This article is more than 5 years old. Information may no longer be current.
PFS predicted OS in metastatic melanoma trials
PFS appeared to be a valuable surrogate for OS in randomized, dacarbazine-controlled trials of patients with metastatic melanoma, results of a meta-analysis suggest.
Keith T. Flaherty, MD, of the department of hematology and oncology at Massachusetts General Hospital, and colleagues systematically reviewed randomized metastatic melanoma trials in which dacarbazine was used as the control arm. They identified trials that reported PFS and OS with a standard HR.
Keith T. Flaherty
The analysis included 12 trials with a combined 4,416 patients. Flaherty and colleagues compared HRs for OS and PFS, and also performed sensitivity analyses based upon the presence of crossover, trial size and dacarbazine dose.
Results showed a strong correlation between the treatment effects for PFS and OS, which appeared independent of treatment type.
Pearson correlation coefficients were 0.71 (95% CI, 0.29-0.90) with random-effects assumption, 0.85 (95% CI, 0.59-0.95) with fixed-effects assumption, and 0.89 (95% CI, 0.68-0.97) with sample-size weighing.
When the researchers limited their analysis to nine trials that did not include treatment crossover, the correlation coefficient was 0.96 (95% CI, 0.81-0.99). This decreased to 0.93 (95% CI, 0.74-0.98) with the addition of trials that included less than 50% crossover. The PFS-to-OS correlation decreased to 0.55 (95% CI, 0.03-0.84) after researchers included mature follow-up data after at least 50% crossover in phase 3 trials for vemurafenib (Zelboraf, Genentech) and dabrafenib (Tafinlar, GlaxoSmithKline).
The PFS-to-OS correlation was 0.85 (95% CI, 0.51-0.96) when trials with no or little crossover were included and a random-effects assumption was used.
Disclosure: The researchers report no relevant financial disclosures.