Skip to main content
. 2019 Oct 2;51(10):115. doi: 10.1038/s12276-019-0319-y

Table 1.

Univariate and multivariate Cox regression analysis of disease-free survival in colorectal cancer (combined with CIT and AUS cohorts)

Variable Univariate Multivariate Multivariate (stepa)
n HR (95% CI) P-value n HR (95% CI) P-value n HR (95% CI) P-value
Gender (male vs. female) 782 0.803 (0.617–1.047) 0.105 721 0.78 (0.583–1.045) 0.096 721 0.783 (0.587–1.045) 0.097
Age (<75 or ≥75) 782 0.986 (0.736–1.321) 0.927 1.203 (0.858–1.685) 0.283
AJCC Stage (I, II, III, or IV) 782 2.837 (2.334–3.447) <0.001 2.267 (1.781–2.885) <0.001 2.327 (1.877–2.886) <0.001
Location (distal, proximal, or rectum) 782 0.765 (0.602–0.973) 0.029 0.809 (0.622–1.053) 0.114 0.823 (0.636–1.066) 0.14
Chemotherapy (No or Yes) 766 1.841 (1.404–2.414) <0.001 1.101 (0.789–1.536) 0.572
CMS subtype (CMS1, 2, 3, or 4) 736 1.274 (1.123–1.446) <0.001 1.134 (0.992–1.296) 0.066
PI (low-risk or high-riskb) 782 1.984 (1.514–2.6) <0.001 1.794 (1.327–2.426) <0.001 1.812 (1.342–2.448) <0.001

HR hazard ratio, CI confidence interval, CMS consensus molecular subtype, PI prognostic index

aA backward-forward step procedure was applied to optimize the multivariate model with the most informative variables

bPredicted outcome in Fig. 2b, c was used for analysis