. 2020 Jan 21;19:5. doi: 10.1186/s12938-019-0744-0

Table 2.

Univariate and multivariable regression analyses of the Radscore with primary clinical features for the histological subtype prediction of NSCLC in the training cohort

Indicators	Univariate analysis				Multivariable analysis
	OR^a	95% CI		p value	OR	95% CI		p value
	OR^a	Lower	Upper	p value	OR	Lower	Upper	p value
Age	2.481	1.368	4.499	< 0.05	3.994	1.421	11.228	< 0.05
Sex	0.068	0.015	0.313	≪ 0.05	0.413	0.015	10.889	0.59
Smoking	0.148	0.053	0.409	≪ 0.05	0.104	0.012	0.947	< 0.05
Side	0.353	0.127	0.981	0.051	–	–	–	–
Location	2.191	0.954	5.028	0.06	–	–	–	–
LD^a	1.991	1.061	3.734	< 0.05	0.048	0.005	0.511	< 0.05
LPD^a	3.246	1.651	6.379	≪ 0.05	9.807	1.420	67.719	< 0.05
CEA^a	0.784	0.635	0.968	< 0.05	0.661	0.488	1.129	0.062
Radscore	11.128	4.100	30.198	≪ 0.01	74.937	9.339	601.32	≪ 0.01

The underlined values indicate statistical significance with p value < 0.05 after the univariate analysis

The italics underlined values indicate statistical significance with p value < 0.05 after the multivariable analysis

^aLD, LPD, CEA and OR indicate the longest diameter, the longest perpendicular diameter, carcinoembryonic antigen, and odds ratio, respectively