Skip to main content

Table 3 Three survival prediction models

From: Identification and validation of a novel survival prediction model based on the T-cell phenotype in the tumor immune microenvironment and peripheral blood for gastric cancer prognosis

 

Regression coefficient∀

Formula

Cut-off&

Model 1

   

 N stage

0.374

1 − S(t)*^exp(w(t)$ − 2.2480#)

0.589

 PD-1+CD8+TILs (%)

1.768

 PD-L1+CD8+TILs (%)

2.518

Model 2

   

 N stage

0.476

1 − S(t)^exp(w(t) + 0.5475#)

0.694

 PD-L1+CD8+PBLs (%)

1.461

 TIM-3+CD8+PBLs (%)

− 2.011

Model 3

   

 TIM-3+CD8+PBLs (%)

− 2.006

1 − S(t)^exp(w(t) − 1.2014#)

0.508

 PD-L1+CD8+TILs (%)

2.821

 PD-1+CD8+TILs (%)

1.960

  1. ∀ Regression coefficient was determined by multivariate Cox regression analysis
  2. *S(t) was defined as three-year survival rate. We calculated the survival rate to be 0.574
  3. $w(t) = \(\sum_{i=0}^{p}\mathrm{\beta iXi}\) βi = regression coefficient for each parameter, Xi = The detection value or classification value of the corresponding parameter
  4. #The constant was calculated according \(\sum_{i=0}^{p}\mathrm{\beta i \overline{X}i}\)
  5. &Cut-off point was the corresponding value of the maximal Youden index. Youden index = sensitivity + specificity − 1