Key PointsQuestion
Are shortened durations of adjuvant chemotherapy associated with decreased survival among patients with stage II and III colon cancer?
Findings
In this systematic review and meta-analysis of 22 studies comprising 43 671 patients, shortened durations of adjuvant chemotherapy were not associated with worse survival in studies involving combination regimens (the current standard of care) among patients with stage III colon cancer. Conversely, the standard 6 months of chemotherapy was associated with improved survival among patients prescribed monotherapy.
Meaning
Shorter durations of adjuvant chemotherapy may not adversely alter survival among patients with stage III colon cancer treated with combination therapy; patients prescribed monotherapy should be encouraged to complete the entire 6 months of treatment.
Importance
The results from the recent International Duration Evaluation of Adjuvant Therapy (IDEA) collaboration have led some clinicians to adopt shorter durations of adjuvant chemotherapy for patients with stage III colon cancer. The extent to which these findings are supported by other data is unknown.
Objective
To conduct a systematic review and meta-analysis of randomized and observational studies investigating the association between the duration of adjuvant chemotherapy and survival among individuals diagnosed as having stage II and III colon cancer (PROSPERO protocol CRD42018108711]).
Data Sources
Abstracts published in English between 2003 and 2018 within the MEDLINE, Embase, CENTRAL, and CINAHL databases were reviewed by 2 authors. Also searched were conference proceedings and the indexes of high-impact oncology journals.
Study Selection
Studies were excluded if they did not present original data; focused on animal populations, on cancers in sites other than the colon, or on patients with stage 0, I, or IV disease; did not examine a 5-flourouracil–based monotherapy or combination therapy; or did not evaluate the association between treatment duration and survival. The search identified 2341 articles, from which 2 randomized trials and 20 observational studies were included in the meta-analysis.
Data Extraction and Synthesis
This study followed the PRISMA and MOOSE reporting guidelines. The risk of bias was assessed by 2 authors using the Cochrane and Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tools. The results were synthesized using a random-effects model.
Main Outcomes and Measures
The primary and secondary outcomes were overall survival and disease-free survival, respectively. It was hypothesized a priori that 3 months of chemotherapy would be as effective as 6 months of chemotherapy.
Results
Twenty-two studies were included in the meta-analysis, representing 43 671 patients. The inclusion of patients with stage II disease or with rectal cancer was identified as a source of heterogeneity. After restricting the analysis to patients with stage III colon cancer, there was no association between the duration of chemotherapy and overall survival among studies involving FOLFOX (leucovorin calcium [folinic acid], fluorouracil, and oxaliplatin) or CAPOX (capecitabine plus oxaliplatin) regimens (hazard ratio [HR], 0.80; 95% CI, 0.58-1.09). Among studies focused exclusively on monotherapy, the standard 6-month regimen relative to a 3-month regimen was associated with improved survival (HR, 0.59; 95% CI, 0.52-0.68).
Conclusions and Relevance
Shortened durations of chemotherapy may reduce survival among patients with stage III colon cancer prescribed monotherapy but not a combination regimen.
Identifying shorter durations of chemotherapy that minimize adverse effects while maintaining clinical efficacy could lead to improved patient outcomes and cost savings. To address this issue, a recent pooled analysis of 6 randomized trials sought to assess the noninferiority of 3 months of adjuvant FOLFOX (leucovorin calcium [folinic acid], fluorouracil, and oxaliplatin) or CAPOX (capecitabine plus oxaliplatin) chemotherapy relative to the 6-month standard among patients with stage III colon cancer.1 Although the results from this collaboration were inconclusive in the entire study population, noninferiority was observed among certain subgroups, namely, patients with “low risk” stage III disease (T1-3 and N1) and patients treated with CAPOX.1 These findings have resulted in clinical practice guidelines recommending shorter durations of adjuvant chemotherapy within this patient population.2 It is important to note that the International Duration Evaluation of Adjuvant Therapy (IDEA) collaboration1 examined disease-free survival and was conducted within a clinical trial setting, which may have limited generalizability. In addition, the results from this investigation are not generalizable to adjuvant monotherapies, which are still prescribed to patients unable to tolerate oxaliplatin. As such, the results from studies examining adjuvant monotherapies or overall survival and from investigations conducted within real-world settings may help to support the findings from the IDEA collaboration1 and strengthen recommendations within clinical practice guidelines.
