© 2004 by Oxford University Press
© 2004 Oxford University Press
ARTICLE |
Randomized Trial of Adjuvant Therapy in Colon Carcinoma: 10-Year Results of NSABP Protocol C-01
Affiliations of authors: NSABP Operations Center, Pittsburgh, PA (RES, NM); NSABP Biostatistical Center, Pittsburgh, PA (LC, HSW, MB)
Correspondence to: Roy Smith, MD, National Surgical Adjuvant Breast and Bowel Project (NSABP), East Commons Professional Bldg., Four Allegheny Center—5th Floor, Pittsburgh, PA 15212-5234 (e-mail: melissa.wolfe{at}nsabp.org)
 |
ABSTRACT |
|---|
 TopNotes Abstract IntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Background: The National Surgical Adjuvant Breast and Bowel Project C-01 trial reported in 1988 that, for patients with adenocarcinoma of the colon, compared with surgery alone, 1) postoperative chemotherapy with 1-(2-chloroethyl)-3-(4-trans-methylcyclohexyl)-1-nitrosourea (i.e., MeCCNU or semustine), vincristine, and 5-fluorouracil was associated with better 5-year disease-free and overall survival and 2) postoperative immunotherapy with bacillus Calmette-Guérin was associated with better 5-year overall, but not disease-free, survival. We now provide a 10-year update of this trial. Methods: Between November 11, 1977, and February 28, 1983, 1166 patients with resected Dukes’ stage B and C adenocarcinoma of the colon were stratified by Dukes’ stage, sex, and age (<65 years or
65 years) and then randomly assigned to receive no further treatment (surgery alone; 394 patients), adjuvant chemotherapy (379 patients), or adjuvant immunotherapy (393 patients). Those eligible for follow-up included 375 (95.2%) patients in the surgery-alone group, 349 (92.1%) patients in the adjuvant-chemotherapy group, and 372 (94.7%) patients in the adjuvant-immunotherapy group. All statistical tests were two-sided. Results: No difference was observed between patients in the chemotherapy group and those in the surgery-alone group in 10-year disease-free survival (hazard ratio [HR] = 1.14, 95% confidence interval [CI] = 0.94 to 1.39;P = .17) or overall survival (HR = 1.12, 95% CI = 0.91 to 1.38; P= .27). Immunotherapy did not appear to prevent tumor relapse after 10 years (for surgery alone versus immunotherapy, relative risk [RR] = 0.99, 95% CI = 0.78 to 1.25; P = .93) but had a beneficial effect on 10-year overall survival (for surgery alone versus immunotherapy, RR = 1.27, 95% CI = 1.03 to 1.56; P = .02) that apparently results from a reduction in deaths associated with comorbidities in the immunotherapy group. Conclusion: The disease-free and overall survival benefit associated with chemotherapy in this patient population is of limited duration, disappearing after 10 years.
|
INTRODUCTION |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
In November 1977, the National Surgical Adjuvant Breast and Bowel Project (NSABP) began a randomized clinical trial (NSABP protocol C-01) to evaluate the efficacy of adjuvant chemotherapy or postoperative immunotherapy after curative surgery in patients with primary colon cancer. The original aim of the study was to determine whether such therapies could lengthen the disease-free interval and improve the overall survival of patients with adenocarcinoma of the colon that invaded through the muscularis propria into the pericolic tissue without regional node involvement or that involved regional nodes with any depth of invasion. An earlier report (1) on protocol C-01 indicated that patients who received postoperative MOF chemotherapy [i.e., 1-(2-chloroethyl)-3-(4-trans-methylcyclohexyl)-1-nitrosourea (MeCCNU or semustine), vincristine, and 5-fluorouracil] showed overall improvement after a mean of 6.4 years of follow-up in both 5-year disease-free survival and overall survival compared with patients who were treated with surgery alone. However, when patients who received postoperative immunotherapy with bacillus Calmette-Guérin (BCG) were compared with those treated with surgery only, a statistically significant difference was found in overall survival but not in disease-free survival (1).
