|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2002 American Society for Clinical Oncology Phase III Study of Gemcitabine in Combination With Fluorouracil Versus Gemcitabine Alone in Patients With Advanced Pancreatic Carcinoma: Eastern Cooperative Oncology Group Trial E2297ByFrom Vanderbilt University, Nashville, TN; Dana-Farber Cancer Institute, Boston, MA; the University of Wisconsin, Madison, WI; Illinois Oncology Research Association, Peoria, and Northwestern University, Chicago, IL; and the University of Pennsylvania, Philadelphia, PA. Address reprint requests to Jordan D. Berlin, MD, 777 Preston Research Building, Nashville, TN 37232-6307; email: jordan.berlin{at}mcmail.vanderbilt.edu
PURPOSE: Gemcitabine is generally considered to constitute first-line therapy for pancreatic cancer. To determine whether the addition of fluorouracil (5-FU) improves on the results from single-agent gemcitabine, the Eastern Cooperative Oncology Group (ECOG) compared gemcitabine plus bolus 5-FU with gemcitabine alone for patients with advanced pancreatic carcinoma. PATIENTS AND METHODS: This trial involved patients with biopsy-proven, advanced carcinoma of the pancreas not amenable to surgical resection. Patients were randomized to receive either gemcitabine alone (1,000 mg/m2/wk) weekly for 3 weeks of every 4 or to receive gemcitabine (1,000 mg/m2/wk) followed by 5-FU (600 mg/m2/wk) weekly on the same schedule. The primary end point of the trial was survival, with secondary end points of time to progression and response rate. RESULTS: Of 327 patients enrolled over 18 months, 322 were eligible. Overall, the median survival was 5.4 months for gemcitabine alone and 6.7 months for gemcitabine plus 5-FU (P = .09). Progression-free survival for gemcitabine alone was 2.2 months, compared with 3.4 months for gemcitabine plus 5-FU (P = .022). Objective responses were uncommon and were observed in only 5.6% of patients treated with gemcitabine and 6.9% of patients treated with gemcitabine plus 5-FU. Most toxicities were hematologic or gastrointestinal; no significant differences were noted between the two treatment arms. CONCLUSION: 5-FU, administered in conjunction with gemcitabine, did not improve the median survival of patients with advanced pancreatic carcinoma compared with single-agent gemcitabine. Further studies with other combinations of gemcitabine and 5-FU are not compelling, and clinical trial resources should address other combinations and novel agents.
PANCREAS CANCER IS the fourth leading cause of cancer deaths in the United States and has the poorest survival of the common malignancies.1 Despite advances in the treatment of other malignancies, improvements in the therapy of pancreas cancer have been few. In advanced disease, the greatest change has been acceptance of gemcitabine as first-line therapy on the basis of the results of a randomized trial comparing gemcitabine with bolus fluorouracil (5-FU).2 Patients enrolled onto this trial had to be symptomatic, with cancer-related pain and a Karnofsky performance status of 50% to 80%. After enrollment, patients underwent a period of 2 to 7 days during which symptoms were to be stabilized before randomization. Of 160 patients enrolled, 34 patients were not eligible to be randomized. In the 126 remaining patients, gemcitabine was superior to 5-FU in terms of median survival (5.6 v 4.4 months, respectively) and 1-year survival (18% v 2%, respectively). In addition, gemcitabine resulted in a superior clinical benefit response—a new symptom-related end point—than did 5-FU (23.8% v 4.8%, respectively). However, gemcitabine has limited measurable antitumor efficacy, with objective response rates of less than 10% and a median survival of less than 6 months.2,3 Combinations of gemcitabine with 5-FU have been evaluated to improve on the results obtained with gemcitabine alone. Two laboratory analyses using fluoropyrimidines in combination with gemcitabine were conducted in HT-29 colon cancer cells.4,5 Whereas one trial studied a sequence of 5-FU before gemcitabine, the second study analyzed the reverse sequence with floxuridine. Both trials suggested at least additive activity. Several clinical trials combining gemcitabine with 5-FU, with or without leucovorin, have been performed with use of different sequences and schedules of administration. One phase I trial evaluated a schedule in which gemcitabine was administered as a 30-minute infusion before bolus leucovorin and bolus 5-FU.6 All three drugs were given weekly for 3 weeks followed by 1 week of rest. The maximum-tolerated doses for the three drugs were gemcitabine 1,000 mg/m2, leucovorin 25 mg/m2, and 5-FU 600 mg/m2. The doses and schedule from this study were used in four separate phase II trials. In Eastern Cooperative Oncology Group (ECOG) protocol E3296, 36 patients with metastatic disease were treated and had a median survival of 4.4 months.7 Three other trials used the same schedule and doses of gemcitabine and 5-FU as E3296 (one also used leucovorin) but also allowed patients with locally advanced disease. The median survival times ranged from 7 to 11 months.8-10 Of the studies using the same regimen, those with proportions of patients with 0%, 46%, and 64% locally advanced disease reported median survival times of 4.4, 7, and 11 months, respectively.7,8,10 Although several factors may be responsible for the apparent differences in median survival with the same regimen in pancreas cancer, the enrollment of patients with locally advanced disease may have had a significant role. To clarify the activity of gemcitabine plus 5-FU compared with gemcitabine alone, the ECOG initiated protocol E2297 for patients with both locally advanced and metastatic carcinoma of the pancreas.
