New Molecular-Based Prognostic Model for Glioblastoma in Temozolomide Era

Key Points

  • A new molecular-based recursive partitioning analysis model was identified, consisting of MGMT protein, c-Met protein, and age.
  • Median overall survival durations in the model’s 3 risk classes were 21.9, 16.6, and 9.4 months in the NRG Oncology RTOG 0520 cohort.

In a study reported in JAMA Oncology, Bell et al derived a molecular-based recursive partitioning analysis model for overall survival in glioblastoma multiforme in the temozolomide era with the aim of refining existing clinically based models.

Study Details

The study involved analysis of 452 specimens from the NRG Oncology RTOG 0525 study for key protein biomarkers in glioblastoma multiforme using a quantitative molecular microscopy–based approach with semiquantitative immunohistochemical validation. Prognostic models were investigated by incorporating protein biomarkers associated with overall survival into a recursive partitioning analysis model consisting of current RTOG recursive partitioning analysis clinical variables (age, Karnofsky performance status, extent of resection, and neurologic function).

Identification and Performance of Novel Model

In the 452-specimen set, markers found to be significantly associated with overall survival on single-marker multivariate analysis consisted of: MGMT (O-6-methylguanine-DNA methyltransferase; hazard ratio [HR] = 1.81, P < .001), survivin (HR = 1.36, P = .02), c-Met (HR = 1.53, P = .02), pmTOR (HR = 0.76, P = .03), and Ki67 (HR = 1.40, P = .007).

Analysis in 166 patients receiving radiation therapy and temozolomide in NRG Oncology RTOG 0520 with data available for all significant proteins showed that higher and lower MGMT protein levels were significantly associated with decreased and increased MGMT promoter methylation (P < .001) and that MGMT protein expression (HR = 1.84, P < .001) had higher overall survival prognostic value than MGMT promoter methylation (HR = 1.77, P < .001).

Recursive partitioning analysis modeling in this population identified a new model (NRG-GBM-RPA) consisting of MGMT protein, c-Met protein, and age, which yielded greater separation of overall survival prognostic classes vs the existing recursive partitioning analysis model and MGMT promoter methylation: class I = MGMT tumor less than median value or MGMT tumor at least median and age < 50 years; class II = MGMT tumor at least median, age ≥ 50 years, and c-Met cytoplasm less than top quartile; and class III = MGMT tumor at least median, age ≥ 50 years, and c-Met cytoplasm at least top quartile. Median overall survival for these 3 classes were 21.9, 16.6 (HR = 1.83, P = .004, vs class I), and 9.4 months (HR = 5.19, P < .001, vs class I).

The prognostic ability of the NRG-GBM-RPA model was confirmed in an independent data set of 176 patients using traditional immunohistochemistry. Among all patients, hazard ratios were 1.46 (P = .04) for class I vs II and 1.88 (P = .006) for class I vs III. Among 87 patients who received radiation therapy and temozolomide in this cohort, hazard ratios were 1.91 (P = .02) for class I vs II and 3.68 (P < .001) for class I vs III.

The investigators concluded: “This new NRG-GBM-RPA model improves outcome stratification over both the current RTOG [recursive partitioning analysis] model and MGMT promoter methylation, respectively, for patients with [glioblastoma multiforme] treated with radiation and temozolomide and was biologically validated in an independent data set. The revised [recursive partitioning analysis] has the potential to contribute to improving the accurate assessment of prognostic groups in patients with [glioblastoma multiforme] treated with radiation and temozolomide and to influence clinical decision making.”

The study was supported by grants from the National Cancer Institute, a Brain Tumor Funders collaborative grant, an award from The Ohio State University Comprehensive Cancer Center, and a grant from Merck & Co.

Arnab Chakravarti, MD, of The Ohio State University Comprehensive Cancer Center, is the corresponding author of the JAMA Oncology article.

The content in this post has not been reviewed by the American Society of Clinical Oncology, Inc. (ASCO®) and does not necessarily reflect the ideas and opinions of ASCO®.


Advertisement

Advertisement



Advertisement