Existing review articles on this topic are limited. A meta-analysis3 published in 2010 found that 9 to 12 months of adjuvant chemotherapy was not more effective than 3 to 6 months. That review article did not compare durations of chemotherapy that were less than 6 months, include patients treated with FOLFOX or CAPOX (the current standards of care among patients with stage III colon cancer),2 or synthesize the results from observational studies. A more recent review article4 examining adjuvant chemotherapy duration was conducted; however, that review article was not systematic and did not include a meta-analysis of the evidence base. To address these limitations, we performed a systematic review and meta-analysis of randomized and observational studies that have examined the association between the duration of adjuvant chemotherapy and overall or disease-free survival among patients with stage II and III colon cancer. We hypothesized a priori that 3 months of chemotherapy would be as effective as 6 months of chemotherapy.
We previously registered a protocol for this systematic review and meta-analysis with PROSPERO (CRD42018108711), the international prospective register of systematic reviews.5 The present study followed the Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA)6 and the Meta-analysis of Observational Studies in Epidemiology (MOOSE)7 guidelines.
To be eligible for inclusion, studies had to report the findings from a randomized clinical trial or observational study investigating the association between the duration of a fluorouracil (including prodrugs, such as capecitabine or tegafur) monotherapy or combination therapy (ie, CAPOX or FOLFOX) and overall or disease-free survival among individuals diagnosed as having stage II or III colon cancer. Studies were excluded if they were a review article, commentary, editorial, protocol, case series, or case report; were conducted in a nonhuman (eg, mice) or laboratory (eg, cell line) population; were focused exclusively on patients with cancers in sites other than the colon; were focused on early-stage (stage 0 or I) or metastatic (stage IV) tumors; failed to examine adjuvant chemotherapy duration or discontinuation; examined a therapy other than those previously described; or if a survival outcome was not reported.
Information Sources and Search Strategy
We systematically searched the MEDLINE, Embase, CENTRAL, and CINAHL databases using a query that was developed by one of us (D.J.B.) in collaboration with a research librarian. This query was previously registered with PROSPERO and is available in eAppendix 1 in the Supplement. The search was conducted on August 10, 2018, and was restricted to abstracts published in English between 2003 and 2018.
In addition to our database search, we scanned the references of the included articles, the conference proceedings of the American Society of Clinical Oncology, and the indexes of the following oncology journals: Annals of Oncology, JAMA Oncology, Journal of Clinical Oncology, Cancer, Journal of Oncology Practice, and Lancet Oncology.
The assessment of eligibility was conducted independently by 2 of us (D.J.B. and C.A.C.). Titles and abstracts were first screened for relevance, and then a full-text screen was undertaken. At both stages, agreement between the 2 reviewers was assessed using the percentage agreement and κ statistics. Disagreements between the 2 were resolved through discussion. Non–English language articles with abstracts written in English that were deemed eligible for full-text review were translated and assessed for eligibility.
Data extraction was carried out by the first author (D.J.B.). The results of this extraction were reviewed by another author (D.E.O.) for accuracy. We attempted to contact authors in cases where the hazard ratio (HR) and corresponding 95% CI were not reported or readily estimable using the available data. These emails were sent on December 6, 2018. Articles were not included in the quantitative synthesis if the authors were unable to provide an estimate suitable for meta-analysis or if the authors could not be reached.
The risk of bias was assessed independently by 2 of us (D.J.B. and D.E.O.) using the Cochrane Collaboration’s tool for assessing the risk of bias in randomized trials8 and the Risk of Bias in Nonrandomized Studies of Interventions (ROBINS-I) tool.9 The final decision regarding the risk of bias for each study was reached through discussion by the 2 authors. Regarding the ROBINS-I tool, we developed a list of confounders a priori that would be relevant to the studies included in this systematic review and meta-analysis. We separated confounders into the following 2 categories according to our certainty that their adjustment was necessary to produce an unbiased effect estimate: (1) important confounders—type of chemotherapy, age, sex, site (colon or rectum), and tumor stage; and (2) potential confounders—comorbidity, performance status, body mass index, tumor location, tumor grade, period of diagnosis, time from surgery to the initiation of chemotherapy, presence of postoperative complications, dose delays, and dose reductions. We did not identify any relevant cointerventions. With respect to bias arising from selection into the study, a low-risk designation was given if the index time was defined as the start of chemotherapy initiation and if the inclusion of participants was not potentially related to the exposure and outcome.
