Development and validation of machine learning models for predicting no. 253 lymph node metastasis in left-sided colorectal cancer using clinical and CT-based radiomic features

Table 2 Odds ratios of the included feature

Characteristic	Odds ratio	CI lower	CI upper	p-value
Tumor site, sigmoid colon	2.237	1.375	3.607	0.001
Endoscopic obstruction, yes	3.390	2.079	5.481	< 0.001
CEA, high	5.080	3.011	8.974	< 0.001
Growth type, ulcerative	2.771	1.476	5.784	0.003
Tumor differentiation, poorly/undifferentiation	3.167	1.814	5.333	< 0.001
Pathological classification, Signet ring adenocarcinoma/Undifferentiated carcinoma	14.103	2.928	54.652	< 0.001
LLN average CT value, HU	13.900	1.027	285.638	0.013
LLN SAD, mm	41.105	6.938	299.895	< 0.001
LLN LAD, mm	20.467	2.25	226.301	0.010
Firstorder_Maximum	647.267	10.982	75783.38	0.005
GLSZM_LargeAreaHighGrayLevelEmphasis	8.20 × 10⁹	4502.913	7.88 × 10¹⁸	0.011
GLSZM_SmallAreaHighGrayLevelEmphasis	9.257	1.527	61.773	0.017
Shape_SurfaceVolumeRatio	0.014	0.001	0.118	< 0.001

LN = lymph node; LLN = largest lymph node; SAD = short-axis diameter; LAD = long-axis diameter; Firstorder_Maximum = the maximum gray level intensity within the region of interest; GLSZM = Gray Level Size Zone Matrix

ISSN: 1470-7330