|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 23, No 15 (May 20), 2005: pp. 3643-3644 © 2005 American Society of Clinical Oncology. DOI: 10.1200/JCO.2005.05.321
Chemotherapy Sensitivity and Response Assays: Are the ASCO Guidelines for Clinical Trial Design Too Restrictive?University of Pittsburgh, Pittsburgh, PA To the Editor: I read with interest the article by Schrag et al1 regarding chemotherapy sensitivity and resistance assays (CSRAs). Although I am in agreement that an emphasis on rigorous evaluation of these assays is important; I would like to express some concerns regarding the task force’s assumptions and recommendations. The panel recommended that trials to evaluate the clinical efficacy of CSRA’s should prospectively "randomly assign patients to the regimens identified on the basis of in vitro testing ("in vitro best regimen") or to empiric therapy with the standard regimen." Although I agree that this is a valid approach, there are other designs that can offer the same scientific rigor and may be more practical in evaluating the clinical utility of CSRAs. For example, suppose two drugs, A and B, are being used to treat a specific disease in a clinical trial. Unbiased estimates of the benefit of assay directed therapy for an outcome (eg, 2-year survival) could be obtained by knowing the estimated survival for the eight categories of chemosensitivity and chemoresistance presented in Table 1.
Then if SA is greater than SB for all four combinations of chemosensitivity and chemoresistance, there is virtually no reason to perform the assay because one would always use treatment A. However, if SA is greater than SB for at least one of the four combinations (eg, SA+– > SB+–) and SB is greater than SA for at least one of the four combinations (eg, SB–+ > SA–+), then there is clear justification for using the assay in clinical practice. Therefore, a perfectly valid way to obtain unbiased estimates of these two-year survivals is to do a randomized trial of drug A versus drug B, to perform the assay on all patients (keeping the patient and physician blinded) and obtain the survival estimates at the end of the trial. In fact, this proposed trial design would provide more information regarding these estimates than the design proposed by Schrag et al. In the Schrag approach, if drug A were the standard of care, presumably patients would only be assigned to drug B if it was shown to be the in vitro best regimen (SB–+). This approach assumes that drug A would be better than drug B in the other three groups. In other words, the Schrag et al design would assign drug A to all patients unless they were only sensitive to drug B, and ignore the important information provided by the patients in the SB++, SB+–, SB– categories. For this reason, we are advocating embedding CSRAs into clinical trials in which two drugs whose relative efficacy is being tested in a randomized clinical trial—a common situation. Second, it should be noted that CSRAs are not proposed as an alternative to empiric therapy, but rather an enhancement to help identify the most effective treatment for an individual when multiple options exist, much like the use of estrogen-receptor expression and tamoxifen. For this reason, I believe that other trial designs more typical of diagnostic assays can provide important clinical information regarding their predictive accuracy. For example, a blind, noninterventional, one-arm assay accuracy trial utilizing two or more therapies known to have similar response rates would determine whether the assay is predictive of response. In this situation, if patients treated with drugs predicted by assay to be chemosensitive had a higher response rate than patients treated with drugs assayed to be chemoresistant, it could then be inferred that patients treated with an assay-sensitive drug would be more likely to respond than those treated with an assay-resistant drug. One would still want to perform a randomized trial to eliminate the potential bias from a historical data comparison. However, data from the one-arm assay trial would already be useful for patients and caregivers because the information could supplement other clinical information in making a treatment decision. In summary, I strongly support the emphasis on randomized trials to evaluate CRSA’s. However, I believe there are alternatives to the trial proposed by Shrag et al, and that in cases where there are several equivalently efficacious treatments available, patients and caregivers can benefit from CRSA results obtained in noninterventional trials as supplements to other clinical data when deciding on a treatment. Author's Disclosure of Potential Conflicts of Interest The following author or their immediate family members have indicated a financial interest. No conflict exists for drugs or devices used in a study if they are not being evaluated as part of the investigation. Consultant: H. Samuel Wieand, Precision Therapeutics, Inc. For a detailed description of these 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 of Information for Contributors found in the front of every issue.
REFERENCE
1. Schrag D, Garewal H, Burstein H, et al: American Society of Clinical Oncology Technology Assessment: Chemotherapy Sensitivity and Resistance Assays. J Clin Oncol 22
: 3631
-3638, 2004
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2005 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
