|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 26, No 21 (July 20), 2008: pp. 3653-3654 © 2008 American Society of Clinical Oncology. DOI: 10.1200/JCO.2008.17.5000
Predicting Success in Phase III Studies From Phase II Results: A New Paradigm Is NeededSmall Dog in the Corner Consulting, Chester, NJ To the Editor: We read with interest the article "Review of Phase II Trial Designs Used in Studies of Molecular Targeted Agents: Outcomes and Predictors of Success in phase III" by El-Maraghi and Eisenhower.1 It correctly calls attention to prediction of phase III success for targeted anticancer agents and attempts to place parameters around a complex process with the goal of improving efficiency and outcomes of future studies. We believe that a few aspects of this study require comment. El-Maraghi and Eisenhauer chose to exclude phase II trials that included combinations of agents. This method introduces uncertainty in assessing the applicability to current drug development practices and may be one of the explanations for drugs with minimal or no activity in phase II studies demonstrating clinically important roles in phase III investigations. We reviewed 122 industry-sponsored phase III studies of targeted therapies included in the ClinicalTrials.gov database with registration between 2003 and March 7, 2008. This brief review reveals that 54% of trials combine the investigational agent with an approved chemotherapy or hormonal therapy regimen. Evaluation of targeted agents that facilitate activity of other drugs may certainly benefit from inclusion of combinations in phase II studies. Predicting eventual phase III and regulatory success based on phase II single-agent response rates has always been difficult, but may be even more challenging in the era of molecularly targeted agents. Ratain2 has pointed out that oncology phase II trials have a high negative predictive value. The data analyzed in that study, however, included only chemotherapeutic agents. In the current article, the authors suggest that drugs with low response rates differ from those with no responses in phase II testing. We believe that this interpretation may be spurious. Half of the 38 trials reviewed in their study that demonstrated objective responses had overall response rates of 1% to 10%. Considering the broad experience in oncology comparing independent and investigator-evaluated response rates in phase II studies, it is likely that many of the studies with less than 10% response rate would be reclassified as negative on independent review. (For example, in the phase II study of bevacizumab in lung cancer, the investigators disagreed with the independent reviewers by absolute values of 12.5%, 6.2%, and 8.5%, respectively, in assessment of the response rates in the three randomized treatment arms.3) The authors suggest that targeted agents in the 1% to 10% response rate category warrant phase III testing, given that four of the six agents they examined are currently US Food and Drug Administration–approved. Considering investigator assessment of response is so important in this setting, one might alternatively conclude that investigator enthusiasm is a necessary requisite for eventual approval of targeted therapies with response rates of 0% to 10%. Many of the agents with response rates of zero may never have proceeded to phase III testing because of absence of clinical proof of concept or lack of an effective champion to pursue their causes. Those compounds with investigator-assessed response rates of 1% to 10% may have had, in reality, equal response efficacy but a stronger scientific rationale or more aggressive advocate to justify phase III testing. The importance of a validated rationale for progression to phase III should not be underestimated in the decision-making process. In conclusion, the authors of this article have demonstrated effectively that low response rates in single-agent phase II studies do not reliably discriminate between clinically useful and useless compounds. The phase II study data available for newer targeted agents are not yet of sufficient quantity or quality to establish a reliable rule for decision making. Because of the cost, time, and human resources expended in phase III studies, advancing all agents with response rates less than 10% from phase II to phase III is neither possible nor desirable. Routine use of randomized phase II studies with progression end points may not be necessary for all agents either. Rather, we propose the following schema for discussion regarding phase II drug development of molecularly targeted agents. First, all new agents with phase II response rates more than 20% and acceptable safety and tolerability might be advanced to phase III without randomized phase II testing. Second, agents in classes with clinically validated targets and in vitro activity could be advanced to phase III testing regardless of single-agent phase II response. Third, agents in new classes, with new targets or with uncertain mechanisms of action should undergo two phase II studies, a single-arm response study and a randomized study, utilizing combination therapy if appropriate, designed to assess both response and progression end points. A decision regarding early progression to phase III could be made if the single-arm study showed a significant response rate. Failure of an agent to show some benefit in a phase II randomized progression–end point study, paired with existing agents when appropriate, should raise concern about procession to phase III testing. Prospective study of phase I and II predictors is necessary to identify more specific and sensitive predictive tests. Considering the high financial stakes involved, pharmaceutical and biotechnology companies need to consider cooperating in prospective studies to collect comprehensive information and answer this question as soon as possible. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES
1. El-Maraghi RH, Eisenhauer EA: Review of phase II trial designs used in studies of molecular targeted agents: Outcomes and predictors of success in phase III. J Clin Oncol 26:1346-1354, 2008 2. Ratain M: Phase II oncology trials: Let's be positive. Clin Cancer Res 11:5661-5662, 2005 3. Johnson DH, Fehrenbacher L, Novotny WF, et al: Randomized phase II trial comparing bevacizumab plus carboplatin and paclitaxel with carboplatin and paclitaxel alone in previously untreated locally advanced or metastatic non–small-cell lung cancer. J Clin Oncol 22:2184-2191, 2004 Related Reply
Related Article
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2008 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