Eligibility criteria, treatment plans, and follow-up procedures have been described previously (1). The purpose of this report is to provide an update of the results from the complete 10-year data. Follow-up was discontinued at 10 years because nearly all recurrences were observed in the first 5 years of the study, and almost no deaths after 10 years would have been related to cancer recurrence.
|
PATIENTS AND METHODS |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Randomization and Patient Characteristics
After stratification by Dukes’ stage, sex, and age (<65 years or 65 years), a total of 1166 patients with colon cancer at institutions in the United States and Canada had curative surgery and then were randomly assigned between November 15, 1977, and February 28, 1983, to no further treatment (394 patients), postoperative MOF chemotherapy (379 patients), or postoperative immunotherapy (393 patients). There was no blinding of treatment, and no patients received placebo. Three hundred seventy-five (95.2%) of those assigned to the surgery-alone group, 349 (92.1%) of those assigned to the chemotherapy group, and 372 (94.7%) of those assigned to the immunotherapy group were considered eligible for follow-up; 95.9% of the patients contributed complete survival information at 10 years, meaning that they had been monitored for 10 years or died before 10 years of follow-up. Of these 1166 patients, 70 (6.0%) patients were ineligible. The reasons for ineligibility for 46 of these patients were previously reported in detail (1). Since that last report, 24 patients have been found to be ineligible: 14 because of questionable informed consent, six because they had Dukes’ stage D disease, three because they had Dukes’ stage A disease, and another patient because he was found to have concurrent rectal cancer. At the time of this analysis, the primary reasons for ineligibility included concurrent cancer in the rectum (19 patients), Dukes’ stage other than B or C (26 patients), and questionable informed consent (14 patients).
The randomization generated comparable treatment groups. Age, sex, lymph node status (positive versus negative), and location of primary tumor were well balanced in the three randomization arms, although there was a previously noted imbalance in that 9% of the patients who received chemotherapy, 6% of those who received immunotherapy, and 30% of those who underwent surgery alone had eight or more positive lymph nodes (1). Notably, there was a prerandomization phase to this trial, and 80 patients who were randomly assigned to treatment during this phase did not accept their assigned treatment arm (13 in the surgery-alone group, 30 in the chemotherapy group, and 37 in the immunotherapy group). An additional 25 patients assigned to chemotherapy and 27 patients assigned to immunotherapy never started therapy. All patients provided written informed consent to participate in the study, for which local internal review boards reviewed and approved the protocol and consent forms.
Statistical Methods
All analyses used in graphs and tables are based on the 1096 (1166–70 ineligible patients) eligible patients as randomly assigned, as are the majority of analyses discussed in the remaining text. We present a few additional analyses in which the patients who did not receive therapy were omitted to verify that the results remained essentially unchanged. Whenever patients who did not receive therapy were omitted, we explicitly state this fact.
Disease-free survival and overall survival were estimated with Kaplan–Meier curves (2). Relapse-free survival curves were also calculated with the Kaplan–Meier method, in which case patients who had not relapsed (i.e., had no recurrence of disease) were censored at the time of a second primary tumor or death. For the primary analyses, a Cox proportional hazards model was used to compare overall survival, disease-free survival, and relapse-free survival among the treatment groups, stratifying by the number of positive lymph nodes (0, 1–4, 5–7, or 8 positive lymph nodes, but the exact number was unknown). We should note that the treatment effect did not maintain a proportional hazard across time, as during the first 5 years the majority of the events are cancer related, which is not the case in the second 5 years. (A test in which a treatment-by-time interaction action was added to the model was statistically significant for all the curves that were presented.) Partly for this reason, we present both 5-year and 10-year results. We performed secondary analyses in which the treatments were included in a Cox proportional hazards model with number of lymph nodes, age, sex, and location. Risk ratios were obtained from this model (3,4). The score test from a Cox model was used to test for differences between treatment groups (5). All statistical tests were two-sided. Follow-up time was measured from the date of surgery, and all follow-up was censored at 10 years. The analyses were conducted with the SAS program package (6). To test for interactions between treatment and prognostic factors, we fit one model with the treatment, the prognostic factors, and the interaction terms. We then fit a model with only treatment and prognostic factors. Finally, we computed the difference between –2-log likelihood for the model with the main effects and interaction terms and –2-log likelihood for the model with only the main effects to determine whether addition of the interaction terms resulted in a statistically significant difference (P<.05). Had these results been statistically significant, we would then have performed further analyses to identify which interactions caused this effect.
