Table 4.
Univariate and multivariate survival analysis performed by Cox proportional hazard regression with factors from univariate analysis p > 0.05. The metric variables were dichotomized according to the median
| HR (95% CI) | p-value | |
|---|---|---|
| univariate | ||
| postoperative KPS, ≥ 70 | 4.25 (1.25–14.45) | 0.021 |
| postoperative MRC-NPS, ≥ 4 | 4.25 (1.25–14.45) | 0.021 |
| radiation therapy brain | 0.25 (0.08–0.80) | 0.019 |
| radiation therapy modality, whole brain | 5.40 (1.27–22.88) | 0.022 |
| preoperative tumor load, > 14.59 cm3 | 1.66 (0.66–4.19) | 0.283 |
| postoperative tumor load, > 0 cm3 | 1.03 (0.41–2.60) | 0.948 |
| multivariate | ||
| postoperative KPS, ≥ 70 | 0.16 (0.02–1.18) | 0.072 |
| postoperative MRC-NPS, ≥ 4 | a | |
| radiation therapy brain | a | |
| radiation therapy modality, whole brain | 4.36 (0.95–20.02) | 0.059 |
HR Hazard ratio, KPS Karnofsky performance status, MRC-NPS Medical Research Council Neurological Performance Score. a degree of freedom reduced due to constant or linear dependent covariates, constant or linearly dependent covariates postoperative MRC-NPS ≥ 4, postoperative KPS ≥ 70