Temozolomide chemotherapy in newly-diagnosed adult glioblastoma: a systematic review using a computational evidence aggregation technique

Abstract

INTRODUCTION: Standard treatment for newly-diagnosed glioblastoma in adults consists of maximal surgical resection, followed by radiotherapy with concurrent and adjuvant temozolomide. Many trials have attempted to improve on this regimen. Conventional methods of metaanalysis are limited by problems with lack of transparency and interactivity. We conducted a systematic review using a novel computational method to synthesise and explore the robustness of evidence for new treatment options. METHOD: We identified all randomised controlled trials in newly-diagnosed adult glioblastoma which included temozolomide in at least one treatment arm. We excluded trials conducted exclusively in the elderly. We extracted data on treatment, overall survival, progression-free survival and toxicities, and assessed the quality of evidence using GRADE criteria. We used a computational evidence aggregation technique to summarise the results in a graphical treatment network. This shows which treatments are superior, the evidence supporting that assessment, and the effect of applying different subjective preference criteria regarding the relative importance of various outcome measures in defining superiority. RESULTS: We analysed data from 24 trials (4756 patients). There is robust evidence supporting the use of standard chemoradiotherapy (7 trials). We identify four additional treatment regimens which may confer a survival advantage: 1) addition of tumour-treating fields, 2) addition of bevacizumab, 3) addition of induction temozolomide, and 4) addition of dendritic cell vaccine. However, the evidence for three of these has significant weaknesses. CONCLUSION: Our computational evidence aggregation technique shows patterns of hierarchical superiority in temozolomide chemotherapy regimens for newly-diagnosed glioblastoma in adults. We identify several treatment regimens which may be superior to standard chemoradiotherapy. The best supported option is the addition of tumour-treating fields. In addition, we provide a novel graphical method for reviewing the evidence base. This allows clear visualisation of the available evidence, both for understanding the current evidence and planning future trials.