Low Levels of Quantified MGMT Methylation Status Impart Positive Impact on Survival

Abstract

MGMT methylation status is a significant prognostic factor for glioblastoma (GBM), particularly in the era of temozolomide (TMZ). However, there is no standardization of MGMT methylation methodology or optimal cutoff criteria. We developed and implemented a quantitative assay for MGMT methylation (mMGMT) and investigated association with clinical outcomes. We retrospectively reviewed GBM patients at our institution treated with standard chemoradiation, from 2008 to 2018. DNA was extracted from formalin-fixed paraffin-embedded (FFPE) tissue. A real-time PCR assay was performed to amplify a 62 base-pair region of MGMT including the first 12 CpG islands after ATG initiation codon for both methylated and unmethylated alleles. PCR products underwent high-resolution melt (HRM) analysis and the fraction of methylated DNA was determined by comparison with a standard curve generated from specimens with known percentages of methylation. Patient demographics, treatment records, and clinical outcomes were collected. Kaplan-Meier and log-rank tests were performed to compare survival. Cox proportional hazard model was used for predicting prognostic factors. Optimal cutoff point was evaluated with maximally selected log-rank statistics. Statistical analysis was conducted on R (v3.5). A total of 108 patients (68 men, 40 women) with newly diagnosed GBM were included in the study. Median age was 61 years old (31-85), and KPS was 80 (50-100). A total of 32 pts (29.6%) had gross total resection (GTR), 61 pts (56.4%) had subtotal resection, and 15 pts (13.9%) had biopsy only. With a median follow-up of 9 months (1-73 months), the median PFS and OS were 6 months and 14.4 months, respectively. At 1 and 2 years, PFS was 25.2% and 17.8% while OS was 58.1% and 23.8%, respectively. A total of 44 patients (40.7%) had detectable MGMT methylation, with median methylation level of 23.36% (range 0.2%-78.2%; IQR 4.77-31.88%). On both univariate and multivariate analysis, detectable MGMT methylation (HR 0.06, 95% CI [0.008-0.4223]; p=0.004) and GTR (HR 0.18, 95% CI [0.07-0.43]; p>0.001) were associated with improved OS, while GTR alone improved PFS (HR 0.38, 95% CI [0.16-0.89]; p=0.028). Patients with any detectable mMGMT had improved OS (19.3 vs 11.1 months; p=0.003). Optimal cutoff point for quantifying MGMT methylation for maximum survival difference was estimated as ∼ 1.9% (p=0.007). MGMT methylation at even low levels may have a positive impact on survival for GBM patients. Further refinement of a cutoff point for our MGMT assay with a validation set is warranted.