|
RESULTS |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Adjuvant Chemotherapy Compared With Surgery Alone
In the previous report on the first 5 years after randomization (1), it was estimated that treatment with surgery alone, compared with surgery plus adjuvant chemotherapy, was associated with an increased risk of treatment failure (relative risk [RR] = 1.29, 95% confidence interval [CI] = 1.03 to 1.61; P = .03) and with an increased risk of dying (RR = 1.31, 95% CI = 1.02 to 1.68; P = .05). At that time, however, not all patients had been monitored for 5 years. As of this report, the 5-year disease-free survival and overall survival were statistically significantly different between the two arms (for disease-free survival, RR = 1.29 [95% CI = 1.03 to 1.61]; P =.03; and for overall survival, RR = 1.29 [95% CI = 1.01 to 1.66]; P =.04). However, during the subsequent 5 years (years 6–10), these differences became less pronounced, and with complete 10-year data, the differences for 10-year disease-free survival and overall survival between the two arms became statistically nonsignificant (for disease-free survival, hazard ratio [HR] = 1.14 [95% CI = 0.94 to 1.39]; P = .17; and for overall survival, HR = 1.12 [95% CI = 0.91 to 1.38]; P = .27) (Fig. 1, A and B). Restricting these analyses to the patients who received therapy had essentially no impact on the results (for disease-free survival, P = .13; for overall survival, P = .29). However, there was an apparent benefit in favor of chemotherapy that was detected as early as 1 year from the date of randomization and that persisted for up to 8 years of follow-up (Fig. 1, A and B). For 10-year relapse-free survival (Fig. 1, C), the benefit occurred during the first 5 years, which is consistent with overall survival differences persisting for 8 years.
|
Late Toxicity Associated With Chemotherapy
The acute toxic effects associated with the chemotherapy regimen used in the C-01 trial have been previously reported (1). Chronic or late chemotherapy–associated toxic effects appeared to be isolated primarily in the hematologic system. Three patients were diagnosed with acute myelocytic leukemia, and five developed myelodysplastic syndromes. All of these patients have since died. Three of four patients who developed prolonged bone marrow hypoplasia have died, and two of three patients who developed non-Hodgkin lymphoma have died. In addition, there was a chemotherapy-related nephrotoxic event in one patient. Of patients who received immunotherapy, three developed non-Hodgkin lymphoma and one died. A surgery-alone patient who presented with chronic lymphocytic leukemia later died.
Adjuvant Immunotherapy Compared With Surgery Alone
In contrast to the chemotherapy and surgery-alone groups, the disease-free survival curves of those who received immunotherapy and those who received surgery alone were virtually superimposable until year 4 after randomization (Fig. 2). Immunotherapy did not appear to prevent tumor relapse (for surgery alone versus immunotherapy, RR = 0.99, 95% CI = 0.78 to 1.25; P = .93) but, nevertheless, had a beneficial effect on 10-year overall survival (for surgery alone versus immunotherapy, RR = 1.27, 95% CI = 1.03 to 1.56; P = .02). These results for 10-year overall survival were essentially unaffected when the analyses were restricted to patients who received their assigned treatment (for relapse-free survival, RR = 0.94 [95% CI = 0.74 to 1.21]; P = .64; and for overall survival, RR = 1.24 [95% CI = 1.00 to 1.54]; P = .05). This improvement in overall survival appeared to be due to a smaller number of deaths from comorbidities rather than to a decrease in deaths after recurrent colon cancer. In the immunotherapy arm, there were fewer cardiac-related deaths, deaths after second primary cancers other than colon cancer, and deaths from other nonmalignant conditions (Table 1). The reduction in deaths after second primary cancer on this arm could not be attributed to any specific site of second primary cancer.
|
|
The first 126 patients assigned to immunotherapy received the Pasteur strain of BCG, and the last 246 received the Connaught strain of BCG. There was a slightly greater separation in the immunotherapy versus surgery-alone survival curves for the Pasteur strain, but the difference between the effects of the two strains was not statistically significant (P = .60; Fig. 3, A and B).
|
Prognostic Variables
An analysis of baseline variables indicates that age, sex, tumor location, and number of lymph nodes involved with tumors were of prognostic importance for survival. However, sex, age, and tumor location were not statistically significant prognostic factors for tumor recurrence (Table 2).
