|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2001 American Society for Clinical Oncology
Clinical Trial Designs for Cytostatic AgentsUniversity of Chicago, Chicago, IL To the Editor:We read with great interest the special article on clinical trial designs for cytostatic agents written by Cancer Therapy Evaluation Program (CTEP) investigators,1 having recently authored a similar article ourselves.2 It is widely recognized that the historical approach to oncology drug development cannot be explicitly used for putative cytostatic agents. This is particularly applicable to phase II trials, where the historical approach of using small cohorts to look for a minimum response rate will clearly fail for drugs for which a classical partial response is not anticipated.3 Korn et al1 suggest that progression-free survival in a 30- to 50-patient cohort can be compared with that of historical controls. In our opinion, this approach is unlikely to be of benefit because of selection bias, lack of standardized data collection in many historical series, as well as the recent change in criteria for response and progression developed in part by CTEP investigators.4 A second approach suggested by the CTEP investigators is the use of small randomized screening studies with large type I errors (alpha = 0.2). This approach will result in a minimum positive rate of 20%. Phase III trials will then be required to sort out true positives from false positives. Historically, far less than 20% of drugs tested in the phase II setting were subsequently proven to be effective. Because phase III trials are notoriously expensive in terms of both financial and patient resources and because a large number of putative cytostatic agents are currently in development, it is unlikely that the current oncologic clinical research environment could support this number of phase III trials. We were particularly concerned by the CTEP investigators’ general recommendation against the randomized discontinuation trial design,5 which has been extensively used outside of oncology, including the Prospective Randomized Study of Ventricular Function and Efficacy of Digoxin trial, demonstrating the efficacy of digoxin in chronic congestive heart failure.6 This design is currently being evaluated in a Cancer and Leukemia Group B trial (sponsored by CTEP) in metastatic renal cancer. Accrual has been brisk, and the study has been well accepted by patients, in contrast to a National Cancer Institute trial of bevacizumab in a similar patient population, where patients are randomized to placebo verses active treatment.7 Korn et al1 are concerned that this design "will lead to an effective agent being declared ineffective if its continued use is not sufficiently better than its initial use." We would argue that a cytostatic agent that is only effective during a very short initial exposure period is highly unlikely to have a significant effect on disease progression or patient survival (as determined in a more standard phase III trial). Given the paucity of success to date in the development of these promising agents, it will be crucial to evaluate a number of different clinical trial designs in a prospective and rigorous manner. Success of a phase II design will depend not only on its ability to stand up to critical statistical analysis but also on its ability to accrue rapidly and on its ability to accurately predict drug benefit in phase III trials. We strongly encourage an open-minded approach that places as much value on ingenuity and originality in trial design as is currently placed on target identification and validation. REFERENCES
1.
Korn EL, Arbuck SG, Pluda JM, et al: Clinical trial designs for cytostatic agents: Are new approaches needed? J Clin Oncol 19: 265-272, 2001 2. Stadler WM, Ratain MJ: Development of target-based antineoplastic agents. Invest New Drugs 18: 7-16, 2000[Medline]
3.
Ratain MJ, Mick R, Schilsky RL, et al: Statistical and ethical issues in the design and conduct of phase I and II clinical trials of new anticancer agents. J Natl Cancer Inst 85: 1637-1643, 1993
4.
Therasse P, Arbuck SG, Eisenhauer EA, et al: New guidelines to evaluate the response to treatment in solid tumors. J Natl Cancer Inst 92: 205-216, 2000 5. Kopec JA, Abrahamowicz M, Esdaile JM: Randomized discontinuous trials: Utility and efficiency. J Clin Epidemiol 46: 959-971, 1993[Medline] 6. Uretsky BF, Young JB, Shahidi FE, et al: Randomized study assessing the effect of digoxin withdrawal in patients with mild to moderate chronic congestive heart failure: Results of the PROVED trial. J Am Coll Cardiol 22: 955-962, 1993[Abstract] 7. Farr NL: Questioning placebo controls Science 288:1747, 2000 (letter)
ResponseNational Cancer Institute, Bethesda, MD In Reply:We, too, are concerned with the possibility of selection bias and lack of standardized data collection when making historical comparisons. That is why we stated, "The historical data required could be the survival or progression-free survival experience for a group of patients with the same stage of disease and amount of prior treatment, similar organ function and performance status, and for whom the same procedures were used for monitoring disease progression. Preferably, this historical experience would come from patients treated at the same institutions with the same referral patterns in a recent era, so that similar diagnostic methodologies and supportive care were available."1 In clinical situations in which such data do not presently exist, we noted one possibility is to acquire the data prospectively in ongoing trials. When the required historical data are not presently available and an agent is ready to be tested, we recommended performing a small screening randomized trial or a large definitive randomized trial and gave some criteria for choosing between the two. Drs Ratain and Stadler are concerned that using screening trials will result in a false-positive rate of 20%, and these false positives will need to be followed up with larger definitive trials. By definition, a screening trial will have more false positives than the 5% we would expect with a definitive trial. However, if the choice is between testing 20% of the agents in large trials or 100% of the agents in large trials, we believe that the oncologic clinical research environment is less likely to support testing 100% of the agents. With many agents currently under development, hard choices of which agents to test in definitive trials will have to be made. We believe the data from randomized screening trials and one-armed trials using valid historical comparisons will help to prioritize which agents should be tested first, especially if these trials can incorporate correlative studies that may confirm effects of the agents on their putative targets. Ratain and Stadler are particularly concerned by our general recommendation against the randomized discontination (enrichment) trial design. Let us consider in more detail the ongoing Cancer and Leukemia Group B trial in metastatic renal cancer mentioned by them. This trial involves the accrual of as many as 335 patients and will test whether an 8-month treatment of carboxyamidotriazole (CAI) leads to longer maintenance of stable disease than a 4-month treatment of CAI among patients who have stable disease after 4 months on the treatment.2 First, note that the required relatively large sample size limits the utility of this type of trial design for screening a large number of agents. Second, if this trial has a positive result, as we all hope, then we will know CAI has some efficacy in treating renal cancer. However, its efficacy in patients who have not already been treated with the agent for 4 months with stable disease will not be known. This may make the indication for use of the agent problematic, especially if only a small proportion of patients treated with CAI have stable disease at 4 months. In addition, it is not obvious what follow-up trials could be performed to clarify this situation after a positive result. However, our major concern with this trial is if it shows there is not a large or statistically significant difference between 8 versus 4 months of treatment for those patients with stable disease. Ratain and Stadler suggest that, in this case, the agent is highly unlikely to have a significant effect on disease progression or patient survival. We know of no evidence on this point and are hesitant to eliminate the development of agents based solely on a negative trial of this sort. Even so, as we noted previously, we could recommend an enrichment design when it is believed to be impossible to conduct a trial with a standard design. In this particular instance, the investigators stated that this was the case, and this was in part the reason that the Cancer Therapy Evaluation Program is sponsoring the trial. Finally, we also encourage the development of new trial designs and nontraditional sequences of types of trials when needed. In fact, the point of our article was to discuss some of the options and their limitations for cytostatic agents. However, Ratain and Stadler seem also to suggest that the paucity of success to date in developing these agents is because of the use of standard trial designs. We would suggest that most of the reason for any lack of success to date has been a result of the ineffectiveness of the agents tested. REFERENCES 1. Korn EL, Arbuck SG, Pluda JM, et al: Clinical trial designs for cytostatic agents: Are new approaches needed? J Clin Oncol 19: 265-272, 2001 2. National Cancer Institute: Cancer Trials. Http://cancertrials.nci .nih.gov This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2001 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
