Table 3.
Univariable logistic regression results for clinical features against cancer status in the training set (n = 586, arranged by adjusted P value).
| Variable | Beta | P | OR (95% CI) |
|---|---|---|---|
| Brock | 0.03 | <0.001 | 1.03 (1.02–1.04) |
| Herder | 0.03 | <0.001 | 1.03 (1.02–1.03) |
| Lung disease | −0.68 | <0.001 | 0.51 (0.36–0.71) |
| Extra-thoracic cancer | 1.10 | <0.001 | 3.00 (1.88–4.79) |
| PET avidity | 1.46 | <0.001 | 4.30 (3.01–6.14) |
| Subsolid or GGO density | 1.45 | <0.001 | 4.25 (2.7–6.66) |
| Age | 0.02 | 0.02 | 1.02 (1.00–1.04) |
| Nodule count | −0.46 | 0.02 | 0.63 (0.45–0.89) |
| Smoking status | 0.38 | 0.02 | 1.46 (1.08–1.98) |
| Spiculation | 0.57 | 0.02 | 1.77 (1.14–2.75) |
| Previous lung cancer | 0.91 | 0.09 | 2.49 (0.92–6.72) |
| Gender | 0.24 | 0.18 | 1.27 (0.91–1.77) |
| Upper lobe | 0.11 | 0.55 | 1.12 (0.80–1.57) |
| FH of lung cancer | 0.04 | 0.94 | 1.04 (0.36–3.00) |
6 features had p values < 0.001 (Wald test) and were selected for the multivariable model. Abbreviations: OR: Odds ratio; CI: confidence interval; GGO: ground-glass opacity; FH: Family history.