Table 2.
Cox proportional hazards 10-year recurrence-free survival models
| n (%) event# | n (%) no-events# | Event occurrence months | Multivariable$ model | |||
|---|---|---|---|---|---|---|
| Mean | Median | HR (95% CI) | P value | |||
| Age | 373 (61.9) | 230 (38.1) | 65.6 | 62.0 | 1.01 (1.00–1.03) | 0.024 |
| Tumor grade | ||||||
| 1 + 2 | 263 (58.2) | 189 (41.8) | 69.9 | 79.0 | Referent group | |
| 3 + 4 | 110 (72.8) | 41 (27.2) | 52.8 | 26.7 | 1.48 (1.17–1.87) | 1.0E−3 |
| Smoking status | ||||||
| Non-smoker | 47 (45.6) | 56 (54.4) | 82.3 | 120.0 | Referent group | |
| Ever-smoker | 326 (65.2) | 174 (34.8) | 62.2 | 51.9 | 1.65 (1.22–2.25) | 1.0E−3 |
| mdNLR* | 373 (61.9) | 230 (38.1) | 65.6 | 62.0 | 1.12 (1.04–1.20) | 3.0E−3 |
HR: hazard ratio, CI: confidence interval, mdNLR: methylation-derived neutrophil to lymphocyte ratio
Stratification was used on sex and BCG treatment status for proportional assumption
$The model controlling for age, sex, tumor grade, smoking status, BCG treatment status, and mdNLR
*Winsorization was used on the top 2% values for fitting linearity assumption
#Initial recurrence or the death whose cause was unknown