P04.10 * PREDICTIVE VALUE OF MGMT PROMOTER METHYLATION QUANTITATIVE ASSESSMENT IN GLIOBLASTOMA, WHICH OPTIMAL CUT-POINT?

Abstract

BACKGROUND: The methylation in the promoter region of the MGMT gene encoding the DNA repair protein O6-methyl guanine-DNA methyl transferase is a predictive marker for benefit from alkylating agent therapy and a prognostic factor in glioblastoma. Pyrosequencing-based assessement of MGMT promoter status is of particularly interest, but analytical methods and cut-points may varied from one center to another. We intend to establish a correlation between quantitative methylation of MGMT promoter values and overall survival (OS) or progression-free survival (PFS). METHODS: The clinical and biological data of 159 newly diagnosed glioblastoma recruited in Lille University Hospital between 2008 and 2011 were retrospectively collected. For the patients, the methylation status has been assessed by pyrosequencing on 5 CpG islands. Statistical analyses were performed using SAS software V9.3. RESULTS: Median age of the patients was 61 years. Sixty three patients were male. Of 159 patients, 156 underwent a surgery (40,4% of complete resection), 3 patients were diagnosed only by biopsies. 123 received concomitant radiotherapy and temozolomide and adjuvant temozolomide. The median PFS and the median OS were 9,9 and 15,9. The most significant cut-off methylation determined on 5 sites mean methylaion value was 12% for both predictive and pronostic evaluations. Site 1 methylation status was the best predictive value for both PFS and OS compared to the other CpG islands. On this site 1, a >3% methylation status was a significant predictive factor for reponse to standard treatment for PFS and a >25% site 1 methylation status was a strong pronostic factor for 2 years survival (OS) (47%). CONCLUSIONS: Even a low MGMT methylation (>3%) status on selected CpG site can already be a predictive factor for response to treatment in glioblastoma and allow us to reduce the cut-off values. Highest methylations status are correlated with the longest overall survival and therefore confirm the pronostic value of MGMT methylation. Establishing correlation between clinical data and biological assessment, such as pyrosequencing providing quantitative values for each CpG, would allow to lower and precise useful cut-points and to chose analysed site for a better predictive and pronostic significativity.