To account for differences in the definitions of chemotherapy completion, we converted each measure of association to an estimate of the change in the log HR per 3 additional months of chemotherapy. We first estimated the median number of months of chemotherapy for each duration category using the middle value of the range of the category. In situations where the number of cycles was used to define categories of duration, clinical standards were used to determine the corresponding duration (ie, 1 cycle of FOLFOX given every 2 weeks, 1 cycle of capecitabine monotherapy or CAPOX given every 3 weeks, or 1 cycle of 5-fluorouracil monotherapy given every 4 weeks). As done in a previous systematic review and meta-analysis,10 we estimated the change in the log HR per 3 additional months of chemotherapy as follows: β = [ln(HR) / (x1 − x0)]*3, where x1 and x0 refer to the median duration (in months) of the upper and lower categories, respectively. We similarly estimated the corresponding SE as follows: SE(β) = {(ln[upper 95% CI] – ln[lower 95% CI]) / ([x1 − x0]*3.92)}*3. When estimating the β coefficient and its SE, the multiplication by 3 was done so that they corresponded to the change in the log HR per 3 additional months of chemotherapy as opposed to the change in the log HR per 1 additional month of chemotherapy. Additional information regarding the derivation of effect estimates in situations where the HR and 95% CI were not directly reported can be found in eAppendix 2 in the Supplement.
The primary and secondary outcomes of interest were overall survival and disease-free survival, respectively. To facilitate the pooling of the results from randomized and observational studies, we focused on per-protocol effect estimates. Given the expected between-study heterogeneity due to differences in patient populations, a random-effects DerSimonian-Laird model was used to estimate the average measure of association. In situations in which a study reported effect estimates for independent subgroups, the subgroups were treated as individual studies in the meta-analysis. In addition to 95% CIs, we also estimated the 95% prediction intervals. As previously discussed, prediction intervals should be reported when conducting random-effects meta-analysis because they provide a range of effect estimates that one would expect to observe in a clinical setting similar to those in the meta-analysis.11 The presence of publication bias was assessed via funnel plots and with the Begg test12 and the Egger test.13 We used the I2 statistic, the τ2 estimate, and a χ2 test to assess heterogeneity within this evidence base.14 To identify sources of heterogeneity, we conducted subgroup analyses by treatment-related variables (chemotherapy regimen and the middle period of diagnosis), patient characteristics (mean age and percentage female), tumor features (percentage with colon cancer and percentage with stage III cancer), and study characteristics (country, median follow-up, and risk of bias). Analyses were performed using the metafor package in R Studio (version 1.1.419).15 Statistical significance was defined at the .05 threshold, and all statistical tests were 2-sided.
A PRISMA flow diagram of the inclusion of studies is shown in eFigure 1 in the Supplement. Our database search identified 2341 publications after removal of duplicates. After an initial screen, 205 articles were reviewed in full, from which 34 studies (9 randomized trials1,16-23 and 25 observational studies24-48) were deemed eligible for inclusion. There was strong agreement between the 2 reviewers at both the initial screen and the full-text screen (percentage agreement >90% and κ > 0.6).
Among the 9 randomized trials, only 3 investigated durations of chemotherapy shorter than the current 6-month standard, specifically the IDEA collaboration1 and the studies by Chau et al17 and by Ito et al.20 Among these studies, per-protocol estimates were only available for the IDEA collaboration1 and for the study by Ito et al.20 The investigation by Chau et al17 did not examine different durations of the same treatment modality. Therefore, we were only able to include the results from the IDEA collaboration1 and from the study by Ito et al20 in the meta-analysis.
