Gliosarcoma vs. glioblastoma: a retrospective case series using molecular profiling

BMC Neurology

Table 5 Logistic regression models, sorted by p value

Model term	Method	p (Wald)	OR	95% CI
Uni-variable models withp<0.05
PD-L1	IHC	0.0025	2.7	1.7 — 4.8
EGFR	CNA	0.0069	0.14	0.03 — 0.54
NF1	NGS	0.015	2.9	1.4 — 6.0
PD-1	IHC	0.017	3.1	2.2 — 5.5
NTRK1	F-RNA	0.022	27	2.0 — 507
LYL1	CNA	0.022	26	1.9 — 498
IDH2	NGS	0.026	24	1.8 — 465
PTCH1	NGS	0.026	24	1.3 — 159
EGFRvIII	FFA	0.04	0.22	0.06 — 0.95
Best two-variable model
Intercept		<0.001	0.04	0.01 — 0.15
EGFR	CNA	0.084	0.26	0.06 — 1.1
PD-1	IHC	0.072	4.1	1.0 — 17.3

Uni-variable models with p<0.05; values for p and OR not shown for the intercept term in these models. The best two-variable model is also shown
Abbreviations: Method = See abbreviations in Table 1, p Probability (p value, Wald statistic), OR Odds ratio, 95% CI 95% confidence interval for OR

ISSN: 1471-2377