Skip to main content
. 2020 Aug;9(8):4762–4769. doi: 10.21037/tcr-19-2039

Table 4. Univariate and multivariate analysis of cancer-related survival.

Risk factor Univariate analysis multivariate analysis
P-valuea Relative risk (95% CI) P-valueb
MVDe low vs. high 0.001 1.8 (1.0–3.2) 0.001
CEACAM1f expression low vs. high 0.048 c 0.362d
pT1–pT2 vs. pT3–pT4 0.064 c 0.089
pN0 vs. pN+ 0.045 c 0.115
G1 vs. G2–G3 0.229 c 0.325
male vs. female 0.354 c 0.562
≤63 years vs. >63 years 0.031 c 0.052

a, P-values of univariate analyses were determined by log-rank test; b, Stepwise multivariate analysis was performed using the Cox proportional-hazard model; c, no estimate of relative risk is given, since the variable was not significant on multivariate analysis; d, the prognostic impact of CEACAM1 expression is dependent on the prognostic impact of MVD, because when CEACAM1 expression enters multivariate analysis with MVD, the significance of CEACAM1 expression observed by univariate analysis disappears significantly. While MVD itself was still a significant prognostic factor for adverse cancer-related survival; e, MVD was categorized according to the median MVD of all tumors into low (<30 microvessels/400× microscopic field of view) and high (≥30 microvessels/400× microscopic field of view); f, CEACAM1 expression was categorized into low (<66% of CEACAM1 positive tumor cells) and high (≥66% of CEACAM1 positive tumor cells).