Research Article

A Novel Pseudogene Methylation Signature to Predict Temozolomide Outcome in Non-G-CIMP Glioblastomas

Table 2

Univariate and multivariate Cox regression analyses in non-G-CIMP GBMs with RT/TMZ or RT monotherapy.

VariablesUnivariate Cox modelMultivariate Cox model
HR95% CIvalueHR95% CI value

Combined discovery cohorts (RT/TMZ)
Patient age (increasing years)1.0371.011-1.0630.0051.0371.012-1.0630.003
The RISK-score signature (low vs. high)0.1990.106-0.372<0.0010.1800.092-0.350<0.001
MGMT methylation status (unmethylated vs. methylated)2.2031.232-3.9380.0081.8260.994-3.3550.052
Gene expression subtypes (nonproneural vs. proneural)1.2760.659-2.4720.469
Dataset source (TCGA vs. GSE60274)1.2230.729-2.0520.446
Combined discovery cohorts (RT monotherapy)
Patient age (increasing years)1.0250.989-1.0630.175
The RISK-score signature (low vs. high)2.3251.047-5.1660.038
MGMT methylation status (unmethylated vs. methylated)1.2740.650-2.4970.480
Gene expression subtypes (nonproneural vs. proneural)1.0330.416-2.5690.944
Dataset source (TCGA vs. GSE60274)1.6890.773-3.6930.189
RAUH cohort (RT/TMZ)
Patient age (increasing years)1.0321.003-1.0620.0291.0351.002-1.0690.039
Pre-adjuvant therapy KPS (≤ 70 vs. >70)1.3190.602-2.8870.489
Extent of surgery (biopsy vs. partial vs. total)1.0340.689-1.5500.872
The RISK-score signature (low vs. high)0.4410.249-0.7790.0050.5280.285-0.9810.043
TERT promoter mutation (no vs. yes)0.3670.144-0.9320.0350.5000.178-1.4040.188
MGMT methylation status (unmethylated vs. methylated)2.4231.334-4.4010.0042.6851.366-5.2770.004
Gene expression subtypes (nonproneural vs. proneural)1.0400.569-1.8980.889

RAUH = Rennes and Angers University Hospitals; TCGA = the Cancer Genome Atlas; G-CIMP = glioma-CpGs island methylator phenotype; MGMT = the O-6-methylguanine-DNA methyltransferase; GBM = glioblastoma; KPS = Karnofsky performance score; TMZ = temozolomide; RT = radiotherapy; TERT = telomerase reverse tranase. Italics were significant results.