A Risk Stratification Model for Lung Cancer Based on Gene Coexpression Network and Deep Learning

Table 1

Univariate and multivariate Cox regression analysis of the risk stratification model and clinicopathological variables.

Variables	Univariate analysis		Multivariate analysis
	Hazard ratio (95% CI)	value	Hazard ratio (95% CI)	value
Training set
Gene network prognostic score, high risk group	2.51 (1.65–3.81)	<0.001	3.057 (1.556–6.006)	0.001
Age, older than 60	1.66 (0.96–2.86)	0.068
Sex, male	1.28 (0.86–1.88)	0.221
Smoking status, ex-smoker	0.59 (0.31–1.12)	0.108
Smoking status, never smoker	0.51 (0.22–1.19)	0.120
T stage: II	2.50 (1.31–4.79)	0.006	1.266 (0.599–2.674)	0.537
T stage: III	13.32 (2.89–61.32)	0.001	5.895 (1.189–29.237)	0.030
N stage: I	2.27 (1.28–4.05)	0.005	1.762 (0.943–3.294)	0.076
Test set 1
Gene network prognostic score, high risk group	4.39 (1.92–10.06)	0.0004	2.97 (1.25–7.09)	0.01
Age, older than 60	1.27 (0.65–2.48)	0.49
Sex, male	1.52 (0.78–2.96)	0.22
Smoking status, never smoker	0.61 (0.31–1.19)	0.15
Stage: II	4.23 (2.17–8.24)	0.00002
EGFR mutation +	0.47 (0.24–0.93)	0.03	2.74 (1.36–5.54)	0.005
KRAS mutation +	0.87 (0.27–2.85)	0.82	0.64 (0.32–1.27)	0.20
Test set 2
Gene network prognostic score, high risk group	1.81 (0.98–3.36)	0.06
Age, older than 60	1.33 (0.73–2.43)	0.35
Sex, male	0.83 (0.40–1.75)	0.63
Stage: T2	1.65 (0.84–3.25)	0.14