Table 3. Multivariate analyses of variables which affect survival in patients with Hodgkin lymphoma (HL) who developed non-small cell lung cancer (NSCLC).
Localized NSCLC | Regional NSCLC | Distant NSCLC | |
Multivariate analysis | |||
Age at HL diagnosis | ND | ND | |
P value | 0.68 | ||
Hazard ratio | 0.98/year | ||
95% confidence interval | 0.91-1.06 | ||
HL subtype (mixed cellularity: no)‡ | ND | ND | |
P value | 0.006** | ||
Hazard ratio | 3.63 | ||
95% confidence interval | 1.44-9.18 | ||
Radiation for HL (radiation: yes) | ND | ND | |
P value | 0.025** | ||
Hazard ratio | 2.08 | ||
95% confidence interval | 1.09-3.94 | ||
Age at NSCLC diagnosis (older) | ND | ND | |
P value | 0.15 | ||
Hazard ratio | 1.07/year | ||
95% confidence interval | 0.98-1.17 | ||
Calendar year of NSCLC diagnosis (earlier) | |||
P value | 0.21 | 0.021** | 0.16 |
Hazard ratio | 1.06/year | 1.08/year | 1.03/year |
95% confidence interval | 0.96-1.19 | 1.01-1.15 | 1.01-1.08 |
Radiation for NSCLC | ND | ND | |
P value | 0.14 | ||
Hazard ratio | 2.38 | ||
95% confidence interval | 0.75-7.50 | ||
Race (non-white) | ND | ND | |
P value | 0.019** | ||
Hazard ratio | 3.70 | ||
95% confidence interval | 1.23-11.1 |
For the Cox proportional hazards multivariate regression analyses, the initial model included year of NSCLC diagnosis, age at NSCLC diagnosis and any variables with P value <0.2 in the univariate analyses. For patients with regional or distant NSCLC, the variable ‘age at NSCLC diagnosis’ was not significant with univariate or multivariate models (multivariate analysis not shown); therefore, the multivariate model was re-run without that variable (multivariate analysis shown). **P value <0.05 in multivariate analysis. Variables associated with more adverse survival are shown in italics; ‡Mixed cellularity subtype versus all others; Abbreviations: HR=hazard ratio, ND=not done, since P value was >0.2 with univariate analysis, OS=overall survival.