Patient Population Eligible patients had microscopically confirmed, measurable, or assessable pancreatic carcinoma (adenocarcinoma or poorly differentiated) not amenable to curative resection. Other eligibility requirements included age  18 years, ability to understand the nature of the trial and provide informed consent, ECOG performance status of 0 to 2, and no other active malignancy or active disease of any type that would preclude safe administration or evaluation of the effect of chemotherapy. Patients must have had adequate organ function, defined as WBC count 3,500/mm3, platelet count 125,000/mm3, bilirubin less than 2.0 mg/dL, and AST  3.0 times the institutional reference value. Patients could have received prior radiation therapy, providing it was completed at least 4 weeks before enrollment. Assessable or measurable disease had to be present outside the radiation ports, or there had to be evidence of postradiation progression within the radiation field. Patients treated previously with adjuvant chemotherapy were eligible if there was an interval of longer than 6 months between completion of adjuvant therapy and diagnosis of recurrent disease. Patients could not have received prior chemotherapy for advanced disease. Pregnant and lactating women were excluded because of risks of fetal or infant injury from therapy. Human subjects committee (institutional review board) approval was required at each participating institution before patients were enrolled.
Chemotherapy Tables for dose modifications based on toxicity were used to ensure uniform treatment at all institutions. Although all supportive measures appropriate for patient care were allowed, use of granulocyte colony-stimulating factor and granulocyte-macrophage colony-stimulating factor was discouraged. Patients were allowed to remain on treatment providing there was no evidence of progressive disease. Patients could be removed from the protocol if intolerable toxicity not amenable to dose modification occurred. Patients could withdraw from the study at any time.
Measurable Disease/Response
Statistical Considerations Fisher’s exact test was used to compare categorical patient characteristics and objective responses between treatment arms. The midrank Wilcoxon test for ordered categorical outcomes was used to compare the distributions of toxicity between treatment arms. Survival curves were estimated by the method of Kaplan and Meier,11 and univariate time-to-event comparisons were performed with the log-rank test.12 Proportional hazards regression models13 of progression-free and overall survival were used to provide adjusted treatment comparisons and identify simultaneous significant prognostic covariates.
Patient Characteristics Between April 1998 and November 1999, 327 patients were enrolled onto the trial. Of these, five were found to be ineligible: four in the gemcitabine plus 5-FU arm and one in the gemcitabine-alone arm. Reasons for ineligibility were incorrect diagnosis (n = 3), synchronous breast primary tumor (n = 1), and no assessable or measurable disease (n = 1). Six patients (three in each arm) did not receive the assigned therapy. Patient characteristics are listed in Table 1. Notably, only 10% of patients were defined as having locally advanced disease. Two imbalances occurred in the randomization. First, the gemcitabine plus 5-FU arm had more patients with a performance status of 1 (P = .046). Although there were correspondingly more patients with a performance status of 0 on the gemcitabine arm, this was not statistically significant. Second, although tumor differentiation and cell type were similar between treatment arms, there was an imbalance in the sites of the original primary tumor. More patients receiving gemcitabine plus 5-FU had a primary tumor in the body of the pancreas. More patients on the gemcitabine arm had primary tumors that encompassed more than one site. The P value for this difference was .005 when all primary tumor sites were evaluated.
