Phase III Study of Fluorouracil, Leucovorin, and Levamisole in High-Risk Stage II and III Colon Cancer: Final Report of Intergroup 0089
http://www.100md.com
《临床肿瘤学》
the University of Pennsylvania, Philadelphia, PA
Dana-Farber Cancer Institute, Department of Biostatistics, Boston, MA
St Vincent's Hospital and Medical Center, New York, NY
Cancer Centers of the Carolinas-Eastside, Greenville, SC
Dean Medical Center, Madison, WI
Ashville Hematology and Oncology, Asheville, NC
ABSTRACT
PURPOSE: In 1990, fluorouracil (FU) plus levamisole for 1 year became standard adjuvant treatment for patients with high-risk stages II and III colon cancer. Intergroup (INT) 0089 assessed the relative contributions of leucovorin and levamisole in such patients.
PATIENTS AND METHODS: From 1988 to 1992, 3,794 patients were randomly assigned. Experimental treatment consisted of one of three chemotherapy regimens: the low-dose leucovorin plus FU (Mayo Clinic; LDLV) regimen, the high-dose leucovorin plus FU (Roswell Park; HDLV) regimen, and the low-dose leucovorin plus levamisole plus FU (LDLV plus LEV) regimen, each administered for 30 to 32 weeks. The control arm was levamisole plus FU (LEV) for 1 year.
RESULTS: After a median follow-up of 10 years, of 3,561 eligible patients, 1,691 (47%) have died and 1,330 (37%) have experienced disease recurrence; 137 (10%) of those experiencing recurrence are still alive. A total of 481 patients (13%) died without evidence of recurrence, and 1,723 (48%) are alive and disease free. Although there were toxicity differences among the four arms, none was statistically superior in disease-free or overall survival.
CONCLUSION: The 6- to 8-month regimens of LDLV and HDLV without levamisole used in this trial, rather than the previous standard regimen of 12 months of LEV, have become widely used. INT-0089 has long-term follow-up of the largest clinical trial of patients with high-risk colon cancer, documenting not only the durability of the treatment effects, but also the natural history of patients with high-risk colon cancer, and analyses of treatment based on age, race, and comorbid conditions such as obesity, diabetes, and second primary cancers.
INTRODUCTION
Although fluorouracil (FU) was considered the standard of care for metastatic colorectal cancer throughout the 1970s and 1980s, its role in the adjuvant treatment of high-risk, curatively resected colon cancer remained uncertain for at least two decades. In part, this was due to a series of underpowered, poorly designed trials, which used dose schedules that were likely ineffective. In 1988, a meta-analysis determined that the benefits of adjuvant therapy with FU-based chemotherapy for high-risk colon cancer had not been established adequately, despite numerous trials involving many thousands of patients.1 In that report, it was concluded that there could have been a potentially worthwhile clinical benefit from adjuvant chemotherapy, but that larger, prospectively planned studies, with adequate power to detect clinically relevant end points, were warranted.
On the basis of early reports from small studies using the antihelminthic drug levamisole, alone and in combination with FU, a trial of adjuvant chemotherapy of 1 year of postoperative FU plus levamisole (LEV) compared with levamisole alone and with surgery alone was initiated by the North Central Cancer Treatment Group (NCCTG).2 As these data matured, there was also evolving evidence that biochemically modulated regimens of LEV were superior to FU alone in metastatic colorectal cancer, suggesting that such combinations might also be beneficial in the adjuvant setting, compared with previously used regimens that contained FU alone.3 When the preliminary results of the NCCTG trial suggested benefit in both relapse-free and overall survival for both postoperative levamisole alone and for the combination of LEV compared with surgery alone, a confirmatory trial, Intergroup (INT) 0035 was begun, with the same comparisons as in the NCCTG trial, but with separate studies for stage III (node positive) and stage II (node negative) patients.
While these trials were maturing, a number of studies, including the subject of this report, INT-0089, were initiated comparing surgery alone with various regimens containing FU plus leucovorin and/or levamisole for high-risk stage II and all stage III colon cancer.4,5 In addition, there was particular interest in assessing the roles of two commonly used regimens of FU and leucovorin, one in which chemotherapy was administered consecutively for 5 days every month (the Mayo Clinic regimen) and the other with weekly chemotherapy (the Roswell Park regimen). INT-0089 was designed in 1988 to compare the contributions of various regimens of FU, leucovorin, and levamisole in the adjuvant treatment of high-risk colon cancer. The control arm in this study became the 12-month regimen of LEV, which demonstrated a 40% relative reduction in recurrence and a 33% relative risk of death at 6.5 years of follow-up for high-risk, stage III patients.6,7 The goal of the study was overall survival as the main end point; the study was designed to detect a difference of 10 percentage points in overall survival at 5 years.
PATIENTS AND METHODS
Patients for this trial were required to have histologic proof of adenocarcinoma of the colon, and to have undergone curative resection. Patients had to have at least one of the following indicators of poor prognosis: for American Joint Committee on Cancer (AJCC) stage II disease evidence, evidence of bowel obstruction or perforation, adherence to or invasion of adjacent organs or tumor perforation; any AJCC stage III disease; or stage II or III disease and en bloc resection with regional peritoneal or mesenteric implants. There must have been no evidence of distant metastases, and the gross inferior margin of the primary tumor must have been completely above the peritoneal reflection. Patients whose inferior tumor margin was at or below the peritoneal reflection were considered to have had rectal cancer and were ineligible. Patients had to have been randomly assigned onto the study between 21 and 35 days postoperatively and to have provided signed informed consent. All patients must have been maintaining oral nutrition and have Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2, with adequate renal and hepatic functions (bilirubin, AST, alkaline phosphatase, creatinine < 3x upper limit of normal). Patients must have also had a total white blood count 3,500/μL and platelets 100,000/μL.
Patients could not have received concurrent radiation or chemotherapy and could not have had prior exposure to FU. Any patient with a concurrent second malignant disease or previous malignant tumor within the previous 3 years, except for superficial squamous or basal cell carcinoma of the skin, or in situ carcinoma of the cervix, were considered ineligible.
Statistical Design and Analysis
INT-0089 was initially activated in August 1988, with three arms comparing surgery alone with low-dose leucovorin plus FU (LDLV) and high-dose leucovorin plus FU (HDLV). In September 1989, when the results of INT-0035 showed a disease-free survival advantage for treatment with FU plus levamisole compared with surgery alone, the surgery-alone arm of INT-0089 was terminated and random assignment of patients continued to the remaining two arms of LDLV and HDLV.
In January 1990, the study was revised to the final four-arm trial design, given that the results of INT-0035 continued to demonstrate benefit from LEV, which became the control arm. The final statistical design formally identified five treatment comparisons of interest: LEV versus LDLV, LEV versus HDLV, LEV versus LDLV plus LEV, LDLV versus LDLV plus LEV, and LDLV versus HDLV. To maintain an overall 5% type I error, the design called for testing the last three comparisons with one-sided log-rank tests at a significance level of .01 and testing the first two comparisons with two-sided log-rank tests at a significance level of .02. On the basis of the results of INT-0035, the total accrual goal was increased to 3,470 patients in March 1992 to allow at least 3,300 eligible and assessable patients. The study terminated accrual on July 30, 1992, with 3,794 registrations (3,759 assessable patients for the four arms, and excluding 35 ECOG patients who had been randomly assigned to the original surgery-alone arm). In August 1994, an interim monitoring plan was added to the study, calling for two interim analyses at 60% and 80% information, and a final analysis at 100% information.
