Table 1. The classification performance of the CNN model using datasets acquired at different combinations of image phases.

Dataset	Accuracy	Precision	Sensitivity	F1 score	AUC
C1 (all 6 phases)	0.833	0.800	0.889	0.842	0.92
C2 (pre, arterial, washout)	0.900	0.835	1.0	0.909	0.95
C3 (C2 + hepatobiliary phase)	0.833	0.803	0.889	0.843	0.92
C4 (pre, me, mw)	0.789	0.814	0.756	0.780	0.91
C2 without transfer learning	0.767	0.777	0.778	0.767	0.90

pre, pre-contrast; arterial, arterial phase; washout, washout phase; me, maximum enhancement phase; mw, maximum washout phase; CNN, convolutional neural network; AUC, area under the receiver operating characteristic curve.