Toxicity Table 2 highlights the more common side effects seen. Ninety-eight percent of patients on both arms had at least one grade 1 or higher side effect. The most common grade 3 or higher side effects were hematologic and gastrointestinal. Although toxicities were more common on the gemcitabine plus 5-FU arm, the differences were not significant. Overall, 3% of patients withdrew from the trial on the gemcitabine-alone arm because of side effects, whereas 8% withdrew from the gemcitabine plus 5-FU arm because of intolerability. Overall, more patients experienced higher worst-degree toxicity on the gemcitabine plus 5-FU arm, but this was not statistically significant (P = .33). There were two deaths (one on each arm) on the trial that may have been related to therapy. On the gemcitabine-alone arm, one patient had bacteremia listed as the cause of death but may also have had a pulmonary toxicity. On the gemcitabine plus 5-FU arm, a single patient died as a result of renal failure. An association with therapy could not be definitely ruled out in either case.
Thromboembolism occurred in 3.1% of patients on the trial overall: 2.5% receiving gemcitabine alone and 3.7% receiving gemcitabine and 5-FU. This does not include patients with a history of thromboembolism before study entry or patients listed with phlebitis in Table 2.
Antitumor Effect Of the 322 eligible patients, 318 were assessable for progression-free survival (PFS). Four patients were not included because they remain alive at the time of this article without data regarding progression. PFS was longer for gemcitabine plus 5-FU than for gemcitabine alone (3.4 v 2.2 months; P = .022; Fig 1).
Because of the rapid pace of accrual, the first interim analysis planned at 66.7% information (180 deaths) occurred 4 months after study closure. At that time, the log-rank statistic of 0.48 crossed the lower asymmetric boundary of 1.09 to reject the alternative hypothesis, indicating that the study was likely not to find a difference in survival at full information. At that time, the results were unblinded. As of March 2001, 291 of the 322 patients had died. Figure 2 indicates the overall survival curves by treatment arm. The median survival was 6.7 months for patients receiving 5-FU plus gemcitabine; this was not statistically significantly different from the median survival of 5.4 months for gemcitabine alone (P = .09).
By log-rank testing, several factors predicted for longer median survival, including performance status, site of primary tumor, liver involvement, degree of tumor differentiation, age, and appetite loss (Table 3). Of note, in this analysis, stage of disease (locally advanced v metastatic) did not predict for survival, but only a small proportion of patients had locally advanced disease. Several proportional hazards models were tested for PFS and overall survival. Although these models did not affect PFS calculations, models that included adjustments for the imbalances in performance status caused overall survival differences by treatment arm to reach statistical significance (P = .037), even when the models included adjustments for primary tumor site and other variables.
E2297 demonstrated that the addition of 5-FU to gemcitabine did not add significantly to the survival of patients with pancreas cancer when compared with gemcitabine alone. Although the observed survival difference for E2297 does not seem substantially different from that obtained on the gemcitabine versus 5-FU randomized trial2 that led to the approval of gemcitabine (1.3 v 1.2 months, respectively), there are substantive differences in the overall results. In the trial of gemcitabine versus 5-FU, gemcitabine was superior to 5-FU, with prolonged median survival (5.6 v 4.4 months, respectively; P = .0025), improved 1-year survival (18% v 2%, respectively), and improvement in clinical benefit response (23.8% v 4.8%, respectively; P = .0022).2 On E2297, the median survival differences were not significant (P = .09). The 1-year survival for both arms on E2297 was less than 20% and was not significantly different. Clinical benefit response data were not collected. What was learned from E2297? In comparing the single-agent gemcitabine arm from E2297 with the gemcitabine arm on the randomized trial comparing gemcitabine with 5-FU,2 many similarities appear. The two trials had some differences in eligibility criteria, and the Burris trial called for a short symptom stabilization period that was not required on E2297, whereas E2297 enrolled patients with better performance status.2 Nonetheless, the median survival for gemcitabine alone on E2297 and the Burris trial seems similar (5.3 v 5.65 months). The same is true for 1-year survival and response rates. Thus, with these data, the activity level of single-agent gemcitabine seems to be confirmed in a large cooperative group trial. This is further supported by the median survival and 1-year survival (5.6 months and 19%, respectively) for the gemcitabine arm of another multicenter randomized