Table 2.

CNN-based prediction performances under models using different inputs.

Input	ACC (%)	SEN (%)	SPE (%)	AUC
T1	72.52 ± 1.12***	73.10 ± 1.80**	72.03 ± 1.72***	0.7904 ± 0.0214***
CEST	74.01 ± 1.15***	75.70 ± 3.00*	72.50 ± 1.14***	0.8022 ± 0.0147***
T1 + CEST	81.75 ± 1.98	79.64 ± 3.33	83.68 ± 3.03	0.8689 ± 0.0145*
T1 + annotation	75.76 ± 2.02**	74.47 ± 3.05**	76.70 ± 3.74*	0.8293 ± 0.0221**
CEST + annotation	74.88 ± 2.58**	74.37 ± 4.24**	75.28 ± 2.27*	0.8192 ± 0.0216**
T1 + CEST + annotation	82.94 ± 1.23	82.35 ± 1.87	83.45 ± 2.50	0.8868 ± 0.0055

The data are reported in form of “Mean ± STD”. * “CEST + T1 + annotation” is significantly higher than the indicated method at p < 0.05 level. ** p < 0.01. *** p < 0.001.