Among the 25 observational studies identified by our search strategy, 5 were excluded from the meta-analysis. Specifically, we excluded the studies by Ahn et al25 and by Yaich et al47 because we were unable to obtain an effect estimate suitable for inclusion in the meta-analysis and we were unable to reach these authors for further details regarding their findings. In addition, we excluded the studies by Satkunam et al40 and by Tashiro et al43 due to clinical heterogeneity in the exposure definition. Specifically, these researchers did not strictly examine the duration of the prescribed chemotherapy regimen; rather, they studied a composite exposure that included dose reductions and dose modifications, such as a periodic dose omission. Last, we excluded the study by Dobie et al28 because they explored cancer-specific mortality and did not report an estimate specific to overall survival or disease-free survival.
In total, 2 randomized trials and 20 observational studies were included in the meta-analysis,1,20,24,26,27,29-39,41,42,44-46,48 representing 43 671 patients diagnosed between 1987 and 2015. Of these 22 studies, 19 examined overall survival, and 12 examined disease-free survival. Descriptive characteristics are listed in Table 1. The complete data extracted from the individual studies are summarized in eTable 1 in the Supplement.
The results from the study quality assessment are listed in Table 2, and the additional details regarding the reasons for the risk of bias assignment within individual domains are listed in eTable 2 in the Supplement. For the 2 randomized trials included in this systematic review and meta-analysis, we had some concerns regarding the risk of bias in the per-protocol effect estimate due to nonadherence because potential confounding arising from between-arm differences in adherence (there were higher levels of adherence in the shorter-duration arms) was not addressed using appropriate statistical methods.49 The risk of bias in the other domains was deemed to be low. Among the 20 observational studies included in the meta-analysis, 12 were judged to have a serious risk of bias due to confounding (n = 9) or selection into the study (n = 3). The remaining 8 studies were deemed to have a moderate risk of bias. None of the studies in this evidence base were judged to have a high or critical risk of bias.
The results from the meta-analysis of overall survival and disease-free survival are summarized in a forest plot (Figure 1). For both outcomes, there was a suggestion that 6 months of adjuvant chemotherapy relative to 3 months was associated with improved survival; however, the 95% prediction intervals included the null: the HR for overall survival was 0.57 (95% CI, 0.49-0.67 and 95% prediction interval, 0.33-1.00), and the HR for disease-free survival was 0.72 (95% CI, 0.60-0.85 and 95% prediction interval, 0.45-1.14). There was no evidence of publication bias for overall survival according to a qualitative examination of the funnel plots (eFigure 2 in the Supplement) or the Begg test (Kendall τ = −0.14; P = .37) or the Egger test (z = −0.71; P = .48). With respect to disease-free survival, there was a suggestion of possible publication bias arising from a lack of small studies with estimates centered around the null (Kendall τ = −0.10; P = .67 for the Begg test and z = −2.06; P = .04 for the Egger test) (eFigure 3 in the Supplement).
There was moderate to substantial heterogeneity within this evidence base for the primary outcome (I2 = 75.0%; τ2 = 0.07; P < .001). The results from the subgroup analysis are summarized in Table 3. The inclusion of patients with stage II disease in these studies was a statistically significant source of heterogeneity. Specifically, the magnitude of the association was attenuated among studies focused exclusively on patients with stage III disease. Although not considered statistically significant sources of heterogeneity, subgroup analyses suggested that the period of diagnosis, mean age, inclusion of patients with rectal cancer, country of study, and median follow-up were also sources of heterogeneity. Specifically, the magnitude of the association appeared to be more pronounced in studies that were conducted (1) within more recent periods, (2) among older populations, (3) exclusively in patients with colon cancer, (4) in Asian populations, and (5) in those with less follow-up. The chemotherapy regimen, proportion of women, and risk of bias did not appear to be sources of heterogeneity according to these subgroup analyses.
