|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 16 (June 1), 2008: pp. 2787-2788 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2008.17.1785
Stopping Trials for Benefit Can (Sometimes) Benefit PatientsNational Cancer Institute of Canada Clinical Trials Group, Queen's University, Kingston, Ontario, Canada
Massachusetts General Hospital Cancer Center, Boston, MA
National Cancer Institute of Canada Clinical Trials Group, Queen's University, Kingston, Ontario, Canada To the Editor: The commentary by Wilcox et al1 provides a useful perspective on the difficulty of interpreting trials stopped early for efficacy, although we think it would best be read in the context of the more comprehensive discussion in their earlier article2 and, in particular, the editorial accompanying that article by Stuart Pocock.3 As Pocock comments, because the decision to stop a trial for efficacy is a difficult one, it is critical that it be made by an independent data safety monitoring committee that is well aware of the clinical context as well as statistical landmines. In their commentary, Wilcox et al1 stress the fact that they found a "strong correlation between the apparent treatment effect and the number of events" in trials stopped early and go on to argue that, "it would seem prudent to perform interim analyses only after a sufficient number of events have occurred so as to reduce the likelihood of overestimating the true treatment effect." However, they seem to have overlooked the fact that such a correlation is an expected outcome of applying rules that were devised to prevent premature closure. As Pocock3 points out, a large difference would need to be present to generate a significance level that crossed a boundary of P = .001 at a first interim analysis when there were few events. However, at later analyses with more events, a smaller difference might meet that criterion. This inverse relationship between the number of events that have occurred and the size of the treatment effect needed to cross a boundary is accentuated further when, as in many oncology trials, O'Brien-Fleming boundaries, which require more stringent significance levels at earlier analyses, are used. Thus, Wilcox et al's observation is not surprising and does not necessarily imply that decisions to stop trials in the presence of large treatment effects and highly significant differences, but relatively fewer events, are made any less thoughtfully. We agree that the first interim analysis should be based on more than a minimal amount of information. The current guideline for cooperative group trials is to conduct the first interim analysis when 25% of the events required at the time of the final analysis have occurred. The key issue, though, as Wilcox et al1 state, is what to make of the treatment effect observed when a trial is stopped early for efficacy. The earlier article argued for the use of statistical procedures to adjust the magnitude of the claimed benefit and its statistical significance, but, in his editorial, Pocock3 pointed out that such adjustments are unlikely to be useful in a particular circumstance. Thus, more empiric data addressing this issue would be valuable. As again noted by Pocock, such data may frequently be available because often the decision to stop is made when some data are still outstanding. Thus, in MA17 (one of the trials included in the review), there were 207 events when the decision to terminate the study was made4; when all data available before unblinding were eventually assembled, an additional 40 events had been reported.5 The observed hazard ratio was 0.57 (P = .00008) in the first instance and 0.58 (P = .00004) in the latter. Re-examination of data from a closed trial may also shed light on the question of what might have happened if the trial had stayed open, particularly when patients on the study have been exposed to different durations of therapy at the time of closure. In MA17, such an analysis indicated that the relative benefit of letrozole versus placebo improved with a longer duration of administration, suggesting that if the trial had continued to its final analysis, the difference observed at the interim analysis might have increased.6 Therefore, we look forward to the results of the authors’ future investigations. More information on the eventual outcomes of trials stopped early for benefit will help not only those who must act on their results, but also the data safety monitoring committees who must balance the interests of patients on the trial with those of future patients who may be harmed, but also might benefit, from early implementation of study results. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST Although all authors completed the disclosure declaration, the following author(s) indicated a financial or other interest that is relevant to the subject matter under consideration in this article. Certain relationships marked with a "U" are those for which no compensation was received; those relationships marked with a "C" were compensated. For a detailed description of the disclosure categories, or for more information about ASCO's conflict of interest policy, please refer to the Author Disclosure Declaration and the Disclosures of Potential Conflicts of Interest section in Information for Contributors. Employment or Leadership Position: None Consultant or Advisory Role: None Stock Ownership: None Honoraria: None Research Funding: Paul Goss, Novartis; Ralph Meyer, Novartis Expert Testimony: None Other Remuneration: None REFERENCES
1. Wilcox RA, Djulbegovic BJ, Guyatt GH, et al: Randomized trials in oncology stopped early for benefit. J Clin Oncol 26:18-19, 2008 2. Montori VM, Devereaux PJ, Adhikari NK, et al: Randomized trials stopped early for benefit: A systematic review. JAMA 294:2203-2209, 2005 3. Pocock SJ: When (not) to stop a clinical trial for benefit. JAMA 294:2228-2230, 2005 4. Goss PE, Ingle JN, Martino S, et al: A randomized trial of letrozole in postmenopausal women after five years of tamoxifen therapy for early-stage breast cancer. N Engl J Med 349:1793-1802, 2003 5. Goss PE, Ingle JN, Martino S, et al: Randomized trial of letrozole following tamoxifen as extended adjuvant therapy in receptor positive breast cancer: Updated findings from NCIC CTG MA.17. J Natl Cancer Inst 97:1262-1271, 2005 6. Ingle JN, Tu D, Pater JL, et al: Duration of letrozole treatment and outcomes in the placebo-controlled NCIC CTG MA.17 extended adjuvant therapy trial. Breast Cancer Res Treat 99:295-300, 2006[CrossRef][Medline]
Related Reply
Related Article
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
