Table 2.

Summary of multivariable conditional logistic regression results for lung cancer risk prediction

VariablesLLP subjects with both EPI variables and SEZ6L SNP (n = 388)Original LLP risk model using LLP case-control subjects (n = 1,275)Model with SEZ6L only using all subjects genotyped (n = 463)
Model with EPI variables onlyModel with EPI + SEZ6L variables
OR (95% CI)OR (95% CI)OR (95% CI)OR (95% CI)
Smoking duration (y)
    1-19 1.84 (0.79-4.24)1.69 (0.72-3.94)2.16 (1.22-3.82)
    20-39 5.32 (2.48-11.41)5.08 (2.34-11.01)4.28 (2.63-6.96)
    40+13.72 (6.54-28.76)13.37 (6.33-28.22)12.39 (7.50-20.46)
Pneumonia (Yes)1.25 (0.68-2.31)1.24 (0.67-2.33)1.82 (1.26-2.63)
Asbestos exposure (Yes)2.35 (1.35-4.08)2.32 (1.32-4.06)1.89 (1.35-2.63)
Previous tumor (Yes)1.62 (0.61-4.32)1.58 (0.58-4.29)1.97 (1.23-3.15)
Family history of cancer
    Early onset (<60 y)2.16 (0.82-5.67)2.34 (0.88-6.20)2.02 (1.18-3.46)
    Late onset (>60 y)0.87 (0.46-1.67)0.93 (0.49-1.79)1.19 (0.80-1.78)
SEZ6L genotype marker
    TG1.17 (0.72-1.92)1.19 (0.78-1.82)
    TT4.51 (1.56-13.02)3.79 (1.45-9.86)
Goodness-of-fit statistic
    Hosmer-Lemeshow (P)6.24 (0.62)5.54 (0.70)
    AIC418.12413.38

Abbreviations: EPI, epidemiological; AIC, Aikaike information criteria.