Overall survival was the main end point, with disease-free survival as a supplementary end point. At the first interim analysis in August 1994, no interim boundaries were crossed for any of the five pairwise treatment comparisons. In August 1995, a second interim analysis was performed, at which time the lower boundary had been crossed for the null hypothesis for the comparisons of LEV versus HDLV, LDLV versus LDLV plus LEV, and LDLV versus HDLV, for both disease-free and overall survival. The results of this analysis were reported. In August 1996, a third interim analysis was conducted for the two remaining comparisons of the study, at which time the comparison of LDLV versus LEV for both disease-free and overall survival crossed the lower boundaries in favor of the null hypothesis, and these results were also reported. In November 1997, the final unblinded comparison of LEV versus LDLV plus LEV also reached full information, and overall survival was then statistically superior for LDLV plus LEV. The 5-year overall survival was 67% for LDLV plus LEV, compared with 63% for LEV (P = .0074). This difference became insignificant with longer follow-up.
In this trial, patients were stratified by extent of bowel wall invasion (limited to bowel wall but not into serosa; into or through serosa; perforation; adherence to adjacent organs; or invasion of adjacent organs), obstruction (present v absent), regional peritoneal or mesenteric implants (present v absent), and extent of regional lymph node metastases (none; one to four; > four). Stratification by these four factors was done separately within each of the three cooperative groups participating in the trial, which included the ECOG, the Cancer and Leukemia Group B, and the Southwest Oncology Group. The treatments were randomly assigned using the ECOG permuted blocks algorithm.8
The distribution of patient characteristics was evaluated using the 2 test for association. The mid-rank Wilcoxon test for ordered categoric outcomes was used to compare the distributions of toxicity grade among the treatment arms, sex, and stage of disease. The Johnckheere-Terpstra test for testing two ordinal categories was used to compare the distributions of toxicities with ordinal predictors.9 Logistic regression for ordinal responses was used in the multivariate analysis of toxicity and tumor involvement.
Survival curves were estimated by the method of Kaplan and Meier with differences assessed by the log-rank test.10,11 Proportional hazards regression models of disease-free and overall survival were based on the likelihood ratio test to identify significant simultaneous prognostics covariates.12
Although nominal P values are reported, univariate pairwise treatment comparisons are judged against the .01 one-sided and .02 two-sided significance levels as specified in the statistical design of the study protocol. A value of P .05 was considered statistically significant for all other comparisons. No adjustments to the treatment comparisons have been performed for interim monitoring or analysis, given that all comparisons are now beyond full information.
Treatment
Experimental treatment consisted of one of three chemotherapy regimens: the LDLV (Mayo Clinic) regimen, the HDLV (Roswell Park) regimen, and the LDLV plus LEV regimen. The control arm was LEV for 1 year. The LDLV regimen consisted of leucovorin 20 mg/m2 and FU 425 mg/m2 on days 1 through 5 by intravenous (IV) push, repeated at 4 weeks, 8 weeks, and every 5 weeks, for a total of six courses. The HDLV regimen consisted of leucovorin 500 mg/m2 by a 2-hour infusion and FU 500 mg/m2 by IV push at 1 hour after the start of the leucovorin infusion, repeated six times weekly, followed by a 2-week rest period, for a total of four 8-week courses. The LDLV plus LEV regimen consisted of levamisole 50 mg orally every 8 hours for 3 days, repeated every 14 days for 6 months, administered with the same chemotherapy as the LDLV treatment regimen. The control arm was the superior arm of INT-0035, the LEV regimen, which consisted of levamisole 50 mg orally every 8 hours for 3 days repeated every 14 days for 1 year and FU 450 mg/m2 by IV push for days 1 through 5, with 450 mg/m2 of FU by IV push weekly beginning on day 29, for a total of 1 year of therapy.
Patient Characteristics
Patient characteristics by treatment are outlined in Tables 1 and 2. Of the 3,759 patients, 198 were ineligible (5.3%). The most common reasons for ineligibility were patients who had stage II disease but were without evidence of perforation, obstruction, adherence, or invasion (67 patients); noncurative resection (17 patients); rectal primary (15 patients); and either duplicate registration or random assignment before 21 days or after 35 days from surgery (36 patients). A total of 50 patients never received assigned therapy, 37 of whom had refused random assignment.
There were no significant imbalances across treatment arms. The median age of patients was 63.7 years, 54.6% were male, and 85.4% were white. Only 3.4% of patients had a performance status of 2. A total of 80.5% of patients had stage III disease, and the remainder had high-risk stage II disease as defined by the protocol. When participation by cooperative groups was examined, most patient characteristic comparisons were nonsignificant, with the exception that the Southwest Oncology Group contributed more minority patients, with lower performance status. Because there was no arbitrary upper age limit for participation, 883 (24.8%) patients were age 70 years or older. When site of colon involvement was evaluated, both sex (P = .005) and age (P < .001) showed statistically significant associations. Males had slightly more left-sided tumor than females, and patients younger than 70 years had a higher rate of left-sided tumors. Treatment and stage were not associated with tumor location.
RESULTS
Toxicity
Toxicity was evaluated for all patients, including ineligible patients, with available toxicity data for 3,695 patients. The 64 patients without toxicity data were primarily from those not receiving assigned therapy. Table 3 summarizes by treatment arm the incidence of common grade 3 or higher treatment-related toxicities. Overall, grade 3 or greater toxicity occurred more frequently for the LDLV arm and the LDLV plus LEV arm. Overall toxicity was not significantly different between the HDLV arm and the LEV arm, but LDLV plus LEV was more significantly toxic than LDLV. The number of deaths potentially attributable to treatment was not significantly different across treatment arms, with five deaths as a result of LDLV, eight deaths as a result of HDLV, four deaths as a result of LEV, and nine deaths as a result of LDLV plus LEV. When the three most common toxicities were analyzed, the incidences of stomatitis, leukopenia, and diarrhea were assessed for age and sex, and separately for each treatment arm. Older age groups and females had significantly higher stomatitis and leukopenia. For stomatitis, the age trends were present for LDLV (P = .010), LEV (P = .015), and LDLV plus LEV (P < .0001). Females had more stomatitis with LDLV (P = .0001), LEV (P = .001), and LDLV plus LEV (P < .0001). For leukopenia, there were age trends in arms LDLV (P = .004) and LDLV plus LEV (P = .008), and females had more leukopenia in all four arms. For diarrhea, older age at entry was not a significant risk, except for those receiving LEV. After controlling for other covariates, stage III patients also had higher risks of stomatitis, leukopenia, and diarrhea. Neurologic symptoms were more commonly reported in the treatment arms containing levamisole, particularly in the control arm, which comprised 12 months of therapy.
