Mar 23 2016
By Shreeya Nanda
Progression-free survival (PFS) only moderately correlates with overall survival (OS) at the individual and trial level in patients with HER2-positive metastatic breast cancer treated with HER2-targeted agents, say researchers.
They believe that the lack of a stronger correlation between PFS and OS for anti-HER2 agents in this patient population "could be related to the crossover that was allowed in several of the included trials, the administration of 2nd or [3rd]-line treatments and the relatively long post-progression time, a feature not common to the studies of other malignancies where PFS has been validated as surrogate for OS".
This meta-analysis included individual patient data from 1839 HER2-positive patients enrolled in eight (six first-line) phase II or III randomised controlled trials assessing either trastuzumab or lapatinib, six of which assessed the agents as first-line treatments.
A total of 1462 patients experienced a PFS event while 1072 died from any cause during a median follow-up of 28 and 33 months, respectively. Median PFS was 5.7 months and median OS was 22.0 months.
PFS correlated moderately with OS at the individual patient level, with a Spearman coefficient of 0.67 corresponding to a squared correlation value of 0.45.
And at the trial level, the correlation between treatment effects on PFS and OS was also moderate, with a squared correlation value of 0.51, report Stefan Michiels (Institut Gustave Roussy, Villejuif, France) and colleagues.
They note that there is no consensus on the minimum trial-level squared correlation value to validate a surrogate endpoint and cutoffs ranging from 0.60 to 0.75 have been used in previous studies.
However, Michiels et al say that "[n]o matter what cut-off is chosen, based on the current results we cannot accept PFS as validated surrogate for OS."
And they conclude in the Annals of Oncology: "[T]he current findings provide only modest support for considering PFS as a surrogate for OS in HER2+ metastatic breast cancer; PFS does not completely substitute for OS in this setting."
Licensed from medwireNews with permission from Springer Healthcare Ltd. ©Springer Healthcare Ltd. All rights reserved. Neither of these parties endorse or recommend any commercial products, services, or equipment.