{"entity":{"id":"gbm-agile","kind":"trial","name":"GBM AGILE","aka":["GBM-AGILE","Glioblastoma Adaptive Global Innovative Learning Environment"],"tldr":"GBM AGILE is the glioblastoma trial that never ends: new drugs enter, are tested against the same standard treatment in newly diagnosed and recurrent patients at once, and either graduate to a confirmatory stage or are dropped. Its first drug, regorafenib, did not help, which the trial established faster and more cheaply than a stand-alone study would have.","summary":"GBM AGILE was designed between 2015 and 2018 by a consortium of neuro-oncologists and statisticians, including Timothy Cloughesy, Brian Alexander, Meredith Buxton and Donald Berry, and is sponsored by the Global Coalition for Adaptive Research, a non-profit created to run it. The design paper appeared in Clinical Cancer Research in 2018 and the trial opened in July 2019 in the United States, since extended to Canada, Europe and China. Patients with newly diagnosed glioblastoma, split by MGMT promoter methylation, and patients with recurrent disease all enter one protocol; the control arm receives temozolomide with radiotherapy (newly diagnosed) or lomustine or another standard (recurrent). Randomisation is response-adaptive within each population, overall survival is the endpoint, and an arm that reaches a pre-set probability of success moves seamlessly into a second, confirmatory stage sized to support approval.\n\nThe first experimental arm, regorafenib, enrolled from 2019 and its results were reported at the 2024 European neuro-oncology meeting and published in the Journal of Clinical Oncology in 2026: regorafenib did not improve overall survival in any of the populations, closing the question raised by the small REGOMA trial. Paxalisib and VAL-083 completed their first stages without meeting the graduation criteria. Troriluzole, VT1021 and azeliragon entered later, and the platform continues to add arms.\n\nGBM AGILE has changed how glioblastoma drugs are tested: companies can enter a single arm without building their own registration trial, patients in newly diagnosed and recurrent settings are studied together, and negative answers arrive in about two years. Whether a graduated arm will be accepted for approval on platform data alone is the test the trial has not yet faced, and the JCO editorial on its regorafenib experience debates how much a shared, evolving control can be trusted over a decade.","status":"recruiting","asOf":"2026-09-17","links":[{"label":"ClinicalTrials.gov NCT03970447","url":"https://clinicaltrials.gov/study/NCT03970447"},{"label":"Global Coalition for Adaptive Research: GBM AGILE","url":"https://www.gcaresearch.org"},{"label":"Alexander et al., Clinical Cancer Research 2018: Adaptive Global Innovative Learning Environment for Glioblastoma, GBM AGILE","url":"https://doi.org/10.1158/1078-0432.CCR-17-0764"},{"label":"Evaluation of regorafenib in newly diagnosed and recurrent glioblastoma: GBM AGILE phase II/III Bayesian randomized platform trial, Journal of Clinical Oncology 2026","url":"https://doi.org/10.1200/JCO-25-01137"},{"label":"Lessons from GBM AGILE's regorafenib experience, Journal of Clinical Oncology 2026","url":"https://doi.org/10.1200/JCO-26-00444"}],"tags":[],"related":[],"cancers":["glioblastoma","brain-tumours"],"sections":[],"technologies":["kinase-inhibitors","digital-twins-trials"],"targets":[],"drugs":["regorafenib","temozolomide","lomustine"],"companies":[],"institutions":["ucla-jonsson"],"pathways":[],"terms":["master-protocol","seamless-adaptive","interim-analysis"],"trials":[],"people":["timothy-cloughesy"],"bottlenecks":["b-trial-design"],"keyPapers":[],"journals":[],"dependsOn":[],"notes":[],"nct":"NCT03970447","phase":"platform","setting":"Newly diagnosed glioblastoma (MGMT methylated and unmethylated) and recurrent glioblastoma: a Bayesian adaptive platform that screens experimental drugs against a shared standard-of-care control on overall survival and lets a winning arm expand seamlessly into a confirmatory stage","sponsor":"Global Coalition for Adaptive Research","result":"Regorafenib, the first arm, did not improve overall survival in newly diagnosed or recurrent glioblastoma; paxalisib and VAL-083 did not graduate; later arms continue.","yearReported":2024,"outcomes":[]},"route":"/trials/gbm-agile/","neighbours":{"cancer":[{"id":"brain-tumours","kind":"cancer","name":"Brain and spinal cord tumours (all types)","route":"/cancers/brain-tumours/"},{"id":"glioblastoma","kind":"cancer","name":"Glioma & glioblastoma","route":"/cancers/glioblastoma/"}],"technology":[{"id":"digital-twins-trials","kind":"technology","name":"Digital twins and virtual control arms","route":"/technologies/digital-twins-trials/"},{"id":"kinase-inhibitors","kind":"technology","name":"Small-molecule kinase inhibitors","route":"/technologies/kinase-inhibitors/"}],"drug":[{"id":"lomustine","kind":"drug","name":"Lomustine (CCNU)","route":"/drugs/lomustine/"},{"id":"regorafenib","kind":"drug","name":"Regorafenib","route":"/drugs/regorafenib/"},{"id":"temozolomide","kind":"drug","name":"Temozolomide","route":"/drugs/temozolomide/"}],"institution":[{"id":"ucla-jonsson","kind":"institution","name":"UCLA Jonsson Comprehensive Cancer Center","route":"/institutions/ucla-jonsson/"}],"term":[{"id":"interim-analysis","kind":"term","name":"Interim analysis, readout and data cut-off","route":"/terms/interim-analysis/"},{"id":"master-protocol","kind":"term","name":"Master protocol (platform, basket and umbrella trials)","route":"/terms/master-protocol/"},{"id":"seamless-adaptive","kind":"term","name":"Seamless, adaptive and Bayesian trial designs","route":"/terms/seamless-adaptive/"}],"person":[{"id":"timothy-cloughesy","kind":"person","name":"Timothy F. Cloughesy","route":"/people/timothy-cloughesy/"}],"bottleneck":[{"id":"b-trial-design","kind":"bottleneck","name":"Trial design, endpoints and cost","route":"/bottlenecks/b-trial-design/"}]}}