148 Using real-world data from GlioCova to assess the feasibility and revise the design of the ARISTOCRAT trial in recurrent glioblastoma

Résumé fourni par la source

Abstract Introduction ARISTOCRAT is a blinded trial of cannabinoids with standard-of-care temozolomide in recurrent glioblastoma, which initially aimed to recruit 234 patients over 18 months. Recruitment started on 03-Feb-2023, and it soon became apparent, from screening and recruitment data, that our initial site feasibility data was not reflected in actual recruitment. Estimating the eligible population size for studies is a recognised challenge. National cancer data is difficult to access and does not include the detail required to judge specific trial eligibility. In view of these problems, we explored alternative methods to estimate recruitment. Methods GlioCova is a comprehensive national dataset of all ∼56k adult primary brain tumour patients in England (2013-2018). We used this to estimate the number of eligible patients based on trial eligibility criteria for previous treatment (received sufficient adjuvant temozolomide and radiotherapy), IDH WT and MGMT methylation status. Results We estimated that <100 patients/year would be eligible for screening for ARISTOCRAT. Based on previous studies in this population, we would not anticipate more than 47% of screened patients to be randomised. Taken together, these data demonstrated that our initial recruitment targets were unrealistic. This prompted a statistical redesign of the trial from a rolling phase II/III to a conventional randomised phase II. This allowed revision of recruitment targets (96-120 patients) whilst maintaining 80% power to detect a 0.65 HR based on a 10% one-sided significance level within a practical and agreeable timeframe. Conclusions To the best of our knowledge, this is the first attempt to robustly address brain tumour trial recruitment using real-world data, and we believe this approach should be considered for future trials to avoid optimistic but unrealistic feasibility assessments that could lead to study failure.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.