trial.14 E2297 also demonstrated that gemcitabine in combination with 5-FU is a tolerable regimen. Although there were some differences in the toxicity profiles of the gemcitabine and gemcitabine plus 5-FU arms, there was no statistical difference in the overall toxicity of the two arms, particularly with respect to grade 3 or 4 toxicity. Because 5-FU, as used in E2297, added little to the toxicity profile of gemcitabine, it is possible that administering more 5-FU, modulating 5-FU, or giving infusional 5-FU may increase the efficacy of the regimen enough to reach both clinical and statistical significance. It is unlikely that significantly higher doses of 5-FU could be given on the same schedule as E2297, because this was the maximum-tolerated dose on a phase I trial and was similar to that obtained in a second trial.6,15 Although both phase I trials contained leucovorin, modulating 5-FU is unlikely to add significantly to 5-FU activity in this setting. A literature search produced only one trial comparing 5-FU with 5-FU plus leucovorin in pancreas cancer. In that trial, patients were randomized to 5-FU, 5-FU plus leucovorin, or one of two doses of octreotide. There were no significant differences in time to progression or survival between 5-FU (n = 26) and 5-FU plus leucovorin (n = 27).16 Another approach would be changing to an infusional schedule for 5-FU. Several different phase I or phase II trials have been performed, with median survival times of 6.2 to 10.3 months.17-20 These trials had limited patient numbers, a variety of schedules, and differing patient populations. Although no trials have compared bolus 5-FU with infusional 5-FU in pancreatic carcinoma, these schedules have been compared extensively in colorectal carcinoma. In colorectal carcinoma, 5-FU has clearly demonstrated measurable antitumor activity as opposed to pancreas cancer, where response rates are as low as 0%.2 Despite activity in colon cancer, even with meta-analysis of data from more than 1,200 patients, the difference in overall survival for infusional versus bolus 5-FU was less than 1 month.21 Therefore, data do not support any of these strategies for gemcitabine plus 5-FU. Although the ECOG may consider adding a third active agent to gemcitabine plus 5-FU, the current strategy is to perform a randomized trial comparing standard gemcitabine with a 10-mg/m2/min infusion of gemcitabine and with gemcitabine plus oxaliplatin. After we determined in the overall analysis that the survival was not significantly different, the regression analysis suggested that this may not have been the case had patients been stratified by performance status. With correction for imbalances in performance status, the P values decreased to just below the .05 level. However, these regression models should not be used to change the interpretation of the study results, because the observed differences in overall survival remain clinically small. This trial was originally designed to evaluate for a 50% increase in median survival because this was believed to represent a clinically significant benefit. The 1.3-month improvement in median survival for the gemcitabine plus 5-FU arm was only 24% higher than that observed for the gemcitabine-alone arm. Regardless of P value, the original goal for survival was not reached. The regression analysis did represent an opportunity to define the prognostic factors that require stratification on future trials. On E2297, stratification was supported for performance status and site of primary disease. It is likely that stage II/III versus stage IV disease should be stratified for as well, but this factor did not seem to be associated with prognosis in this trial. This was possibly due to the small proportion of patients with early-stage disease. In conclusion, E2297 demonstrated that the addition of 5-FU to gemcitabine should not replace gemcitabine alone as the standard of care for patients with pancreas cancer. Although there are several different ways to modulate 5-FU or modulate the 5-FU schedule, it is unlikely that any of these changes alone would improve therapy of pancreas cancer significantly. If there is a possibility that an incremental improvement in survival was produced with the addition of 5-FU, then adding other, more active, new agents to improve on the gemcitabine plus 5-FU arm on E2297 may be considered. Although several novel agents and new cytotoxic agents are being studied in pancreas cancer, the ideal agent to add to either gemcitabine or gemcitabine plus 5-FU would potentially synergize with either or both agents without increasing toxicity.
Supported in part by Public Health Service grant nos. CA23318, CA66636, CA21115, CA49957, CA21076, CA13650, CA15488, and CA17145 and by the National Cancer Institute, the National Institutes of Health, and the Department of Health and Human Services. We thank Deb Warren and Carol Caillouette for their tireless work on this trial.
This article’s contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute.