Given the presence of heterogeneity and the identification of tumor stage and site as sources of heterogeneity, we repeated our primary analyses after restricting the analysis to study populations consisting exclusively of patients with stage III colon cancer. Because combination therapy is the current standard of care for patients with stage III colon cancer,2 we stratified these analyses by whether the investigation included patients treated with FOLFOX or CAPOX or was conducted exclusively among patients prescribed monotherapy (Figure 2). With respect to overall survival, there was little to no heterogeneity and no evidence of an association among studies involving combination regimens (HR, 0.80; 95% CI, 0.58-1.09; 95% prediction interval, 0.50-1.27). However, among studies of adjuvant monotherapy, 6 months of chemotherapy was associated with improved overall survival relative to 3 months (HR, 0.59; 95% CI, 0.52-0.68; 95% prediction interval, 0.52-0.68). With respect to disease-free survival, there remained some heterogeneity among studies of combination therapy, so we also further stratified by type of regimen. Only 1 study examined the association between monotherapy duration and disease-free results. In general, the results for disease-free survival were similar to those for overall survival.
Herein, we present the results from the first-ever systematic review and meta-analysis to date of randomized and nonrandomized studies examining durations of adjuvant chemotherapy shorter than 6 months and their association with survival among patients with stage II and III colon cancer. Given the large degree of heterogeneity, it would have been inappropriate to use a single estimate to characterize the results from the underlying evidence base. Because tumor stage and site were identified as sources of heterogeneity, we restricted our analyses to estimates specific to patients with stage III colon cancer and further stratified by chemotherapy regimen. The results from these subgroup analyses suggested that 3 months of chemotherapy relative to 6 months may not be associated with diminished survival among patients with stage III colon cancer prescribed adjuvant FOLFOX or CAPOX, but it may be associated with reduced survival among patients prescribed an adjuvant monotherapy. Estimates from observational studies examining the duration of FOLFOX (n = 2) or CAPOX (n = 1) and disease-free survival among patients with stage III colon cancer were imprecise but consistent with the subgroup analyses of the IDEA collaboration.1
These findings support recommendations made by the National Comprehensive Cancer Network2 of 3 to 6 months of adjuvant combination therapy, depending on the regimen and tumor substage, or 6 months of adjuvant monotherapy if the patient is not a suitable candidate for oxaliplatin therapy. When revising these guidelines to account for the recent findings from the IDEA collaboration,1 concerns were raised regarding the underestimation of the treatment effect arising from the use of disease-free survival as opposed to overall survival.2 Although disease-free survival has been shown to be a good surrogate end point for overall survival in this patient population,50 we found that the magnitude of the average measure of association tended to be larger when examining overall survival as opposed to disease-free survival, which provides some support for these concerns. This discrepancy may be larger if publication bias was present among the estimates of disease-free survival as suggested in our analyses. However, differences between studies with respect to the degree of bias and precision may also account for this discrepancy.
Additional research could further strengthen current clinical practice guidelines. Mature results from the IDEA collaboration1 focused on overall survival will help to address concerns regarding the reliance on disease-free survival. Ancillary per-protocol analyses of the IDEA collaboration1 that address nonadherence and censoring using modern statistical techniques, such as structural nested failure-time models and inverse probability of censoring weights, would also help to strengthen the evidence.49 Future observational studies from large real-world data sets could examine the association between chemotherapy duration and overall survival among strata defined within the subgroup analyses of the IDEA collaboration.1 Such granularity from well-conducted studies could further strengthen the evidence surrounding the targeting of adjuvant chemotherapies for patients with stage III colon cancer. Although adjuvant chemotherapy is offered to high-risk patients with stage II colon cancer,51 additional research regarding the optimal duration of adjuvant treatment within this patient population is needed.