Compliance with treatment was high overall, with 88.4% of patients receiving defined treatment. There were no significant differences in compliance among treatment groups, although patients younger than 40 years had significantly better compliance than older patients (P = .019), with no differences between patients age 40 to 70 years and 70 years and older, and between females and males.
Disease-Free and Overall Survival
Overall survival (OS) was defined as time from study entry to death as a result of any cause. Disease-free survival (DFS) was defined as time from study entry to recurrence or death in the absence of data on recurrence. Second colon cancers were counted as events in the disease-free survival analysis. With a median follow-up of 10 years, of 3,561 eligible patients, 1,691 (47%) have died and 1,330 (37%) have had a documented recurrence, with 137 (10%) of those with recurrence still alive. A total of 481 patients (13%) died without evidence of recurrence, and 1,723 patients (48%) are alive and disease free. Table 4 lists 5-year disease-free and overall survival by patient characteristics. Age, sex, performance status, gross feature of the primary tumor, extent of invasion, nodal status, and presence of regional implants were all significant prognostic factors for both disease-free and overall survival. Table 5 lists the 5-year disease-free and overall survival data by treatment arm, and Table 6 lists the same information for 10-year follow-up by treatment arm, AJCC stage, and nodal status. Figure 1 displays these data by treatment arm.
There are no statistically significant differences among treatment arms at final analysis with 100% information; this is in contrast to the interim analysis in November 1997, which showed a slight survival advantage for LDLV plus LEV compared with LEV. The analyses by TNM stage and nodal status document the stability of the proportionate differences in outcome for these high-risk patients during long-term follow-up. In particular, the effect of nodal status shows the wide variability in DFS and OS outcomes in high-risk colorectal cancer, as recently documented in pooled analyses of such patients.13-16 Treatment comparisons by sex and race showed no statistically significant differences between males and females, or between whites and blacks; this latter comparison has been more fully reported elsewhere.17 The univariate analyses for DFS and OS were confirmed in a proportional hazards regression model for age, sex, performance status, obstruction, nodal status, and regional implants (Table 7).
DISCUSSION
This large INT study compared four FU-based adjuvant chemotherapy regimens for high-risk stage II and III colon cancer. The study was designed with adequate power to detect small differences in disease-free and overall survival, and found that none of the formal treatment comparisons was statistically significant after adjustment for multiple comparisons, nor were there differences seen within each stage. Although there were no significant survival differences among the four treatments, the results of this study showed that patients could derive as much benefit from 6 to 8 months of therapy, rather than 12 months, and could accomplish this without the additional toxicity of levamisole. Other contemporary trials also showed that similar shorter treatment programs were as efficacious as 12 months of therapy, and other trials documented that levamisole was not a necessary component of adjuvant chemotherapy, when FU and leucovorin were administered.4,5,18,19 Therefore, both in practice and in clinical trials, the HDLV and LDLV regimens used in this trial, without levamisole, have been in common use since the initial interim analyses of this study and as other studies were reported. In addition, the equivalency of the LDLV and HDLV arms allow for options for patients and physicians, based on toxicity profiles and scheduling. Because of the toxicity profile of the HDLV regimen, which is limited primarily to diarrhea, this regimen has most often been adopted in practice and in subsequent cooperative group trials as the standard control arm. This study also supported the concept that, in the adjuvant setting, efficacy may be predicted by testing in the metastatic population, given that a large trial in metastatic disease of these two regimens showed similar outcomes in response rate, progression-free survival, and overall survival.20
Although 5-year overall survival has been the benchmark of success of adjuvant chemotherapy for colon cancer in research and in practice, this trial and other recent reports suggest two other outcomes should be addressed. In a large meta-analysis, Sargent et al21 has demonstrated that 3-year DFS is an excellent surrogate for 5-year overall survival in clinical trials, given that the vast majority of deaths at 5 years in trials using fluoropyrimidines alone are preceded by relapse. Recently, these data were updated, suggesting that even 2-year DFS may be a valid surrogate for clinical trials, and that the time of relapse may differ between stage II and III patients, with later recurrences in stage II disease.22 As in INT-0089, long-term follow-up of National Surgical Adjuvant Breast and Bowel Project C-03 shows that about 10% of recurrences occur between 5 and 10 years after diagnosis, although survival advantages do not significantly attenuate over time.23 Although these observations may not affect the overall results of clinical trials, they may influence surveillance practices in terms of duration and frequency, especially if the predictive value of surrogacy for 3-year DFS is altered by the introduction of combination adjuvant chemotherapy programs.
In addition, this trial now has long-term follow-up in the single largest population of patients with high-risk colon cancer, allowing for the assessment of findings other than the initial primary end points of the study—disease-free and overall survival. Because there were no outcome differences among the four treatment comparisons, the long-term follow-up of 24,498 life-years for patients receiving adjuvant chemotherapy has also allowed for the determination of the treatment effects and the natural history of patients with high-risk colon cancer. Of this group, 485 new occurrences of cancers have been reported, including both colorectal (n = 64) and other primary tumors (prostate [n = 96], breast [n = 48], and lung [n = 36] were the three most commonly reported new cancers).
Secondary analyses of this trial also have been used to assess various predictive and prognostic features in patients with high-risk colon cancer treated with standard adjuvant treatment regimens, with long-term follow-up. Alone, and pooled with other trials, these data have demonstrated that the elderly are as likely to tolerate and benefit from adjuvant chemotherapy as are younger patients.24,25 In addition, a cohort study of body mass index showed that obese women who received adjuvant chemotherapy had a significant increase in overall mortality, and a borderline significant increase in disease recurrence.26 A similar cohort study of patients with diabetes mellitus who entered this trial demonstrated that these patients also had a significantly higher rate of disease recurrence and mortality.27 The clinical benefit outcomes in black and white patients were found to be similar, as opposed to results suggested in other trials, and black patients overall had less (rather than more) toxicity than white patients.17
Patients in this study were encouraged to have regular follow-up with surveillance for new primary colon cancers with either colonoscopy or sigmoidoscopy with barium enema. A subset analysis of this trial for such surveillance showed a higher-than-predicted likelihood of new primary colon cancers at 5 years, compared with patients in the general population undergoing colonoscopy, and also compared with the incidence seen in the National Polyp Study.28 This suggests that guidelines for patients in the general population or those derived in patients with polyps may not apply to patients with a previous history of colon cancer. In addition, data from this trial have documented that colon cancer survival is associated not only with the accuracy of lymph node sampling and the number of positive lymph nodes, but on an increasing number of lymph nodes obtained and analyzed at surgery.29,30 Analysis of data from INT-0089 did not show that survival was associated with hospital volume, as had been documented previously for rectal primary cancers.31
Therefore, the data from this trial have and can be used to plan and analyze current and future studies; these data have also helped to establish the current standard of therapy for the adjuvant treatment of high-risk colorectal cancer. Data from recently completed trials of combination chemotherapy with FU, leucovorin, and oxaliplatin have documented that patients with high-risk disease benefit from combination treatment compared with FU plus leucovorin alone.32 It is likely also to be true that some patients with low- to moderate-risk disease may continue to receive the simpler regimens used in this and other protocols, with or without some of the newer targeted drugs, such as bevacizumab and cetuximab, which are now being studied in research trials.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported in part by Public Health Service Grants No. CA23318, CA66636, CA21115, CA32291, and CA15488, and from the National Cancer Institute, National Institutes of Health and the Department of Health and Human Services.
