg SA (eds): Cancer: Principles and Practice of Oncology. Philadelphia, PA, Lippincott-Raven, 1997, pp3075-3085
Kabbinavar F, Hurwitz HI, Fehrenbacher L, et al: Phase II, randomized trial comparing bevacizumab plus fluorouracil (FU)/leucovorin (LV) with FU/LV alone in patients with metastatic colorectal cancer. J Clin Oncol 21:60-65, 2003
Hurwitz H, Fehrenbacher L, Novotny W, et al: Bevacizumab plus irinotecan, fluorouracil, and leucovorin for metastatic colorectal cancer. N Engl J Med 350:2335-2342, 2004
Oken MM, Creech RH, Tormey DC, et al: Toxicity and response criteria of the Eastern Cooperative Oncology Group. Am J Clin Oncol 5:649-655, 1982
Saltz LB, Kanowitz J, Kemeny NE, et al: Phase I clinical and pharmacokinetic study of irinotecan, fluorouracil, and leucovorin in patients with advanced solid tumors. J Clin Oncol 14:2959-2967, 1996
Petrelli N, Douglass HO Jr, Herrera L, et al: The modulation of fluorouracil with leucovorin in metastatic colorectal carcinoma: A prospective randomized phase III trial—Gastrointestinal Tumor Study Group. J Clin Oncol 7:1419-1426, 1989
Camptosar [package insert]. Kalamazoo, MI, Pharmacia & Upjohn Company, 2002
Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors: European Organization for Research and Treatment of Cancer, National Cancer Institute of the United States, National Cancer Institute of Canada. J Natl Cancer Inst 92:205-216, 2000
Yang JC, Haworth L, Sherry RM, et al: A randomized trial of bevacizumab, an anti-vascular endothelial growth factor antibody, for metastatic renal cancer. N Engl J Med 349:427-434, 2003
Kabbinavar FF, Schulz J, McCleod M, et al: Addition of bevacizumab to bolus fluorouracil and leucovorin in first-line metastatic colorectal cancer: Results of a randomized phase II trial. J Clin Oncol 23 (in press)
Meta-Analysis Group in Cancer: Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer. J Clin Oncol 16:301-308, 1998
Goldberg RM, Sargent DJ, Morton RF, et al: A randomized controlled trial of fluorouracil plus leucovorin, irinotecan, and oxaliplatin combinations in patients with previously untreated metastatic colorectal cancer. J Clin Oncol 22:23-30, 2004(Herbert I. Hurwitz, Louis)