Convection-enhanced delivery of a transforming growth factor- 2 inhibitor trabedersen for recurrent high-grade gliomas: efficacy real or imagined?, in reference to Bogdahn et al. (Neuro-Oncology 2011;13:132-142)

Abstract

Dr Bogdahn et al. reported on the convection-enhanced delivery (CED) of the transforming growth factor-β2 (TGF-β2) inhibitor trabedersen for recurrent high-grade glioma.1 The trial was designed as a randomized phase 2b study comparing two doses of trabedersen to chemotherapy, either temozolomide (TMZ) or procabazine-lomustine-vincristine (PCV). Several comments regarding the study I believe are warranted. Firstly, the trial was never properly powered to detect a difference in trabedersen dose vs. best standard chemotherapy. Notwithstanding this serious limitation, the trial was clearly negative with respect to recurrent glioblastoma. Much more problematic are the conclusions with respect to recurrent anaplastic astrocytoma (AA) in which the authors assert there is compelling evidence of trabedersen effectiveness to proceed to a randomized phase 3 trial (SAPPHIRE). Concerning in particular is the remarkably small numbers of recurrent AAs studied (39 randomized to 1 of 3 treatment arms) from which these conclusions are based. Furthermore, efficacy of trabedersen is based upon a secondary endpoint (progression-free survival at 14 months: PFS-14) not the prespecified primary endpoint (progression-free survival at 6 months: PFS-6). To my knowledge, PFS-14 has never been validated as a glioma trial endpoint and, consequently, it is uncertain what this construct represents.2,3 Remarkably, from 1/3 to 2/3 of all patients with recurrent AA had not been treated with chemotherapy before study entry, creating differences in the treatment groups and in effect subdividing an already small group of patients into 2 separable populations (chemotherapy-experienced or -naive). At least in the United States, it would be highly unlikely to find such a chemotherapy-naive population, as TMZ is widely used in the upfront treatment of such tumors. How to assess a study performed in non-traditional neuro-oncology centers where standards of care likely differ from European and U.S. centers are challenging, though admittedly is becoming more common as financial and regulatory considerations shift new drug studies offshore. What was not a study inclusion criteria, prespecifying that all recurrent AA patients had been treated with either PCV or TMZ would have resulted in more comparable subgroups, a design utilized in the German NOA-04 recurrent anaplastic glioma trial.4 Additionally and not stated was the salvage therapy employed after study treatment failure, a highly relevant issue with respect to recurrent AA, as salvage therapy has a significant impact in overall survival. Confusingly, the recurrent AA treatment groups had no significant difference in tumor control at 6 months (presumably though not explicitly stated PFS-6) but did differ with respect to median duration of response favoring trabedersen. There also was an improvement favoring low-dose trabedersen in median time to progression. Notwithstanding the considerable investment in the trial, no central radiology review was performed, a process that is blinded to treatment allocation and without bias (unlike investigator-assessed response determinations). Without specifying endpoint definitions these descriptions appear confusing and contradictory. Not mentioned lastly is the prior lack of success utilizing CED as a means of drug delivery exemplified by the PRECISE and TransMID trials for recurrent glioblastoma. The failure of these trials begs the question of whether CED is a practical or feasible drug delivery system. What appears needed based upon this study is a properly designed and statistically powered phase 2 study of trabedersen in the treatment of recurrent AA that utilizes acceptable endpoints and a similarly pretreated cohort of patients. This critique is not intended to diminish the considerable effort by the study investigators but rather draw attention to prematurely moving investigational agents into randomized phase 3 trials without first establishing robust phase 2 trial results.