Table 2.
Cox regression model
| Variable | Coefficient | SE | χ2 statistics | P‐value | Hazard ratio | ||
|---|---|---|---|---|---|---|---|
| Estimate | 95% confidence interval | ||||||
| Lower limit | Upper limit | ||||||
| LN‐ad † | 0.83 | 0.41 | 4.02 | 0.0450 | 2.28 | 1.02 | 5.12 |
| Stage | 0.81 | 0.29 | 7.61 | 0.0058 | 2.25 | 1.27 | 4.00 |
| IFN‐γ | –1.25 | 0.46 | 7.31 | 0.0068 | 0.29 | 0.12 | 0.71 |
Three sets of parameters on patient survival were estimated by the Cox regression model. The result of the model that gave minimum Akaike's Information Criterion (AIC) is shown here.
†
The minimum value of AIC suggested that the most appropriate model was with lymphadenectomy (LN‐ad) being reduced to two categories, i.e. ‘complete LN‐ad’ and ‘LN‐ad or partial LN‐ad’, although stages were not reduced.