MRI-guided adaptive radiotherapy for high grade glioma (UNITED): a single-centre, single-arm, non-inferiority, phase 2 trial
Mené sur 98 patients atteints d'un glioblastome de haut grade (durée médiane de suivi : 14,2 mois ; 46 % de femmes), cette essai de phase II évalue la sécurité d'une radiothérapie adaptative guidée par IRM permettant de réduire les marges d'irradiation
Background: Radiotherapy volumes for patients with glioblastoma have remained unchanged for decades and result in large volumes of irradiated brain. We aimed to show the safety of a small-margin, MRI-guided adaptive radiotherapy approach for glioblastoma.
Methods: This single-arm, phase 2 trial was done at Sunnybrook Health Sciences Centre, Toronto, ON, Canada, and included patients with pathologically confirmed glioblastoma and planned for concurrent daily chemoradiotherapy up to 60 Gy in 30 daily fractions over 6 weeks (long course) or 40 Gy in 15 daily fractions over 3 weeks (short course). Eligible patients were adults (≥18 years) with an expected life expectancy greater than 12 weeks and an Eastern Cooperative Oncology performance status less than or equal to 2. A 5 mm clinical target volume (CTV) was applied, with the allowance of associated T2-weighted fluid-attenuated inversion recovery hyperintense regions in a personalised way at the discretion of the treating physician. All patients were treated with 1·5 T MRI–guided linear accelerator (MR-Linac) incorporating weekly gadolinium-enhanced online adaptive fractions. The primary outcome was the risk of marginal failure powered for non-inferiority with a margin of 10% compared with a historical control. All patients were included in the primary and safety analyses. This study is registered with ClinicalTrials.gov, NCT04726397.
Findings: Between April 28, 2021, and May 12, 2023, 109 patients were assessed for eligibility and 98 were enrolled and received treatment on the UNITED protocol with 59 [60%] patients receiving long-course radiotherapy and 39 [40%] receiving short-course radiotherapy. Median follow-up was 14·2 months (IQR 8·9–19·7). 53 (54%) of 98 patients were male and 45 (46%) were female. The observed risk of marginal failure was 4% (95% CI 0–8). The most common grade 3–4 adverse events were lymphopenia (eight [11%] of 72 patients without baseline lymphopenia) and thrombocytopenia (four [4%] of 98 patients). Serious adverse events occurred in one (1%) patient due to febrile neutropenia and there were no treatment-related deaths.
Interpretation: MRI-guided adaptation for glioblastoma enables margin de-escalation and resulted in a low rate of marginal failure. A randomised trial comparing this technique to a standard large-margin radiotherapy approach will establish whether toxicity and quality-of-life improvements can be realised.
The Lancet Oncology , résumé, 2026