Evidence from several studies supports that the epigenetic, transcriptional and translational regulation and expression of O6-methylguanine-methyltransferase (MGMT) is relevant for prognostic and predictive considerations in glioblastoma patients. MGMT status is being used as a stratifying factor or eligibility criterion in ongoing and accruing clinical glioblastoma trials. In some cases, there is also interest in MGMT assessment of glioblastoma tissue in the day-to-day clinical setting. However, a number of different methods and protocols have been used for MGMT analysis and it is unclear which methods harbour the greatest potential for translation into routine clinical use. This article reviews methods that have been used for MGMT assessment at DNA-, RNA- and protein-level in glioblastoma with a focus on their potential clinical utility. Conclusions. (1) DNA-based methods for MGMT analysis seem more promising for translation into the clinical setting than RNA- or protein-based methods. However, at present there is lack of data to base recommendations for a specific method or protocol for MGMT testing on. There is a strong need for systematic comparisons and validation of intra- and interlaboratory reproducibility and clinical performance of different methods for MGMT assessment to identify the best method for clinical MGMT testing. (2) The current practice of formalin-fixation of neurosurgical specimens considerably limits the spectrum of methods that can be applied for molecular diagnosis in clinical neuro-oncology. Further studies may be helpful to establish more appropriate protocols for tumour tissue preservation (e.g. identification of alternative fixatives that do not deteriorate DNA and RNA quality).