BACKGROUND: Molecular testing for MGMT promoter methylation has become of clinical importance in the diagnostic assessment of malignant gliomas since test results may guide therapeutic decision making, in particular in elderly patients with glioblastoma. However, the patterns and extent of MGMT promoter methylation may vary from tumor to tumor, and standardized approaches for its routine diagnostic assessment are lacking. Thus, external quality assessment (EQA) measures are required to ensure accuracy and reproducibility of results across different laboratories. METHODS: We performed an interlaboratory comparison of MGMT promoter methylation analysis involving twenty-three academic institutions in Germany, Austria and the Netherlands. Two different test rounds were carried out, the first one using high molecular weight DNA extracted from frozen tissue samples of 20 tumors and the second one using formalin-fixed paraffin-embedded tissue sections of 16 tumors. All samples were centrally retrieved from the CNS tumor tissue bank at Heinrich Heine University Düsseldorf. Each participating center evaluated the same set of samples using the locally established methods. Results were centrally collected, together with information on the individual assays and the number of tests carried out per year. RESULTS: Methylation specific-PCR was the most commonly used method at the participating centers. Other less common techniques included pyrosequencing of bisulfite-modified DNA, MethyQESD (methylation-quantification of endonuclease-resistant DNA), MLPA (Multiplex Ligation-dependent Probe Amplification), and PCR-based fragment analysis. MGMT testing results showed a good overall concordance across the participating laboratories for those tumors that either had strongly methylated or clearly unmethylated MGMT promoter sequences. However, poor concordance was obtained for cases with only weak or partial MGMT promoter methylation. CONCLUSIONS: Our study provides an overview of the current practice of MGMT testing in glioma diagnostics in a large number of different institutions. The results underscore the importance of EQA measures to assure optimal quality and interlaboratory reproducibility of test results in order to avoid wrong treatment decisions. The assessment of tumors with only partialy or weakly methylated MGMT promoter sequences is problematic and may need to be improved by testing algorithms involving more than a single technique. SECONDARY CATEGORY: Clinical Neuro-Oncology.