|
No statistically significant treatment interaction was identified with any of the prognostic variables included in the analysis for immunotherapy or chemotherapy. For tests of interaction, all lymph node–positive patients were included in one group because the patient numbers were small in the categories with higher numbers of positive lymph nodes. Global tests for interaction were not statistically significant in the chemotherapy group (overall survival, P = .58; disease-free survival, P = .57; and relapse-free survival, P = .84) and in the immunotherapy group (overall survival, P = .61; disease-free survival, P = .92; and relapse-free survival, P = .83).
|
DISCUSSION |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
Although a disease-free survival and overall survival advantage in favor of the chemotherapy group compared with the surgery-alone group was demonstrated after the first 5 years of follow-up, this benefit became statistically nonsignificant after 10 years of follow-up. That chemotherapy attenuated the course of disease in patients with either Dukes’ stage B or C disease seems likely because a statistically significant disease-free survival benefit was apparent from as early as 1 year after randomization to as long as 8 years of follow-up. Although the mechanism responsible for this early disease-free survival advantage remains speculative, analysis of relapse-free survival indicates that there was an early delay in recurrence. Similarly, there was an overall survival benefit evident after the first 5 years of follow-up that disappeared after an additional 5 years of follow-up. This observation evidently resulted from a combination of late disease recurrence (>5 years) and subsequent colon cancer–related mortality, deaths from second primary cancers, and deaths from other non–cancer-related comorbidities. It should be noted, however, that MOF, which was used in this trial, is no longer a recommended chemotherapy for colon cancer because of the results of trials that were begun after NSABP Protocol C-01 (7–9). In particular, NSABP Protocol C-03 compared 5-fluorouracil and leucovorin chemotherapy with MOF chemotherapy and demonstrated a superior benefit with the 5-fluorouracil and leucovorin regimen (9). Thus, it is possible that the newer chemotherapy regimens would be better than surgery alone.
The results of this study also demonstrated a continued advantage at 10 years for patients treated with nonspecific BCG immunotherapy, although this difference remained statistically significant only for overall survival. It is of interest that, unlike what we saw in the comparison of the chemotherapy-treated group with the surgery-alone group, the curves for disease-free survival and overall survival between the immunotherapy and surgery-alone groups did not begin to separate until after 4 years of follow-up. This observation raises the possibility that immunotherapy alters the natural history of the comorbidities commonly associated with this patient population whose members, at the time of this analysis, were predominantly older than 65 years. In the 5-year report (1), calculations that omitted deaths without disease recurrence demonstrated that immunotherapy had no direct benefit on disease-free survival, and this result is also supported by our 10-year re-analysis of relapse-free survival and disease-free survival.
In conclusion, the findings from this analysis after 10 years of follow-up continue to indicate that the disease-free survival and overall survival benefit from the use of MOF chemotherapy in this patient population is of limited duration. The overall survival advantage associated with immunotherapy when compared with surgery alone appears to be unrelated to a beneficial alteration in the natural course of the primary malignancy. This observed difference was associated with a reduction in deaths associated with comorbid conditions in this elderly population. It is possible that this observation reflects a real differential effect of immunotherapy on these comorbidities, although we cannot rule out the possibility that this is a chance finding because it was an unanticipated effect, and there were two primary end points and two treatment comparisons.
|
NOTES |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
From the National Surgical Adjuvant Breast and Bowel Project (NSABP) Operations and Biostatistical Centers, Pittsburgh, PA. This work is dedicated to Helen Louise Smith. We thank Barbara C. Good, PhD, for editorial assistance.
This investigation was supported by Public Health Service grants NCI-U10-CA-69651, NCI-U10-CA-12027, NCI-U10-CA-37377, and NCI-U10-CA-69974 from the National Cancer Institute, National Institutes of Health, Department of Health and Human Services, Bethesda, MD.
|
REFERENCES |
|---|
|
TopNotesAbstractIntroductionPatients and MethodsResultsDiscussionReferences |
|---|
1 Wolmark N, Fisher B, Rockette H, Redmond C, Wickerham DL, Fisher E, et al. Postoperative adjuvant chemotherapy or BCG for colon cancer: results from NSABP protocol C-01. J Natl Cancer Inst 1988;80:30–6.