Recommendations for Future Research
Based on our review of the evidence to date, we propose a series of recommendations for improving the conduct of studies of adjuvant chemotherapy duration in real-world settings. Future studies could explore the duration of chemotherapy as an ordinal variable with 3 or more groups to allow for the exploration of a threshold effect. We had initially planned to conduct a nonlinear dose-response meta-analysis, but only 5 studies34,37-39,45 in the evidence base explored 3 or more categories of chemotherapy duration. To facilitate further meta-analysis, the median duration, the number of events, and the total number of patients or person-years at risk within each exposure category should be reported.52 In addition, studies should attempt to minimize the risk of bias due to confounding by adjusting for a greater number of confounders. The identification of sets of confounders that could be controlled for in the analysis can be guided by a directed acyclic graph.53,54 Researchers should be aware that comparisons of different durations of chemotherapy in real-world settings may be limited by immortal time bias.55 Specifically, there will be a period of immortal time among the longer-duration group and thus a survival benefit if the start of follow-up is defined as the date of diagnosis, the date of surgery, or the date of chemotherapy initiation and the exposure is treated as time invariant. To address this issue, we recommend that the index time be defined as the date of chemotherapy initiation and that the duration of treatment be modeled as a time-varying covariate. Among the observational studies included in this evidence base, only 1 study36 modeled chemotherapy duration as a time-varying covariate. Some studies excluded patients lost to follow-up, which may have resulted in selection bias. These patients should be included in the analyses, and inverse probability of censoring weights could be used to address differential loss to follow-up.56 In addition, the number of patients censored and lost to follow-up should be reported so that risk of bias can be better assessed. This information was not available in several of the studies included in this systematic review and meta-analysis. When comparing durations of chemotherapy, differences regarding dose reductions, omissions, or delays should be explored within the analysis. We propose that such differences may be conceptualized as different versions of the treatment or as time-varying confounders.57
This investigation has limitations that should be recognized. First, the results of this meta-analysis are likely biased due to residual confounding. A large proportion of the observational studies included in the meta-analysis (45%) were judged to have a serious risk of bias because they did not adjust for 1 or more of the important confounders we identified a priori. Second, the absence of an association among the subset of studies examining overall survival may be due to a lack of statistical power arising from a limited sample size. Third, the results of our meta-analysis may not be generalizable to patients with stage II disease. Although identified as a source of heterogeneity, there were insufficient data to examine the association of interest among patients with stage II colon cancer. As such, our results should only be generalized to patients with stage III colon cancer.
While additional research is needed, the evidence to date suggests that shortened durations of adjuvant chemotherapy may not negatively alter survival among patients with stage III colon cancer who are prescribed FOLFOX or CAPOX. Conversely, the standard 6 months of adjuvant chemotherapy is associated with improved survival among patients treated who are prescribed monotherapy.
Accepted for Publication: March 26, 2019.
Published: May 17, 2019. doi:10.1001/jamanetworkopen.2019.4154
Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Boyne DJ et al. JAMA Network Open.
Corresponding Author: Darren R. Brenner, PhD, Department of Cancer Epidemiology and Prevention Research, CancerControl Alberta, Alberta Health Services, 3300 Hospital Dr NW, Heritage Medical Research Bldg, Calgary, AB T2N 4Z6, Canada (darren.brenner@ucalgary.ca).
Author Contributions: Dr Brenner had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: Boyne, Hilsden, Friedenreich, Brenner.
Acquisition, analysis, or interpretation of data: Boyne, Cuthbert, O’Sullivan, Sajobi, Cheung, Brenner.
Drafting of the manuscript: Boyne, Brenner.
Critical revision of the manuscript for important intellectual content: Boyne, Cuthbert, O’Sullivan, Sajobi, Hilsden, Friedenreich, Cheung.
Statistical analysis: Boyne.
Administrative, technical, or material support: Boyne, Friedenreich.
Supervision: Sajobi, Hilsden, Friedenreich, Cheung, Brenner.
Conflict of Interest Disclosures: None reported.
Funding/Support: Mr Boyne is supported by the Barrie I. Strafford Doctoral Scholarship for Interdisciplinary Studies on Aging. Mr O’Sullivan is supported by a Queen Elizabeth II Scholarship and an Empire Life Fellowship.
Role of the Funder/Sponsor: The funding sources had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Additional Contributions: Zahra Premji, PhD (Libraries and Cultural Resources, University of Calgary), provided thoughtful feedback pertaining to the search strategy. Yibing Ruan, PhD (Department of Cancer Epidemiology and Prevention Research, Alberta Health Services), translated an article written in Chinese. Neither received compensation.
4.Cruz
JPM, Pales
CGC, Kim
KM, Kim
YW. Adjuvant chemotherapy for high-risk stage II and stage III colon cancer: timing of initiation and optimal duration.
J BUON. 2018;23(3):568-573.
