Predictive value of CAC score combined with clinical features for obstructive coronary heart disease on coronary computed tomography angiography: a machine learning method

Table 2 The comparisons of NRI and IDI between the different models

	NRI (95%CI)	P	IDI (95%CI)	P
RF model versus clinical model	0.18 (0.040–0.327)	0.012	0.03 (0.005–0.059)	0.021
Events	0.08 (− 0.053 to 0.221)	0.218
Non-events	0.09 (0.035–0.158)	0.001
RF model versus CACS model	0.82 (0.696–0.953)	< 0.0001	0.13 (0.10–0.155)	< 0.0001
Events	0.47 (0.352–0.578)	< 0.0001
Non-events	0.35 (0.298–0.406)	< 0.0001
Clinical model versus CACS model	0.64 (0.501–0.767)	< 0.0001	0.09 (0.075–0.117)	< 0.0001
Events	0.28 (0.153–0.393)	< 0.0001
Non-events	0.36 (0.303–0.412)	< 0.0001

CACS model including CAC score. Clinical model including ASCVD risk factors and CAC score
ASCVD, atherosclerotic cardiovascular disease; LR, logistic regression; RF, random forest; NRI

ISSN: 1471-2261