Presented in part at the 33rd Annual Meeting of the American Society of Clinical Oncology, Denver, CO, May 20, 1997.
Study contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. This study was coordinated by the Eastern Cooperative Oncology Group (Robert L. Comis, MD).
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Buyse M, Zeleniuch-Jacquotte A, Chalmers TC: Adjuvant therapy of colorectal cancer: Why we still don't know. JAMA 259:3571-3578, 1988
Laurie JA, Moertel CG, Fleming TR, et al: Surgical adjuvant therapy of large bowel adenocarcinoma: An evaluation of levamisole and the combination of levamisole and 5-FU. J Clin Oncol 7:1447-1456, 1989
Piedbois P, Buyse M: What can we learn from a meta-analysis of trials testing the modulation of 5-FU by leucovorin? Advanced Colorectal Meta-Analysis Project. Ann Oncol 4:15-19, 1993 (suppl 2)
O'Connell MJ, Kahn MJ, Erlichman C, et al: Optimal duration of FU plus levamisole colon cancer surgical adjuvant therapy. J Clin Oncol 14:2887, 1996
Wolmark N, Rockette H, Mamounas E, et al: Clinical trial to assess the relative efficacy of FU and leucovorin, FU and levamisole, and FU, leucovorin, and levamisole in patients with Dukes' B and C carcinoma of the colon: Results from National Surgical Adjuvant Breast and Bowel Project C-04. J Clin Oncol 17:3553-3559, 1999
Moertel CG, Fleming TR, Macdonald JS, et al: Levamisole and FU for adjuvant therapy of resected colon carcinoma. N Engl J Med 322:352-358, 1990
Moertel CG, Fleming TR, Macdonald JS, et al: FU plus levamisole as effective adjuvant therapy after resection of stage III colon carcinoma: A final report. Ann Intern Med 122:321-326, 1995
Zelen M: The randomization and stratification of patients to clinical trials. J Chronic Dis 27:365-375, 1974
Mehta CR, Patel NR, Tsiatis AA: Exact significance testing to establish treatment equivalence with ordered categorical data. Biometrics 40:819-825, 1984
Kaplan EL, Meier P: Non-parametric estimation of incomplete observations. J Am Stat Assoc 53:457-481, 1958
Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50:163-170, 1966
Cox DR: Regression models and life tables. J R Stat Soc 34:187-202, 1972
Gill S, Loprinzi CL, Sargent DJ, et al: Pooled analysis of FU-based adjuvant therapy for stage II and III colon cancer: Who benefits and by how much? J Clin Oncol 22:1797-1806, 2004
Greene FL, Stewart AK, Norton HJ: A new TNM staging strategy for node-positive (stage III) colon cancer: An analysis of 50,042 patients. Ann Surg 236:416-421, 2002
Greene FL, Stewart AK, Norton HJ: New tumor-node-metastasis staging strategy for node-positive (stage III) rectal cancer: An analysis. J Clin Oncol 22:1778-1784, 2004
Gunderson LL, Sargent DJ, Tepper JE, et al: Impact of T and N stage and treatment on survival and relapse in adjuvant rectal cancer: A pooled analysis. J Clin Oncol 22:1785-1796, 2004
McCollum AD, Catalano PJ, Haller DG, et al: Outcomes and toxicity in African-American and Caucasian patients in a randomized adjuvant chemotherapy trial for colon cancer. J Natl Cancer Inst 94:1160-1167, 2002
Efficacy of adjuvant FU and folinic acid in colon cancer: International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) investigators. Lancet 345:939-944, 1995
Comparison of fluorouracil with additional levamisole, higher-dose folinic acid, or both, as adjuvant chemotherapy for colorectal cancer: A randomised trial—QUASAR Collaborative Group. Lancet 355:1588-1596, 2000
Buroker TR, O'Connell MJ, Wieand HS: Randomized comparison of two schedules of FU and leucovorin in the treatment of advanced colorectal cancer. J Clin Oncol 12:14-20, 1994
Sargent DJ, Weiand S, Benedetti J, et al: Disease-free survival (DFS) vs. overall survival (OS) as a primary endpoint for adjuvant colon cancer studies: Individual patient data from 12,915 patients on 15 randomized trials. J Clin Oncol 22:245s, 2004 (suppl; abstr 3502)
Sargent DJ: Endpoints for colon adjuvant trials: Recommendations made from independent patient data from 20898 patients and 18 randomized trials. J Clin Oncol 23:249s, 2005 (suppl; abstr 3512)
O'Connell MJ, Wolmark N, Yothers G, et al: Durable improvement in disease-free survival (DFS) and overall survival (OS) for stage II or III colon cancer treated with leucovorin-modulated fluorouracil (FL): 10-year follow-up of National Surgical Adjuvant Breast and Bowel Project (NSABP) protocol C-03. J Clin Oncol 23:248s, 2005 (suppl; abstr 3511)
Sargent DJ, Goldberg RM, Jacobson SD, et al: A pooled analysis of adjuvant chemotherapy for resected colon cancer in elderly patients. N Engl J Med 345:1091-1097, 2001
Schrag D, Cramer LD, Bach PB, et al: Age and adjuvant chemotherapy use after surgery for stage III colon cancer. J Natl Cancer Inst 93:850-857, 2001
Meyerhardt JA, Catalano PJ, Haller DG, et al: Influence of body mass index on outcomes and treatment-related toxicity in patients with colon carcinoma. Cancer 98:484-495, 2003
Meyerhardt JA, Catalano PJ, Haller DG, et al: Impact of diabetes mellitus on outcomes in patients with colon cancer. J Clin Oncol 21:433-440, 2003
Green RJ, Metlay JP, Propert K, et al: Surveillance for second primary colorectal cancer after adjuvant chemotherapy: An analysis of Intergroup 0089. Ann Intern Med 136:261-269, 2002
Joseph NE, Sigurdson ER, Hanlon AL, et al: Accuracy of determining nodal negativity in colorectal cancer on the basis of the number of nodes retrieved on resection. Ann Surg Oncol 10:213-218, 2003
Le Voyer TE, Sigurdson ER, Hanlon AL, et al: Colon cancer survival is associated with increasing number of lymph nodes analyzed: A secondary survey of intergroup trial INT-0089. J Clin Oncol 21:2912-2919, 2003
Meyerhardt JA, Catalano PJ, Schrag D, et al: Association of hospital procedure volume and outcomes in patients with colon cancer at high risk for recurrence. Ann Intern Med 139:649-657, 2003
Andre T, Boni C, Mounedji-Boudiaf L, et al: Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. N Engl J Med 350:2343-2351, 2004(Daniel G. Haller, Paul J.)
