P01.123 Matched-pair analysis of long- versus short-term survivors of glioblastoma: molecular markers make the difference

Abstract

BACKGROUND: A small population of patients diagnosed with glioblastoma (GB) survives long-term. We performed a matched-pair analysis of patients that survived three years or longer and patients that survived less than three years. MATERIAL AND METHODS: We performed a cross-institutional review of 269 patient records and matched a group of 93 patients surviving 3 years or longer with 176 patients that survived less than 3 years (STS), whereas 83 of all LTS had 2 matches (LTS-2m), and 10 had 1 match (LTS-1m). Matching was performed according to age, Karnofsky performance Status (KPS), initial therapy (surgery/biopsy) and adjuvant therapy (chemo-radiotherapy/maintenance chemotherapy). Factors of interest were tumor location, extent of resection (GTR/STR), radiation dose, GFAP (expressed/partially expressed/not expressed), MGMT promoter hypermethylation status as well as IDH-1 mutation status and ki67 index. RESULTS: There were no statistically relevant differences in tumor location, extent of resection, radiation dose or ki67 index between STS and LTS. In turn, the incidences of MGMT promoter hypermethylation (STS vs. LTS-2m:p=0.008; STS vs. LTS-1m: p=0.013) and IDH-1 mutations (STS vs. LTS-2m: p=0.008; STS vs. LTS-1m: p=0.046) were significantly higher in LTS. GFAP expression, which is frequently lost with increasing grade of de-differentiation, was also seen significantly more often in LTS than in STS (STS vs. LTS-2m: p=0.002; STS vs. LTS-1m: p=0.019). CONCLUSION: This cross-institutional analysis corroborates the prognostic roles of molecular markers such as MGMT and IDH-1 over all other factors that may partially depend on caregivers.