Trends in Phase III Trials in Advanced NSCLC: Decline in Use of Overall Survival as Primary Endpoint, More ‘Positive’ Outcomes
In a study published in the Journal of Clinical Oncology, Sacher et al identified trends in reporting of phase III trials in advanced non–small cell lung cancer (NSCLC) that include reduced use of overall survival as the primary endpoint and an increase in claims of positive outcomes despite absence of statistical significance in the primary outcome.
Reduced Use of Overall Survival
This analysis of reporting trends included a total of 201 phase III trials of systemic therapy in patients with advanced NSCLC conducted between 1980 and 2010. The number of such trials increased from 32 in 1980 to 1990 to 53 in 1991 to 2000 and 118 in 2001 to 2010, with median sample size increasing from 152 to 184 and 413, respectively (P < .001). The use of overall survival as the primary endpoint has decreased from 97% of trials in 1980 to 1990 and 96% in 1991 to 2000 to 81% in 2001 to 2010 (P = .002), with progression-free survival being used as the primary endpoint in 13% of trials in the latter period.
‘Positive’ Outcomes Without Statistical Proof
In the period 2001 to 2010, 53% of trials were reported as positive, compared with 31% during 1980 to 1990. The proportions of trials showing statistically significant improvement in the primary endpoint has remained stable over time, including 29% of trials in 1980 to 1990 and 31% in 2001 to 2010. However, the proportion of trial reports claiming a positive outcome despite the absence of statistical significance in the primary outcome has increased significantly from 30% in 1980 to 1990 and 24% in 1991 to 2000 to 53% in 2001 to 2010 (P < .001).
Positive outcomes in these trials were claimed on the basis of improvements in secondary endpoints, a claim of noninferiority despite absence of a statistically appropriate noninferiority design, or recommendations for further study on the basis of nonsignificant trends in primary endpoints. The only trials that were reported as negative because of insufficient magnitude of clinical benefit despite statistical significance were two conducted in 1980 to 1990.
Among trials reporting statistically significant improvement in overall survival, the median net survival benefit was 3.9 months in 1980 to 1990, 2.4 months in 1991 to 2000, and 2.5 months in 2001 to 2010 (P = .11). Among all trials that claimed positive outcome, median net survival benefit was 3.9, 2.0, and 0.9 months, respectively (P < .001).
The investigators concluded, “A significant shift has occurred over the past three decades in the design and interpretation of phase III trials in advanced NSCLC. The use of survival as the primary measure of benefit is declining, as is the magnitude of benefit deemed clinically relevant…. Clear standards of acceptable minimal clinical differences in advanced NSCLC and appropriate clinically relevant outcomes should be established and endorsed.”
Natasha B. Leighl, MD, MMSc, of Princess Margaret Cancer Centre/University Health Network, is the corresponding author for the Journal of Clinical Oncology article.
The study was supported by the Division of Medical Oncology/Hematology, Princess Margaret Cancer Centre/University Health Network. The study authors reported no potential conflicts of interest.
The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.