Dana-Farber Cancer Institute, Department of Biostatistics, Boston, MA
St Vincent's Hospital and Medical Center, New York, NY
Cancer Centers of the Carolinas-Eastside, Greenville, SC
Dean Medical Center, Madison, WI
Ashville Hematology and Oncology, Asheville, NC
ABSTRACT
PURPOSE: In 1990, fluorouracil (FU) plus levamisole for 1 year became standard adjuvant treatment for patients with high-risk stages II and III colon cancer. Intergroup (INT) 0089 assessed the relative contributions of leucovorin and levamisole in such patients.
PATIENTS AND METHODS: From 1988 to 1992, 3,794 patients were randomly assigned. Experimental treatment consisted of one of three chemotherapy regimens: the low-dose leucovorin plus FU (Mayo Clinic; LDLV) regimen, the high-dose leucovorin plus FU (Roswell Park; HDLV) regimen, and the low-dose leucovorin plus levamisole plus FU (LDLV plus LEV) regimen, each administered for 30 to 32 weeks. The control arm was levamisole plus FU (LEV) for 1 year.
RESULTS: After a median follow-up of 10 years, of 3,561 eligible patients, 1,691 (47%) have died and 1,330 (37%) have experienced disease recurrence; 137 (10%) of those experiencing recurrence are still alive. A total of 481 patients (13%) died without evidence of recurrence, and 1,723 (48%) are alive and disease free. Although there were toxicity differences among the four arms, none was statistically superior in disease-free or overall survival.
CONCLUSION: The 6- to 8-month regimens of LDLV and HDLV without levamisole used in this trial, rather than the previous standard regimen of 12 months of LEV, have become widely used. INT-0089 has long-term follow-up of the largest clinical trial of patients with high-risk colon cancer, documenting not only the durability of the treatment effects, but also the natural history of patients with high-risk colon cancer, and analyses of treatment based on age, race, and comorbid conditions such as obesity, diabetes, and second primary cancers.
INTRODUCTION
Although fluorouracil (FU) was considered the standard of care for metastatic colorectal cancer throughout the 1970s and 1980s, its role in the adjuvant treatment of high-risk, curatively resected colon cancer remained uncertain for at least two decades. In part, this was due to a series of underpowered, poorly designed trials, which used dose schedules that were likely ineffective. In 1988, a meta-analysis determined that the benefits of adjuvant therapy with FU-based chemotherapy for high-risk colon cancer had not been established adequately, despite numerous trials involving many thousands of patients.1 In that report, it was concluded that there could have been a potentially worthwhile clinical benefit from adjuvant chemotherapy, but that larger, prospectively planned studies, with adequate power to detect clinically relevant end points, were warranted.
On the basis of early reports from small studies using the antihelminthic drug levamisole, alone and in combination with FU, a trial of adjuvant chemotherapy of 1 year of postoperative FU plus levamisole (LEV) compared with levamisole alone and with surgery alone was initiated by the North Central Cancer Treatment Group (NCCTG).2 As these data matured, there was also evolving evidence that biochemically modulated regimens of LEV were superior to FU alone in metastatic colorectal cancer, suggesting that such combinations might also be beneficial in the adjuvant setting, compared with previously used regimens that contained FU alone.3 When the preliminary results of the NCCTG trial suggested benefit in both relapse-free and overall survival for both postoperative levamisole alone and for the combination of LEV compared with surgery alone, a confirmatory trial, Intergroup (INT) 0035 was begun, with the same comparisons as in the NCCTG trial, but with separate studies for stage III (node positive) and stage II (node negative) patients.
While these trials were maturing, a number of studies, including the subject of this report, INT-0089, were initiated comparing surgery alone with various regimens containing FU plus leucovorin and/or levamisole for high-risk stage II and all stage III colon cancer.4,5 In addition, there was particular interest in assessing the roles of two commonly used regimens of FU and leucovorin, one in which chemotherapy was administered consecutively for 5 days every month (the Mayo Clinic regimen) and the other with weekly chemotherapy (the Roswell Park regimen). INT-0089 was designed in 1988 to compare the contributions of various regimens of FU, leucovorin, and levamisole in the adjuvant treatment of high-risk colon cancer. The control arm in this study became the 12-month regimen of LEV, which demonstrated a 40% relative reduction in recurrence and a 33% relative risk of death at 6.5 years of follow-up for high-risk, stage III patients.6,7 The goal of the study was overall survival as the main end point; the study was designed to detect a difference of 10 percentage points in overall survival at 5 years.
PATIENTS AND METHODS
Patients for this trial were required to have histologic proof of adenocarcinoma of the colon, and to have undergone curative resection. Patients had to have at least one of the following indicators of poor prognosis: for American Joint Committee on Cancer (AJCC) stage II disease evidence, evidence of bowel obstruction or perforation, adherence to or invasion of adjacent organs or tumor perforation; any AJCC stage III disease; or stage II or III disease and en bloc resection with regional peritoneal or mesenteric implants. There must have been no evidence of distant metastases, and the gross inferior margin of the primary tumor must have been completely above the peritoneal reflection. Patients whose inferior tumor margin was at or below the peritoneal reflection were considered to have had rectal cancer and were ineligible. Patients had to have been randomly assigned onto the study between 21 and 35 days postoperatively and to have provided signed informed consent. All patients must have been maintaining oral nutrition and have Eastern Cooperative Oncology Group (ECOG) performance status of 0, 1, or 2, with adequate renal and hepatic functions (bilirubin, AST, alkaline phosphatase, creatinine < 3x upper limit of normal). Patients must have also had a total white blood count 3,500/μL and platelets 100,000/μL.
Patients could not have received concurrent radiation or chemotherapy and could not have had prior exposure to FU. Any patient with a concurrent second malignant disease or previous malignant tumor within the previous 3 years, except for superficial squamous or basal cell carcinoma of the skin, or in situ carcinoma of the cervix, were considered ineligible.
Statistical Design and Analysis
INT-0089 was initially activated in August 1988, with three arms comparing surgery alone with low-dose leucovorin plus FU (LDLV) and high-dose leucovorin plus FU (HDLV). In September 1989, when the results of INT-0035 showed a disease-free survival advantage for treatment with FU plus levamisole compared with surgery alone, the surgery-alone arm of INT-0089 was terminated and random assignment of patients continued to the remaining two arms of LDLV and HDLV.
In January 1990, the study was revised to the final four-arm trial design, given that the results of INT-0035 continued to demonstrate benefit from LEV, which became the control arm. The final statistical design formally identified five treatment comparisons of interest: LEV versus LDLV, LEV versus HDLV, LEV versus LDLV plus LEV, LDLV versus LDLV plus LEV, and LDLV versus HDLV. To maintain an overall 5% type I error, the design called for testing the last three comparisons with one-sided log-rank tests at a significance level of .01 and testing the first two comparisons with two-sided log-rank tests at a significance level of .02. On the basis of the results of INT-0035, the total accrual goal was increased to 3,470 patients in March 1992 to allow at least 3,300 eligible and assessable patients. The study terminated accrual on July 30, 1992, with 3,794 registrations (3,759 assessable patients for the four arms, and excluding 35 ECOG patients who had been randomly assigned to the original surgery-alone arm). In August 1994, an interim monitoring plan was added to the study, calling for two interim analyses at 60% and 80% information, and a final analysis at 100% information.
