Table 5. Univariate and multivariate cox proportional hazards models for the predictors of overall survival (results from mean value cutoffs).
Univariate analysis
|
Multivariate analysis
a
|
|||||
---|---|---|---|---|---|---|
Variables | HR | 95% CI | P-value | HR | 95% CI | P-value |
CD3b | 0.627 | 0.410–0.627 | 0.031 | 0.813 | 0.523–1.264 | 0.358 |
CD8b | 0.583 | 0.358–0.950 | 0.030 | 0.574 | 0.347–0.949 | 0.031 |
CD45ROb | 0.530 | 0.331–0.850 | 0.008 | 0.612 | 0.377–0.995 | 0.048 |
CD3CD8c | 0.522 | 0.303–0.900 | 0.019 | 0.572 | 0.325–1.010 | 0.054 |
CD3CD45ROc | 0.441 | 0.252–0.772 | 0.004 | 0.565 | 0.317–1.009 | 0.054 |
CD8CD45ROc | 0.416 | 0.221–0.785 | 0.007 | 0.436 | 0.227–0.838 | 0.013 |
CD3CD8CD45ROd | 0.393 | 0.202–0.766 | 0.006 | 0.471 | 0.236–0.937 | 0.032 |
CI=confidence interval; HR=hazard ratio.
The results were according to the density groups of TIL with mean value cutoffs.
Univariate and multivariate Cox proportional hazards models were used as statistical methods.
Tumour stage (T stage), histologic classification (WHO), presence of lymph node metastasis, and presence of lymphatic invasion which were proved to have a prognostic significance in univariate analysis were adopted as covariates in each multivariate analysis.
High vs low.
High/high vs low/low.
High/high/high vs low/low/low.