Table 2. Meta-regression of the Log (pooled DOR) of AI-assisted CT diagnostic technology^†.

Variable	Estimate (95% CI)	Standard error	P value
Intercept	9.4338 (7.4538, 11.4137)	0.9818	<0.0001
Algorithms (control = deep belief network)
Support vector machine (1: yes, 0: no)	−3.1780 (−4.5967, −1.7587)	0.7036	<0.0001
Decision tree (1: yes, 1: no)	−2.3370 (−3.8411, −0.8320)	0.7460	0.0031
Convolutional neural networks (1: yes, 1: no)	−2.1640 (−3.8342, −0.4931)	0.8284	0.0124
Artificial neural network (1: yes, 0: no)	−2.9970 (−6.1486, 0.1548)	1.5628	0.0618
Others (1: yes, 0: no)	−3.0740 (−4.6794, −1.4692)	0.7959	0.0004
No. of nodules (1: ≥150, 0: <150)	−0.8420 (−1.8731, 0.1889)	0.5112	0.1068
China (1: yes, 0: no)	−2.0550 (−3.6124, −0.4980)	0.7722	0.0109

^†, a multilevel linear regression model (method = REML, weight = 1/variance of odds) was used to control for the study random effects. DOR, diagnostic odds ratio; AI, artificial intelligence.