Table 4.

Multivariable logistic regression model incorporating LN-RPV and clinical features in the training set (n = 586, arranged by p value).

Variable	Beta	OR (95% CI)	P value
LN-RPV	0.25	1.28 (1.21–1.35)	<0.001
Herder	0.004	1.00 (1.00–1.01)	<0.001
Density – subsolid	0.23	1.26 (1.16–1.37)	<0.001
Density – GGO	0.21	1.24 (1.09–1.40)	<0.001
Extra-thoracic cancer	0.08	1.09 (1.01–1.18)	0.03
Absent lung disease	0.06	1.06 (1.00–1.13)	0.09
Brock	−0.001	1.00 (1.00–1.00)	0.63

The Brock score and absent lung disease lost significance in multivariable testing. LN-RPV, subsolid and GGO density had the highest beta coefficients. Abbreviations: GGO, Ground-glass opacity; OR, Odds ratio; CI, Confidence interval.