2 Cutler SJ, Ederer F. Maximum utilization of the life table method on analyzing survival. J Chronic Dis 1958;8:699–712.[CrossRef][Medline]
3 Mantel N. Evaluation of survival data and two new rank order statistics arising in its consideration. Cancer Chemother Rep 1966;50:163–70.[Medline]
4 Peto R, Peto J. Asymptotically efficient rank invariant test procedures. J R Stat Soc (A) 1972;135:185–206.
5 Hankey BF, Myers MH. Evaluating differences in survival between two groups of patients. J Chronic Dis 1971;24:523–31.[CrossRef][Web of Science][Medline]
6 SAS Institute Inc. SAS/STAT user's guide, version 6. Vol. 1. 4th ed. Cary (NC): SAS Institute Inc.; 1989.
7 Moertel CG, Fleming TR, Macdonald JS, Haller DG, Laurie JA, Goodman PJ, et al. Levamisole and fluorouracil for adjuvant therapy of resected colon carcinoma. N Engl J Med 1990;322:352–8.[Abstract]
8 O'Connell MJ, Mailliard JA, Kahn MJ, Macdonald JS, Haller DG, Mayer RJ, et al. Controlled trial of fluorouracil and low-dose leucovorin given for 6 months as postoperative adjuvant therapy for colon cancer. J Clin Oncol 1997;15:246–50.
9 Wolmark N, Rockette H, Fisher B, Wickerham DL, Redmond C, Fisher ER, et al. The benefit of leucovorin-modulated fluorouracil as postoperative adjuvant therapy for primary colon cancer: results from National Surgical Adjuvant Breast and Bowel Project Protocol C-03. J Clin Oncol 1993;11:1879–87
Manuscript received August 22, 2003; revised May 25, 2004; accepted June 9, 2004.
Correspondence about this Article
Editorial about this Article
Related Memo to the Media
CiteULike 
Connotea 
Del.icio.us What's this?
J Natl Cancer Inst 2004 96: 1794.
J Natl Cancer Inst 2004 96: 1794.
J Natl Cancer Inst 2004 96: 1116-1117.
J Natl Cancer Inst 2004 96: 1115.
This article has been cited by other articles:
|
|
 |
|
  L. K. Mell, R. R. Weichselbaum, J. A. Bonner, N. Azarnia, and E. K. Rowinsky More on Cetuximab in Head and Neck Cancer N. Engl. J. Med., November 22, 2007; 357(21): 2201 - 2203. [Full Text] [PDF]
|
|
|
|
 |
|
 C. J. A. Punt, M. Buyse, C.-H. Kohne, P. Hohenberger, R. Labianca, H. J. Schmoll, L. Pahlman, A. Sobrero, and J.-Y. Douillard Endpoints in Adjuvant Treatment Trials: A Systematic Review of the Literature in Colon Cancer and Proposed Definitions for Future Trials J Natl Cancer Inst, July 4, 2007; 99(13): 998 - 1003. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 I Chau and D Cunningham Adjuvant therapy in colon cancer--what, when and how? Ann. Onc., September 1, 2006; 17(9): 1347 - 1359. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Mima, S. Yamamoto, M. Ito, R. Tomoshige, Y. Tabata, K. Tamai, and Y. Kaneda Targeted chemotherapy against intraperitoneally disseminated colon carcinoma using a cationized gelatin-conjugated HVJ envelope vector. Mol. Cancer Ther., April 1, 2006; 5(4): 1021 - 1028. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Alikhan, V. Phooskooru, and M. Kohli Re: Randomized Trial of Adjuvant Therapy in Colon Carcinoma: 10-Year Results of NSABP Protocol C-01 J Natl Cancer Inst, December 1, 2004; 96(23): 1794 - 1794. [Full Text] [PDF]
|
|
|
|
|
|
R. E. Smith RESPONSE: Re: Randomized Trial of Adjuvant Therapy in Colon Carcinoma: 10-Year Results of NSABP Protocol C-01 J Natl Cancer Inst, December 1, 2004; 96(23): 1794 - 1794. [Full Text] [PDF]
|
|
|
|
|
|
J. Grem Adjuvant Therapy for Colon Cancer: A Historical Perspective J Natl Cancer Inst, August 4, 2004; 96(15): 1116 - 1117. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||