PubMedGoogle Scholar 7.Stroup
DF, Berlin
JA, Morton
SC,
et al; Meta-analysis of Observational Studies in Epidemiology (MOOSE) Group. Meta-analysis of Observational Studies in Epidemiology: a proposal for reporting.
JAMA. 2000;283(15):2008-2012. doi:
10.1001/jama.283.15.2008PubMedGoogle ScholarCrossref 8.Higgins
JPT, Sterne
JAC, Savović
J,
et al. A revised tool for assessing risk of bias in randomized trials. In: Chandler
J, McKenzie
J, Boutron
I, Welch
V, eds.
Cochrane Methods. London, England: Cochrane Database of Systematic Reviews; 2016. doi:
10.1002/14651858.CD201601.
10.Biagi
JJ, Raphael
MJ, Mackillop
WJ, Kong
W, King
WD, Booth
CM. Association between time to initiation of adjuvant chemotherapy and survival in colorectal cancer: a systematic review and meta-analysis.
JAMA. 2011;305(22):2335-2342. doi:
10.1001/jama.2011.749PubMedGoogle ScholarCrossref 16.André
T, Quinaux
E, Louvet
C,
et al. Phase III study comparing a semimonthly with a monthly regimen of fluorouracil and leucovorin as adjuvant treatment for stage II and III colon cancer patients: final results of GERCOR C96.1.
J Clin Oncol. 2007;25(24):3732-3738. doi:
10.1200/JCO.2007.12.2234PubMedGoogle ScholarCrossref 17.Chau
I, Norman
AR, Cunningham
D,
et al. A randomised comparison between 6 months of bolus fluorouracil/leucovorin and 12 weeks of protracted venous infusion fluorouracil as adjuvant treatment in colorectal cancer.
Ann Oncol. 2005;16(4):549-557. doi:
10.1093/annonc/mdi116PubMedGoogle ScholarCrossref 18.Haller
DG, Catalano
PJ, Macdonald
JS,
et al. Phase III study of fluorouracil, leucovorin, and levamisole in high-risk stage II and III colon cancer: final report of Intergroup 0089.
J Clin Oncol. 2005;23(34):8671-8678. doi:
10.1200/JCO.2004.00.5686PubMedGoogle ScholarCrossref 19.Yamaguchi
S, Kunieda
K, Sato
T,
et al. Phase III trial of 24 weeks vs. 48 weeks capecitabine adjuvant chemotherapy for patients with stage III colon cancer: final results of JFMC37-0801.
Ann Oncol. 2016;27(suppl_6):vi149-vi206. doi:
10.1093/annonc/mdw370.18Google Scholar 20.Ito
K, Kato
T, Koike
A,
et al. Optimum duration of oral adjuvant chemotherapy of HCFU for colorectal cancer: review of 5-year follow-up.
Anticancer Res. 2000;20(6C):4681-4686.
PubMedGoogle Scholar 21.Nakamura
T, Ohno
M, Tabuchi
Y,
et al; Kansai Carmofur Study Group. Optimal duration of oral adjuvant chemotherapy with Carmofur in the colorectal cancer patients: the Kansai Carmofur Study Group trial III.
Int J Oncol. 2001;19(2):291-298.
PubMedGoogle Scholar 23.Sadahiro
S, Tsuchiya
T, Sasaki
K,
et al. Randomized phase III trial of treatment duration for oral uracil and tegafur plus leucovorin as adjuvant chemotherapy for patients with stage IIB/III colon cancer: final results of JFMC33-0502.
Ann Oncol. 2015;26(11):2274-2280. doi:
10.1093/annonc/mdv358PubMedGoogle ScholarCrossref 26.Cespedes Feliciano
EM, Lee
VS, Prado
CM,
et al. Muscle mass at the time of diagnosis of nonmetastatic colon cancer and early discontinuation of chemotherapy, delays, and dose reductions on adjuvant FOLFOX: the C-SCANS study.
Cancer. 2017;123(24):4868-4877. doi:
10.1002/cncr.30950PubMedGoogle ScholarCrossref 29.Figer
A, Nissan
A, Shani
A,
et al. Mature results of a prospective randomized trial comparing 5-flourouracil with leucovorin to 5-flourouracil with levamisole as adjuvant therapy of stage II and III colorectal cancer: the Israel Cooperative Oncology Group (ICOG) study.
