Table 4.

Radiologist with AI assistance performance

Author(s)	Dataset	Sample size		Operator		Tests
		HCC	non-HCC	Group	Years experience	Sensitivity	Specificity	Accuracy	NPV	PPV
[23]	Internal test 1	54	41	Radiologist-1 + AI	8–10	0.927	0.815	0.863	0.936	0.792
				Radiologist-2 + AI	8–10	0.927	0.852	0.884	0.939	0.826
				Radiologist-3 + AI	8–10	0.854	0.889	0.874	0.889	0.854
	External test 1	46	36	Radiologist-1 + AI	8–10	0.750	0.826	0.793	0.809	0.771
				Radiologist-2 + AI	8–10	0.833	0.913	0.878	0.875	0.882
				Radiologist-3 + AI	8–10	0.889	0.891	0.890	0.911	0.865
	Internal test 2	26	16	Radiologist-1 + AI	8–10	0.731	0.938	0.810	0.682	0.950
				Radiologist-2 + AI	8–10	0.808	0.938	0.857	0.750	0.955
				Radiologist-3 + AI	8–10	0.731	0.938	0.810	0.682	0.950
	External test 2	23	27	Radiologist-1 + AI	8–10	0.826	0.778	0.800	0.840	0.760
				Radiologist-2 + AI	8–10	0.913	0.593	0.740	0.889	0.656
				Radiologist-3 + AI	8–10	0.913	0.926	0.920	0.926	0.913
[24]	Internal	252	632	Radiologist-1 + AI	5	0.623	0.859	0.794
				Radiologist-2 + AI	> 10	0.885	0.888	0.887
	External	140	452	Radiologist-1 + AI	5	0.693	0.892	0.845
				Radiologist-2 + AI	> 10	0.821	0.914	0.892
[25]	External test 1	64	1161	Radiologist-1 + AI	5–10	0.903	0.988			0.800
				Radiologist-2 + AI	5–10	0.806	0.997			0.926
				Radiologist-3 + AI	5–10	0.710	0.992			0.815
				Radiologist-4 + AI	10–20	0.871	0.997			0.931
				Radiologist-5 + AI	10–20	0.871	0.990			0.818
				Radiologist-6 + AI	10–20	0.833	0.997			0.926