Overall survival was the main end point, with disease-free survival as a supplementary end point. At the first interim analysis in August 1994, no interim boundaries were crossed for any of the five pairwise treatment comparisons. In August 1995, a second interim analysis was performed, at which time the lower boundary had been crossed for the null hypothesis for the comparisons of LEV versus HDLV, LDLV versus LDLV plus LEV, and LDLV versus HDLV, for both disease-free and overall survival. The results of this analysis were reported. In August 1996, a third interim analysis was conducted for the two remaining comparisons of the study, at which time the comparison of LDLV versus LEV for both disease-free and overall survival crossed the lower boundaries in favor of the null hypothesis, and these results were also reported. In November 1997, the final unblinded comparison of LEV versus LDLV plus LEV also reached full information, and overall survival was then statistically superior for LDLV plus LEV. The 5-year overall survival was 67% for LDLV plus LEV, compared with 63% for LEV (P = .0074). This difference became insignificant with longer follow-up.
In this trial, patients were stratified by extent of bowel wall invasion (limited to bowel wall but not into serosa; into or through serosa; perforation; adherence to adjacent organs; or invasion of adjacent organs), obstruction (present v absent), regional peritoneal or mesenteric implants (present v absent), and extent of regional lymph node metastases (none; one to four; > four). Stratification by these four factors was done separately within each of the three cooperative groups participating in the trial, which included the ECOG, the Cancer and Leukemia Group B, and the Southwest Oncology Group. The treatments were randomly assigned using the ECOG permuted blocks algorithm.8
The distribution of patient characteristics was evaluated using the 2 test for association. The mid-rank Wilcoxon test for ordered categoric outcomes was used to compare the distributions of toxicity grade among the treatment arms, sex, and stage of disease. The Johnckheere-Terpstra test for testing two ordinal categories was used to compare the distributions of toxicities with ordinal predictors.9 Logistic regression for ordinal responses was used in the multivariate analysis of toxicity and tumor involvement.
Survival curves were estimated by the method of Kaplan and Meier with differences assessed by the log-rank test.10,11 Proportional hazards regression models of disease-free and overall survival were based on the likelihood ratio test to identify significant simultaneous prognostics covariates.12
Although nominal P values are reported, univariate pairwise treatment comparisons are judged against the .01 one-sided and .02 two-sided significance levels as specified in the statistical design of the study protocol. A value of P .05 was considered statistically significant for all other comparisons. No adjustments to the treatment comparisons have been performed for interim monitoring or analysis, given that all comparisons are now beyond full information.
Treatment
Experimental treatment consisted of one of three chemotherapy regimens: the LDLV (Mayo Clinic) regimen, the HDLV (Roswell Park) regimen, and the LDLV plus LEV regimen. The control arm was LEV for 1 year. The LDLV regimen consisted of leucovorin 20 mg/m2 and FU 425 mg/m2 on days 1 through 5 by intravenous (IV) push, repeated at 4 weeks, 8 weeks, and every 5 weeks, for a total of six courses. The HDLV regimen consisted of leucovorin 500 mg/m2 by a 2-hour infusion and FU 500 mg/m2 by IV push at 1 hour after the start of the leucovorin infusion, repeated six times weekly, followed by a 2-week rest period, for a total of four 8-week courses. The LDLV plus LEV regimen consisted of levamisole 50 mg orally every 8 hours for 3 days, repeated every 14 days for 6 months, administered with the same chemotherapy as the LDLV treatment regimen. The control arm was the superior arm of INT-0035, the LEV regimen, which consisted of levamisole 50 mg orally every 8 hours for 3 days repeated every 14 days for 1 year and FU 450 mg/m2 by IV push for days 1 through 5, with 450 mg/m2 of FU by IV push weekly beginning on day 29, for a total of 1 year of therapy.
Patient Characteristics
Patient characteristics by treatment are outlined in Tables 1 and 2. Of the 3,759 patients, 198 were ineligible (5.3%). The most common reasons for ineligibility were patients who had stage II disease but were without evidence of perforation, obstruction, adherence, or invasion (67 patients); noncurative resection (17 patients); rectal primary (15 patients); and either duplicate registration or random assignment before 21 days or after 35 days from surgery (36 patients). A total of 50 patients never received assigned therapy, 37 of whom had refused random assignment.
There were no significant imbalances across treatment arms. The median age of patients was 63.7 years, 54.6% were male, and 85.4% were white. Only 3.4% of patients had a performance status of 2. A total of 80.5% of patients had stage III disease, and the remainder had high-risk stage II disease as defined by the protocol. When participation by cooperative groups was examined, most patient characteristic comparisons were nonsignificant, with the exception that the Southwest Oncology Group contributed more minority patients, with lower performance status. Because there was no arbitrary upper age limit for participation, 883 (24.8%) patients were age 70 years or older. When site of colon involvement was evaluated, both sex (P = .005) and age (P < .001) showed statistically significant associations. Males had slightly more left-sided tumor than females, and patients younger than 70 years had a higher rate of left-sided tumors. Treatment and stage were not associated with tumor location.
RESULTS
Toxicity
Toxicity was evaluated for all patients, including ineligible patients, with available toxicity data for 3,695 patients. The 64 patients without toxicity data were primarily from those not receiving assigned therapy. Table 3 summarizes by treatment arm the incidence of common grade 3 or higher treatment-related toxicities. Overall, grade 3 or greater toxicity occurred more frequently for the LDLV arm and the LDLV plus LEV arm. Overall toxicity was not significantly different between the HDLV arm and the LEV arm, but LDLV plus LEV was more significantly toxic than LDLV. The number of deaths potentially attributable to treatment was not significantly different across treatment arms, with five deaths as a result of LDLV, eight deaths as a result of HDLV, four deaths as a result of LEV, and nine deaths as a result of LDLV plus LEV. When the three most common toxicities were analyzed, the incidences of stomatitis, leukopenia, and diarrhea were assessed for age and sex, and separately for each treatment arm. Older age groups and females had significantly higher stomatitis and leukopenia. For stomatitis, the age trends were present for LDLV (P = .010), LEV (P = .015), and LDLV plus LEV (P < .0001). Females had more stomatitis with LDLV (P = .0001), LEV (P = .001), and LDLV plus LEV (P < .0001). For leukopenia, there were age trends in arms LDLV (P = .004) and LDLV plus LEV (P = .008), and females had more leukopenia in all four arms. For diarrhea, older age at entry was not a significant risk, except for those receiving LEV. After controlling for other covariates, stage III patients also had higher risks of stomatitis, leukopenia, and diarrhea. Neurologic symptoms were more commonly reported in the treatment arms containing levamisole, particularly in the control arm, which comprised 12 months of therapy.
Compliance with treatment was high overall, with 88.4% of patients receiving defined treatment. There were no significant differences in compliance among treatment groups, although patients younger than 40 years had significantly better compliance than older patients (P = .019), with no differences between patients age 40 to 70 years and 70 years and older, and between females and males.
