Table 3.

Panel of prediction factors selected in the training data set.

Features		Importance	Remain in model after correlation-based feature selection	Regression coefficients in final model to predict AL	P values in final model to predict AL	Odds ratio for AL (95% CI)
Preoperative factors
	Aortic calcification	29.8	Yes	1.0203	0.0069	2.77(1.32-5.81)
	Celiac trunk calcification	32.4	Yes	1.0275	0.0167	2.79(1.20, 6.48)
	Forced vital capacity ratio (FEV1%)	48.5	Yes	-0.6653	0.0177	0.51(0.30-0.89)
	Transfer factor for carbon monoxide (TLCO) by single-breath (SB) method (%)	51.8	Yes	-0.5820	0.0111	0.56(0.27-1.18)
	Peripheral vascular disease	17.6	Yes	1.6026	0.0109	4.97(1.44-17.07)
Intra-operative and postoperative factors
	Laparoscope	40.3	Yes	1.3665	0.0209	3.92(1.23-12.51)
	Postoperative hospital stay	219.6	Yes	0.1571	0.0000	1.17(1.13-1.21)
	Tumor vascular permeability (laboratory test)	20.1	Yes	-0.7663	0.0507	0.46(0.14-1.48)
	Incision fat liquefaction and infection	23.5	Yes	1.4718	0.0007	4.36(1.86-10.21)