MPTH-25. THE PROGNOSTIC VALUE OF A NOVEL QUANTITATIVE MGMT PROMOTER METHYLATION SCORE FOR PATIENTS WITH GLIOBLASTOMA

Abstract

Evidence continues to solidify the central role molecular markers play in gliomas. Methylation status of MGMT promoter has been shown as a key marker in glioblastoma, predicting response to temozolomide. Traditional testing of MGMT promoter methylation, performed using a methylation specific PCR, yields a binary result; however it fails to consider the multiple potential methylation sites in MGMT’s promoter region. A novel quantitative bisulfite Sanger DNA sequencing assay evaluates methylation status at each of the 17 potential methylation sites in the promoter, resulting in a methylation score of 0-17. This novel assay was evaluated for its prognostic role in a cohort of 180 consecutive patients undergoing treatment for newly diagnosed glioblastoma at UCSF from April 2014 to August 2015. All patients received concurrent chemoradiation with temozolomide. Univariate and multivariate models for overall survival were built along with other established prognostic variables such as age, KPS and extent of resection. The methylation score was predictive of overall survival as a linear model. Stratification into partition groups revealed 3 groups (in order of increasing survival): A)Age >70, B)Age <70 and KPS <70 or Age <70 and KPS >70 and MGMT score <8 and C)Age <70 and KPS >70, and MGMT score >8. Brier score revealed lower predictive error for the methylation score compared to the traditional binary MGMT promoter methylation assessment, suggesting the methylation score is a more reliable model predicting survival. The methylation score, derived from the bisulfite sequencing analysis, provides a quantitative representation of the degree of methylation in the MGMT promoter. Survival models and partitioning support its role as a prognostic marker for overall survival, and the methylation score being superior to the traditional MGMT promoter methylation analysis in predicting survival in patients with newly diagnosed glioblastoma. Prospective validation and comparison to older/newer techniques including next-generation sequencing are warranted.