Disease-Free and Overall Survival
Overall survival (OS) was defined as time from study entry to death as a result of any cause. Disease-free survival (DFS) was defined as time from study entry to recurrence or death in the absence of data on recurrence. Second colon cancers were counted as events in the disease-free survival analysis. With a median follow-up of 10 years, of 3,561 eligible patients, 1,691 (47%) have died and 1,330 (37%) have had a documented recurrence, with 137 (10%) of those with recurrence still alive. A total of 481 patients (13%) died without evidence of recurrence, and 1,723 patients (48%) are alive and disease free. Table 4 lists 5-year disease-free and overall survival by patient characteristics. Age, sex, performance status, gross feature of the primary tumor, extent of invasion, nodal status, and presence of regional implants were all significant prognostic factors for both disease-free and overall survival. Table 5 lists the 5-year disease-free and overall survival data by treatment arm, and Table 6 lists the same information for 10-year follow-up by treatment arm, AJCC stage, and nodal status. Figure 1 displays these data by treatment arm.
There are no statistically significant differences among treatment arms at final analysis with 100% information; this is in contrast to the interim analysis in November 1997, which showed a slight survival advantage for LDLV plus LEV compared with LEV. The analyses by TNM stage and nodal status document the stability of the proportionate differences in outcome for these high-risk patients during long-term follow-up. In particular, the effect of nodal status shows the wide variability in DFS and OS outcomes in high-risk colorectal cancer, as recently documented in pooled analyses of such patients.13-16 Treatment comparisons by sex and race showed no statistically significant differences between males and females, or between whites and blacks; this latter comparison has been more fully reported elsewhere.17 The univariate analyses for DFS and OS were confirmed in a proportional hazards regression model for age, sex, performance status, obstruction, nodal status, and regional implants (Table 7).
DISCUSSION
This large INT study compared four FU-based adjuvant chemotherapy regimens for high-risk stage II and III colon cancer. The study was designed with adequate power to detect small differences in disease-free and overall survival, and found that none of the formal treatment comparisons was statistically significant after adjustment for multiple comparisons, nor were there differences seen within each stage. Although there were no significant survival differences among the four treatments, the results of this study showed that patients could derive as much benefit from 6 to 8 months of therapy, rather than 12 months, and could accomplish this without the additional toxicity of levamisole. Other contemporary trials also showed that similar shorter treatment programs were as efficacious as 12 months of therapy, and other trials documented that levamisole was not a necessary component of adjuvant chemotherapy, when FU and leucovorin were administered.4,5,18,19 Therefore, both in practice and in clinical trials, the HDLV and LDLV regimens used in this trial, without levamisole, have been in common use since the initial interim analyses of this study and as other studies were reported. In addition, the equivalency of the LDLV and HDLV arms allow for options for patients and physicians, based on toxicity profiles and scheduling. Because of the toxicity profile of the HDLV regimen, which is limited primarily to diarrhea, this regimen has most often been adopted in practice and in subsequent cooperative group trials as the standard control arm. This study also supported the concept that, in the adjuvant setting, efficacy may be predicted by testing in the metastatic population, given that a large trial in metastatic disease of these two regimens showed similar outcomes in response rate, progression-free survival, and overall survival.20
Although 5-year overall survival has been the benchmark of success of adjuvant chemotherapy for colon cancer in research and in practice, this trial and other recent reports suggest two other outcomes should be addressed. In a large meta-analysis, Sargent et al21 has demonstrated that 3-year DFS is an excellent surrogate for 5-year overall survival in clinical trials, given that the vast majority of deaths at 5 years in trials using fluoropyrimidines alone are preceded by relapse. Recently, these data were updated, suggesting that even 2-year DFS may be a valid surrogate for clinical trials, and that the time of relapse may differ between stage II and III patients, with later recurrences in stage II disease.22 As in INT-0089, long-term follow-up of National Surgical Adjuvant Breast and Bowel Project C-03 shows that about 10% of recurrences occur between 5 and 10 years after diagnosis, although survival advantages do not significantly attenuate over time.23 Although these observations may not affect the overall results of clinical trials, they may influence surveillance practices in terms of duration and frequency, especially if the predictive value of surrogacy for 3-year DFS is altered by the introduction of combination adjuvant chemotherapy programs.
In addition, this trial now has long-term follow-up in the single largest population of patients with high-risk colon cancer, allowing for the assessment of findings other than the initial primary end points of the study—disease-free and overall survival. Because there were no outcome differences among the four treatment comparisons, the long-term follow-up of 24,498 life-years for patients receiving adjuvant chemotherapy has also allowed for the determination of the treatment effects and the natural history of patients with high-risk colon cancer. Of this group, 485 new occurrences of cancers have been reported, including both colorectal (n = 64) and other primary tumors (prostate [n = 96], breast [n = 48], and lung [n = 36] were the three most commonly reported new cancers).
Secondary analyses of this trial also have been used to assess various predictive and prognostic features in patients with high-risk colon cancer treated with standard adjuvant treatment regimens, with long-term follow-up. Alone, and pooled with other trials, these data have demonstrated that the elderly are as likely to tolerate and benefit from adjuvant chemotherapy as are younger patients.24,25 In addition, a cohort study of body mass index showed that obese women who received adjuvant chemotherapy had a significant increase in overall mortality, and a borderline significant increase in disease recurrence.26 A similar cohort study of patients with diabetes mellitus who entered this trial demonstrated that these patients also had a significantly higher rate of disease recurrence and mortality.27 The clinical benefit outcomes in black and white patients were found to be similar, as opposed to results suggested in other trials, and black patients overall had less (rather than more) toxicity than white patients.17
Patients in this study were encouraged to have regular follow-up with surveillance for new primary colon cancers with either colonoscopy or sigmoidoscopy with barium enema. A subset analysis of this trial for such surveillance showed a higher-than-predicted likelihood of new primary colon cancers at 5 years, compared with patients in the general population undergoing colonoscopy, and also compared with the incidence seen in the National Polyp Study.28 This suggests that guidelines for patients in the general population or those derived in patients with polyps may not apply to patients with a previous history of colon cancer. In addition, data from this trial have documented that colon cancer survival is associated not only with the accuracy of lymph node sampling and the number of positive lymph nodes, but on an increasing number of lymph nodes obtained and analyzed at surgery.29,30 Analysis of data from INT-0089 did not show that survival was associated with hospital volume, as had been documented previously for rectal primary cancers.31
Therefore, the data from this trial have and can be used to plan and analyze current and future studies; these data have also helped to establish the current standard of therapy for the adjuvant treatment of high-risk colorectal cancer. Data from recently completed trials of combination chemotherapy with FU, leucovorin, and oxaliplatin have documented that patients with high-risk disease benefit from combination treatment compared with FU plus leucovorin alone.32 It is likely also to be true that some patients with low- to moderate-risk disease may continue to receive the simpler regimens used in this and other protocols, with or without some of the newer targeted drugs, such as bevacizumab and cetuximab, which are now being studied in research trials.
Authors' Disclosures of Potential Conflicts of Interest
The authors indicated no potential conflicts of interest.
