Table 2.

Performance of 3D-ResNet on testing set, validation set, training set, and external test set

Performance	Testing set (n = 106) NTM-LD (n = 28) MTB-LD (n = 78)	Validation set (n = 113) NTM-LD (n = 29) MTB-LD (n = 84)	Training set (n = 886) NTM-LD (n = 244) MTB-LD (n = 642)	External test set (n = 80) NTM-LD (n = 40) MTB-LD (n = 40)
AUC (95% CI)	0.86 (0.76, 0.95)	0.88 (0.80, 0.96)	0.90 (0.88, 0.93)	0.78 (0.68, 0.89)
Accuracy	0.83	0.88	0.87	0.69
Specificity	0.57	0.52	0.56	0.63
Sensitivity	0.92	1	0.98	0.75
NTM-LD precision	0.73	1	0.93	0.71
MTB-LD precision	0.86	0.86	0.86	0.67
NTM-LD F1 score	0.64	0.68	0.70	0.67
MTB-LD F1 score	0.89	0.92	0.91	0.70