Table 4.
Multivariate Cox hazard regression models.
| Overall Survival | Disease-Specific Survival | Disease-Free Survival | ||||
|---|---|---|---|---|---|---|
| Events | 38/76 | 14/69 * | 43/76 | |||
| Overall model fit | p < 0.001 | p = 0.028 | p < 0.001 | |||
| HR (95% CI) | p | HR (95% CI) | p | HR (95% CI) | p | |
| Age 65+ vs <65 | - | - | - | - | 3 (1.4–6.4) | 0.0042 |
| Gender M vs F | 3.1 (1.1–8.7) | 0.027 | - | - | - | - |
| Sm vs NSmND | 2.1 (0.4–10.7) | 0.360 | - | - | - | 0 |
| SmD vs NSND | 1.7 (0.7–4.1) | 0.206 | - | - | - | 0 |
| Stage III vs I–II | 1.6 (0.5–4.9) | 0.419 | 3.9 (0.4–38.6) | 0.244 | 1.2 (0.5–3.3) | 0.674 |
| Stage IV vs I–II | 1.1 (0.4–3.5) | 0.839 | 3.5 (0.4–31.9) | 0.265 | 1.4 (0.6–3.7) | 0.464 |
| Perivascular vs no invasion | 2.8 (0.8–9.6) | 0.099 | 3.9 (0.5–30.5) | 0.192 | 2.8 (1–8) | 0.049 |
| Both invasions vs no invasion | 4.3 (1.4–13.6) | 0.014 | 3.9 (0.6–27.8) | 0.169 | 4 (1.5–10.9) | 0.007 |
| Resection edge Not reported vs Clean (>5 mm) | 4.5 (1.5–13.6) | 0.008 | 8.3 (0.7–96.1) | 0.089 | 3.1 (1.2–8) | 0.019 |
| Resection edge Close (<5 mm) vs Clean (>5 mm) | 1.9 (0.7–4.7) | 0.190 | 5 (0.6–42.1) | 0.143 | 1.5 (0.6–3.7) | 0.345 |
| LN yield 1–10 vs 0 | 11.6 (1.2–9.7) | 0.033 | - | - | 10 (1.8–54.2) | 0.008 |
| LN yield 11–20 vs 0 | 5.3 (0.6–46.9) | 0.134 | - | - | 2.7 (0.6–13.4) | 0.216 |
| LN yield 21+ vs 0 | 4.7 (0.6–40.1) | 0.156 | - | - | 5 (1.1–23.7) | 0.041 |
| FC_miR-21 | 1.1 (1–1.2) | 0.011 | - | - | 1.1 (1–1.2) | 0.002 |
| FC miR-106b * | - | - | 1.3 (0.8–1.9) | 0.269 | - | - |
* for the DSS model, only miR-106b was significant in univariate models, thus this miRNA was selected for multivariate models despite reducing the total number of cases evaluated to 69 due to missing data for some cases. HR—hazard ratio, CI—confidence interval, Sm—smoker, NSmND—never smoker/never drinker, SmD—smoker and drinker, LN—lymph node, FC—fold change, NA—not applicable.