Machine learning-based radiomics score improves prognostic prediction accuracy of stage II/III gastric cancer: A multi-cohort study

Table 2 Univariate and multivariate Cox regression

	Univariate		Multivariate
	RR (95%CI)	P value	RR (95%CI)	P value
Gender (female)	0.885 (0.523-1.498)	0.649
Age (yr)	1.002 (0.981-1.024)	0.843
ASA
1	Ref
2	1.338 (0.669-2.675)	0.411
3	0.979 (0.496-1.934)	0.952
SMI	0.958 (0.930-0.986)	0.004	0.968 (0.932-0.991)	0.009
SMD	0.907 (0.872-0.943)	< 0.001	0.936 (0.889-0.975)	0.002
OACRS	3.158 (1.378-5.247)	< 0.001	3.023 (1.896-4.365)	< 0.001
Operation
Proximal	Ref
Distal	1.499 (0.632-3.553)	0.369
Total	1.018 (0.390-2.652)	0.972
CEA (< 5 ng/mL)	2.236 (1.275-3.921)	0.005	3.174 (1.635-6.035)	0.003
Tumor size	1.035 (0.917-1.168)	0.58
Nerve or vascular invasion	1.700 (1.006-2.871)	0.047
TD (No)	2.167 (1.233-3.810)	0.007
Differentiation
Good	Ref
Moderate	1.808 (0.814-4.016)	0.146
Poor	0.996 (0.569-1.744)	0.989
T stage (2/3)	1.472 (0.867-2.499)	0.153
N stage (0-1)	4.501 (1.793-11.299)	0.001
TNM stage (II)	6.325 (3.425-11.680)	< 0.001	3.698 (2.478-7.635)	< 0.001
Chemotherapy (No)	0.410 (0.241-0.697)	0.001	0.496 (0.268-0.768)	0.002

SMI: Skeletal muscle index; SMD: Skeletal muscle density; TD: Tumor deposition; NA: Not available; TNM: Tumor, node, and metastasis; ASA: American Association of Anesthesiologists score; CEA: Carcinoembryonic antigen; OACRS: Overall survival-associated computed tomography image radiomics score.