1. Greenleee RT, Hill-Harmon MB, Murray T, et al: Cancer statistics, 2001. CA Cancer J Clin 51: 15-36, 2001
2. Burris HA III, Moore M, Andersen J, et al: Improvements in survival and clinical benefit with gemcitabine as first-line therapy for patients with advanced pancreas cancer: A randomized trial. J Clin Oncol 15: 2403-2413, 1997 3. Carmichael J, Fink U, Russell RCG, et al: Phase II study of gemcitabine in patients with advanced pancreatic cancer. Br J Cancer 73: 101-105, 1996[Medline] 4. Schulz L, Schalhorn A, Wilmanns W, et al: Synergistic interaction of gemcitabine and 5-fluorouracil in colon cancer cells. Proc Am Soc Clin Oncol 17: 251a, 1998 (abstr) 5. Ren Q, Kao V, Grem JL: Cytotoxicity and DNA fragmentation associated with sequential gemcitabine and 5-fluoro-2'-deoxyuridine in HT-29 colon cancer cells. Clin Cancer Res 4: 2811-2818, 1998[Abstract] 6. Berlin JD, Alberti D, Arzoomanian RZ, et al: A phase I study of gemcitabine, 5-fluorouracil, and leucovorin in patients with advanced, recurrent and/or metastatic solid tumors. Invest New Drugs 16: 325-330, 1999 7. Berlin JD, Adak S, Vaughn DJ, et al: A phase II study of gemcitabine and 5-fluorouracil in metastatic pancreatic cancer: An Eastern Cooperative Oncology Group study (E3296). Oncology 58: 215-218, 2000[CrossRef][Medline] 8. Cascinu S, Silva RR, Barni S, et al: A combination of gemcitabine and 5-fluorouracil in advanced pancreatic cancer: A report from the Italian Group for the Study of Digestive Tract Cancer (GISCAD). Br J Cancer 80: 1595-1598, 1999[CrossRef][Medline] 9. Jovtis S, Marantz A, Almira E, et al: Phase II trial of gemcitabine (GEM), 5-fluorouracil (5-FU) and leucovorin (LV) in advanced pancreatic cancer. Eur J Cancer 35: S157, 1999 (suppl, abstr) 10. Pastorelli D, Pedrazzoi S, Sperti C, et al: Phase II trial with gemcitabine (GEM) + 5-fluorouracil (5-FU) in advanced pancreatic cancer (APC). Proc Am Soc Clin Oncol 19: 284a, 2000 (abstr) 11. Kaplan EL, Meier P: Nonparametric estimation of incomplete observations. J Am Stat Assoc 53: 457-481, 1958[CrossRef] 12. Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50: 163-170, 1996 13. Cox DR: Regression models and life tables. J R Stat Soc B 34: 187-220, 1972
14. Bramhall SR, Rosemurgy A, Brown PD, et al: Marimastat as first-line therapy for patients with unresectable pancreatic cancer: A randomized trial. J Clin Oncol 19: 3447-3455, 2001 15. Poplin E, Roberts J, Tombs M, et al: Leucovorin, 5-fluorouracil, and gemcitabine: A phase I study. Invest New Drugs 17: 57-62, 1999[CrossRef][Medline]
16. Burch PA, Block M, Schroeder G, et al: Phase III evaluation of octreotide versus chemotherapy with 5-fluorouracil or 5-fluorouracil plus leucovorin in advanced exocrine pancreatic cancer: A North Central Cancer Treatment Group study. Clin Cancer Res 6: 3486-3492, 2000 17. Oettle H, Pelzer U, Hochmuth K, et al: Phase I trial of gemcitabine (Gemzar), 24 h infusion 5-fluorouracil and folinic acid in patients with inoperable pancreatic cancer. Anticancer Drugs 10: 699-704, 1999[Medline] 18. Rauch DP, Maurer CA, Aebi S, et al: Activity of gemcitabine and continuous infusion fluorouracil in advanced pancreatic cancer. Oncology 60: 43-48, 2001[CrossRef][Medline]
19. Hidalgo M, Castellano D, Paz-Ares L, et al: Phase I-II study of gemcitabine and fluorouracil as a continuous infusion in patients with pancreatic cancer. J Clin Oncol 17: 585-592, 1999 20. Matano E, Tagliaferri P, Libroia A, et al: Gemcitabine combined with continuous infusion 5-fluorouracil in advanced and symptomatic pancreatic cancer: A clinical benefit-oriented phase II study. Br J Cancer 82: 1772-1775, 2000[CrossRef][Medline] 21. Efficacy of intravenous continuous infusion of fluorouracil compared with bolus administration in advanced colorectal cancer: Meta-analysis Group In Cancer. J Clin Oncol 16:301-308, 1998 Submitted November 30, 2001; accepted April 20, 2002. This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
INTRODUCTION