J Cancer. 2011;2:177-185. doi:
10.7150/jca.2.177PubMedGoogle ScholarCrossref 30.Hassan
AS, Naicker
M, Yusof
KH, Wan Ishak
WZ. Prognostic factors and the role of adjuvant chemotherapy in post-curative surgery for Dukes B and C colon cancers and survival outcomes: a Malaysian experience.
Asian Pac J Cancer Prev. 2015;16(6):2237-2243. doi:
10.7314/APJCP.2015.16.6.2237PubMedGoogle ScholarCrossref 31.Hwang
IG, Lee
JS, Lee
SC,
et al. Association between timing and duration of adjuvant chemotherapy and survival for colorectal cancer in Korea, 2011-2014: a nationwide study based on the database of quality assessment and the health insurance [abstract].
J Clin Oncol. 2017;35(15)(suppl):3605. doi:
10.1200/JCO.2017.35.15_suppl.3605Google Scholar 32.Ji
WB, Hong
KD, Kim
JS, Joung
SY, Um
JW, Min
BW. Effect of a shortened duration of FOLFOX chemotherapy on the survival rate of patients with stage II and III colon cancer.
Chemotherapy. 2018;63(1):8-12. doi:
10.1159/000481566PubMedGoogle ScholarCrossref 33.Kim
CA, Spratlin
JL, Armstrong
DE, Ghosh
S, Mulder
KE. Efficacy and safety of single agent or combination adjuvant chemotherapy in elderly patients with colon cancer: a Canadian cancer institute experience.
Clin Colorectal Cancer. 2014;13(3):199-206. doi:
10.1016/j.clcc.2014.06.002PubMedGoogle ScholarCrossref 36.Laurent
M, Des Guetz
G, Bastuji-Garin
S,
et al. Chronological age and risk of chemotherapy nonfeasibility: a real-life cohort study of 153 stage II or III colorectal cancer patients given adjuvant-modified FOLFOX6.
Am J Clin Oncol. 2018;41(1):73-80. doi:
10.1097/COC.0000000000000233PubMedGoogle Scholar 39.Qiu
MZ, Teng
KY, Ruan
DY,
et al. Impact of adjuvant chemotherapy duration on 3-year disease-free survival of colorectal carcinoma patients after radical resection [in Chinese].
Ai Zheng. 2009;28(7):743-748.
PubMedGoogle Scholar 41.Sgouros
J, Aravantinos
G, Kouvatseas
G,
et al. Impact of dose reductions, delays between chemotherapy cycles, and/or shorter courses of adjuvant chemotherapy in stage II and III colorectal cancer patients: a single-center retrospective study.
J Gastrointest Cancer. 2015;46(4):343-349. doi:
10.1007/s12029-015-9746-8PubMedGoogle ScholarCrossref 42.Sun
J, Ilich
AI, Kim
CA,
et al. Concomitant administration of proton pump inhibitors and capecitabine is associated with increased recurrence risk in early stage colorectal cancer patients.
Clin Colorectal Cancer. 2016;15(3):257-263. doi:
10.1016/j.clcc.2015.12.008PubMedGoogle ScholarCrossref 46.van Erning
FN, Janssen-Heijnen
ML, Creemers
GJ, Pruijt
JF, Maas
HA, Lemmens
VE. Recurrence-free and overall survival among elderly stage III colon cancer patients treated with CAPOX or capecitabine monotherapy.
Int J Cancer. 2017;140(1):224-233. doi:
10.1002/ijc.30423PubMedGoogle ScholarCrossref 47.Yaich
A, Khanfir
A, Bayrouti
MI, Frikha
M. Adjuvant chemotherapy in colon cancer: about 119 cases [in French].
Tunis Med. 2015;93(4):242-247.
PubMedGoogle Scholar 51.Meyers
BM, Cosby
R, Quereshy
F, Jonker
D. Adjuvant systemic chemotherapy for stages II and III colon cancer after complete resection: a clinical practice guideline.
Curr Oncol. 2016;23(6):418-424. doi:
10.3747/co.23.3330PubMedGoogle ScholarCrossref