NOTES
Supported in part by Public Health Service Grants No. CA23318, CA66636, CA21115, CA32291, and CA15488, and from the National Cancer Institute, National Institutes of Health and the Department of Health and Human Services.
Presented in part at the 33rd Annual Meeting of the American Society of Clinical Oncology, Denver, CO, May 20, 1997.
Study contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. This study was coordinated by the Eastern Cooperative Oncology Group (Robert L. Comis, MD).
Authors' disclosures of potential conflicts of interest are found at the end of this article.
REFERENCES
Buyse M, Zeleniuch-Jacquotte A, Chalmers TC: Adjuvant therapy of colorectal cancer: Why we still don't know. JAMA 259:3571-3578, 1988
Laurie JA, Moertel CG, Fleming TR, et al: Surgical adjuvant therapy of large bowel adenocarcinoma: An evaluation of levamisole and the combination of levamisole and 5-FU. J Clin Oncol 7:1447-1456, 1989
Piedbois P, Buyse M: What can we learn from a meta-analysis of trials testing the modulation of 5-FU by leucovorin? Advanced Colorectal Meta-Analysis Project. Ann Oncol 4:15-19, 1993 (suppl 2)
O'Connell MJ, Kahn MJ, Erlichman C, et al: Optimal duration of FU plus levamisole colon cancer surgical adjuvant therapy. J Clin Oncol 14:2887, 1996
Wolmark N, Rockette H, Mamounas E, et al: Clinical trial to assess the relative efficacy of FU and leucovorin, FU and levamisole, and FU, leucovorin, and levamisole in patients with Dukes' B and C carcinoma of the colon: Results from National Surgical Adjuvant Breast and Bowel Project C-04. J Clin Oncol 17:3553-3559, 1999
Moertel CG, Fleming TR, Macdonald JS, et al: Levamisole and FU for adjuvant therapy of resected colon carcinoma. N Engl J Med 322:352-358, 1990
Moertel CG, Fleming TR, Macdonald JS, et al: FU plus levamisole as effective adjuvant therapy after resection of stage III colon carcinoma: A final report. Ann Intern Med 122:321-326, 1995
Zelen M: The randomization and stratification of patients to clinical trials. J Chronic Dis 27:365-375, 1974
Mehta CR, Patel NR, Tsiatis AA: Exact significance testing to establish treatment equivalence with ordered categorical data. Biometrics 40:819-825, 1984
Kaplan EL, Meier P: Non-parametric estimation of incomplete observations. J Am Stat Assoc 53:457-481, 1958
Mantel N: Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 50:163-170, 1966
Cox DR: Regression models and life tables. J R Stat Soc 34:187-202, 1972
Gill S, Loprinzi CL, Sargent DJ, et al: Pooled analysis of FU-based adjuvant therapy for stage II and III colon cancer: Who benefits and by how much? J Clin Oncol 22:1797-1806, 2004
Greene FL, Stewart AK, Norton HJ: A new TNM staging strategy for node-positive (stage III) colon cancer: An analysis of 50,042 patients. Ann Surg 236:416-421, 2002
Greene FL, Stewart AK, Norton HJ: New tumor-node-metastasis staging strategy for node-positive (stage III) rectal cancer: An analysis. J Clin Oncol 22:1778-1784, 2004
Gunderson LL, Sargent DJ, Tepper JE, et al: Impact of T and N stage and treatment on survival and relapse in adjuvant rectal cancer: A pooled analysis. J Clin Oncol 22:1785-1796, 2004
McCollum AD, Catalano PJ, Haller DG, et al: Outcomes and toxicity in African-American and Caucasian patients in a randomized adjuvant chemotherapy trial for colon cancer. J Natl Cancer Inst 94:1160-1167, 2002
Efficacy of adjuvant FU and folinic acid in colon cancer: International Multicentre Pooled Analysis of Colon Cancer Trials (IMPACT) investigators. Lancet 345:939-944, 1995
Comparison of fluorouracil with additional levamisole, higher-dose folinic acid, or both, as adjuvant chemotherapy for colorectal cancer: A randomised trial—QUASAR Collaborative Group. Lancet 355:1588-1596, 2000
Buroker TR, O'Connell MJ, Wieand HS: Randomized comparison of two schedules of FU and leucovorin in the treatment of advanced colorectal cancer. J Clin Oncol 12:14-20, 1994
Sargent DJ, Weiand S, Benedetti J, et al: Disease-free survival (DFS) vs. overall survival (OS) as a primary endpoint for adjuvant colon cancer studies: Individual patient data from 12,915 patients on 15 randomized trials. J Clin Oncol 22:245s, 2004 (suppl; abstr 3502)
Sargent DJ: Endpoints for colon adjuvant trials: Recommendations made from independent patient data from 20898 patients and 18 randomized trials. J Clin Oncol 23:249s, 2005 (suppl; abstr 3512)
O'Connell MJ, Wolmark N, Yothers G, et al: Durable improvement in disease-free survival (DFS) and overall survival (OS) for stage II or III colon cancer treated with leucovorin-modulated fluorouracil (FL): 10-year follow-up of National Surgical Adjuvant Breast and Bowel Project (NSABP) protocol C-03. J Clin Oncol 23:248s, 2005 (suppl; abstr 3511)
Sargent DJ, Goldberg RM, Jacobson SD, et al: A pooled analysis of adjuvant chemotherapy for resected colon cancer in elderly patients. N Engl J Med 345:1091-1097, 2001
Schrag D, Cramer LD, Bach PB, et al: Age and adjuvant chemotherapy use after surgery for stage III colon cancer. J Natl Cancer Inst 93:850-857, 2001
Meyerhardt JA, Catalano PJ, Haller DG, et al: Influence of body mass index on outcomes and treatment-related toxicity in patients with colon carcinoma. Cancer 98:484-495, 2003
Meyerhardt JA, Catalano PJ, Haller DG, et al: Impact of diabetes mellitus on outcomes in patients with colon cancer. J Clin Oncol 21:433-440, 2003
Green RJ, Metlay JP, Propert K, et al: Surveillance for second primary colorectal cancer after adjuvant chemotherapy: An analysis of Intergroup 0089. Ann Intern Med 136:261-269, 2002
Joseph NE, Sigurdson ER, Hanlon AL, et al: Accuracy of determining nodal negativity in colorectal cancer on the basis of the number of nodes retrieved on resection. Ann Surg Oncol 10:213-218, 2003
Le Voyer TE, Sigurdson ER, Hanlon AL, et al: Colon cancer survival is associated with increasing number of lymph nodes analyzed: A secondary survey of intergroup trial INT-0089. J Clin Oncol 21:2912-2919, 2003
Meyerhardt JA, Catalano PJ, Schrag D, et al: Association of hospital procedure volume and outcomes in patients with colon cancer at high risk for recurrence. Ann Intern Med 139:649-657, 2003
Andre T, Boni C, Mounedji-Boudiaf L, et al: Oxaliplatin, fluorouracil, and leucovorin as adjuvant treatment for colon cancer. N Engl J Med 350:2343-2351, 2004(Daniel G. Haller, Paul J.)