2051 Background: MGMT promoter methylation has been described as a prognostic marker in glioblastoma (GBM) and may be associated with chemosensitivity to alkylating agents. This study determined the feasibility of real-time determination of MGMT promoter methylation testing in a large international phase III clinical trial. Methods: Paraffin tumor blocks were obtained from patients registered onto RTOG 0525 then distributed to one of two central pathology labs: MD Anderson (Houston TX, K. Aldape) or CHUV (Lausanne, CH, R. Janzer). After histologic confirmation of GBM, unstained slides (40 microns of tumor tissue) were sent to the testing laboratories. Results were used for randomization into a treatment arm. MGMT methylation was assessed using methylation specific real-time PCR (MSP). The assay determined the number of copies of both methylated MGMT and of beta-actin (ACTB) in the sample. The ACTB copy number was used to assess the quality and quantity of the sample DNA. Results: Results were available for 995 samples. Following MSP samples were categorized into one of five possible results: failed (2, 0.2%), methylated (302, 30.2%), non-methylated (602, 60.2%), indeterminate (62, 6.2%), and invalid (27, 2.7%). Among cases that were evaluable as either methylated or unmethylated (n = 904) the MGMT promoter methylation rate in newly diagnosed GBM was approximately 1/3 (33.4%), a rate that is somewhat lower than prior reports. The average time from receipt of sample at the central pathology laboratory to reporting of results was 9.3 days. The time required decreased over the course of the clinical trial. This was due, in part, to training of the sites to deliver samples just before the start of runs. Conclusions: The results demonstrate the feasibility of performing real-time MGMT methylation testing, a tumor based assay, as a stratification factor in a multinational clinical trial. This study confirms that treatment decisions based on the molecular characteristics of the tumor are feasible, thereby providing opportunities to develop more molecularly-based tumor stratification or selection, a major advance in developing personalized treatment regimens. No significant financial relationships to disclose.