|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Originally published as JCO Early Release 10.1200/JCO.2008.18.4218 on September 8 2008 © 2008 American Society of Clinical Oncology.
SWOG S0023: What Meets the Eye May Be Only Half the Truth
Department of Thoracic/Head and Neck Medical Oncology, The University of Texas M. D. Anderson Cancer Center, Houston, TX To the Editor: We read with interest the results of the Southwest Oncology Group (SWOG) S0023 trial1 because we had earlier tested the concept of gefitinib maintenance in advanced non–small-cell lung cancer (NSCLC).2 Several stakeholders might consider the negative results of the SWOG S0023 trial as yet another nail in the coffin for gefitinib; however, we concur with Keedy et al3 that these results perhaps highlight the gaping void in our understanding of lung cancer biology. In the SWOG S0023 trial, gefitinib's value was tested in the maintenance setting for stage III NSCLC patients who were not preselected for molecular markers and who continued to have complete response, partial response, or stable disease after chemoradiotherapy and consolidation docetaxel. This study design practically filtered out all patients who dropped out, had disease progression, died, or experienced adverse effects before random assignment. Thus, it is surprising that such a favorable population, which had a median survival time of 35 months with placebo, had a 21-month median survival time with the use of gefitinib. The authors and accompanying editorial have extensively discussed the possible imbalance in molecular and other characteristics, which could have negatively impacted the survival.1,3 However, they reasoned that such an imbalance was unlikely because of random assignment, and even if it was there, it could not possibly explain the consequent negative survival impact of gefitinib. In our opinion, survival discrepancy can be explained on the basis of molecular heterogeneity, not because it was disproportionately distributed between the two arms, but because it conferred differential survival risks as a result of treatment effect. Our inferences are based on a seminal work by Betensky et al,4 in which they systematically illuminated, through hypothetical statistical modeling, how failure to account for molecular heterogeneity could weaken the power of randomized studies and make it practically impossible to demonstrate a drug's real intrinsic efficacy for a minority in a diverse population. The S0023 trial was designed using a simple proportional hazards model to compute power and sample size, without taking into consideration the impact of molecular subtypes on treatment effect; hence, there is a possibility of inadvertent misspecification of regression model with consequent errors in sample size/power calculations and analysis.4 As elucidated in the following paragraph, the median survival of patients receiving gefitinib is dependent on molecular subtypes; hence, the correct statistical model should have specified the hazard of death as a function of treatment and molecular subtype. Even if a small subset of patients truly benefited from gefitinib, the trial was not designed to detect that difference. From available literature,5-7 we can assume that the SWOG S0023 population comprised the following three molecularly distinct groups based on potential outcome to gefitinib: 10% to 20% who had EGFR activating mutations, high EGFR copy number, and protein expression, and had survival benefit, who were termed responders; 20% to 30% who harbored KRAS mutations and had survival disadvantage, who were termed negative responders; and 50% to 70% who presumably lacked or had the previously mentioned markers but without any impact on survival, who were termed nonresponders. If all of these molecular subtypes were captured in the study population and random assignment took care of imbalance in known or unknown prognostic factors, then it is logical to assume that all of the three subgroups were proportionately distributed between the two arms. Then how do we explain for a survival differential? Under these circumstances, it is possible that the small subset of patients who benefited from gefitinib were masked by the much larger group of nonresponders, thus diluting the survival gains achieved in responders. At the best, this scenario would explain no difference in survival between the two arms; however, we did see a negative effect in the gefitinib arm, hinting at the presence of a significant number of negative responders in that population. It is quite possible that chemoradiotherapy contributed to the enrichment of the study population with negative predictors such as KRAS mutation that led to adverse survival when treated with gefitinib. Therefore, the already diluted survival benefit was pushed to negative in the gefitinib arm, whereas the placebo arm experienced no such detrimental effect, explaining the gap in survival between the arms. Because this negative impact on survival was not demonstrated in another randomized trial8 where gefitinib was used mostly in metastatic disease refractory or intolerant to chemotherapy, it can be argued that use of radiation in the SWOG S0023 trial in some way altered the epidermal growth factor receptor signaling pathway, conferring resistance to gefitinib. All of these considerations about molecular heterogeneity and its influence on outcome remain speculative, unless results on molecular analyses from this study are available. Yet another door to speculation could have been closed had information on poststudy treatments been collected. It is quite possible that gefitinib-resistant patients were inherently nonresponsive to other therapy, whereas placebo-treated patients still responded. In the era of targeted therapies, the challenge for the clinical investigators is to carefully design randomized trials if there is suspicion of unobserved heterogeneity among patients. Care should be taken to incorporate molecular characterization of the cohort as an important stratification variable to delineate the responders from nonresponders. Until this challenge is met, promising drugs will continue to be martyred on the altar of our own ignorance. AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. NOTES published online ahead of print at www.jco.org on September 8, 2008 REFERENCES
1. Kelly K, Chansky K, Gaspar LE, et al: Phase III trial of maintenance gefitinib or placebo after concurrent chemoradiotherapy and docetaxel consolidation in inoperable stage III non–small-cell lung cancer: SWOG S0023. J Clin Oncol 26:2450-2456, 2008 2. Bhutani M, Pathak AK, Guleria R, et al: A pilot study of sequential therapy with gefitinib following chemotherapy in advanced non-small cell lung cancer (NSCLC). Lung Cancer 49:S236-S237, 2005 (suppl 2) 3. Keedy VL, Arteaga CL, Johnson DH: Does gefitinib shorten lung cancer survival? Chaos redux. J Clin Oncol 26:2428-2430, 2008 4. Betensky RA, Louis DN, Cairncross JG: Influence of unrecognized molecular heterogeneity on randomized clinical trials. J Clin Oncol 20:2495-2499, 2002 5. Wheatley-Price P, Shepherd FA: Epidermal growth factor receptor inhibitors in the treatment of lung cancer: Reality and hopes. Curr Opin Oncol 20:162-175, 2008[Medline] 6. Sequist LV, Martins RG, Spigel D, et al: First-line gefitinib in patients with advanced non–small-cell lung cancer harboring somatic EGFR mutations. J Clin Oncol 26:2442-2449, 2008 7. Lynch TJ, Bell DW, Sordella R, et al: Activating mutations in the epidermal growth factor receptor underlying responsiveness of non-small-cell lung cancer to gefitinib. N Engl J Med 350:2129-2139, 2004 8. Thatcher N, Chang A, Parikh P, et al: Gefitinib plus best supportive care in previously treated patients with refractory advanced non-small-cell lung cancer: Results from a randomised, placebo-controlled, multicentre study (Iressa Survival Evaluation in Lung Cancer). Lancet 366:1527-1537, 2005[CrossRef][Medline]
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
|
