Table 2. Survival outcome analysis in patients with regional metastatic disease: Cox regression analysis of gene expression phenotypes.
| Univariable analysis | Multivariable analysis3 | Confounders3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Events/N | HR | 95% CI | P* | Events/N | HR | 95% CI | P* | ||
| Distant metastasis-free survival1 | |||||||||
| GEX phenotypes4 | 69/119 | 69/118 | |||||||
| High-immune | 1.00 | Ref | (0.01) | 1.00 | Ref | (0.02) | Gender, age | ||
| Pigmentation | 1.9 | 1.05-3.28 | 0.03 | 1.8 | 1.00-3.17 | 0.05 | |||
| Proliferative | 2.8 | 1.43-5.57 | 0.003 | 2.7 | 1.37-5.36 | 0. 004 | |||
| Gender | 72/125 | ||||||||
| Female | 1.00 | Ref | (0.02) | 1.00 | Ref | (0.006) | |||
| Male | 1.9 | 1.13-3.17 | 0.02 | 2.2 | 1.26-3.78 | 0.005 | |||
| Age (ys) | 72/124 | ||||||||
| <60 | 1.00 | Ref | (0.04) | 1.00 | Ref | (0.01) | |||
| ≥60 | 1.7 | 1.01-2.74 | 0.04 | 1.9 | 1.14-3.27 | 0.01 | |||
| Disease specific survival2 | |||||||||
| GEX phenotypes4 | 48/119 | 45/112 | |||||||
| High-immune | 1.00 | Ref | (0.009) | 1.00 | Ref | (0.06) | Gender, age, metastasis type | ||
| Pigmentation | 1.7 | 0.83-3.28 | 0.2 | 1.4 | 0.71-2.92 | 0.3 | |||
| Proliferative | 3.5 | 1.56-7.80 | 0.002 | 2.8 | 1.19-6.65 | 0.02 | |||
| Gender | 51/125 | ||||||||
| Female | 1.00 | Ref | (0.02) | 1.00 | Ref | (0.007) | |||
| Male | 2.2 | 1.13-4.11 | 0.02 | 2.8 | 1.34-5.90 | 0.006 | |||
| Age (ys) | 51/124 | ||||||||
| <60 | 1.00 | Ref | (0.05) | 1.00 | Ref | (0.06) | |||
| ≥60 | 1.8 | 1.01-3.34 | 0.05 | 1.8 | 0.97-3.49 | 0.06 | |||
| Metastasis type | 48/118 | ||||||||
| In-transit | 1.00 | Ref | (0.03) | 1.00 | Ref | (0.3) | |||
| Regional | 0.39 | 0.16-0.92 | 0.03 | 0.63 | 0.25-1.56 | 0.3 | |||
Abbreviations: CI, confidence interval.
1
Follow up starts at disease progression and ends at distant metastasis occurrence (=event).
2
Follow up starts at disease progression and ends at melanoma-specific death (=event).
3
The following confounders were included in the model: Gender, age (dichotomized at 60 years), and metastasis type (in-transit and regional). The confounders were selected based on their significance from the univariable analysis with P≤ 0.05.
4
Not including unclassified observations in the analysis.
*
P-values for the pairwise comparisons were calculated using the Wald-test. Overall P-values (also from the Wald-test) are given within the parentheses.