TABLE II.

Statistics of the compared methods on the test set (22 cases) based on the contours generated by Simultaneous Truth and Performance Level Estimation Algorithm (STAPLE). Average values and their standard deviations were reported.

		Score		DSC		ASSD
Methods	Modalities	Single	Dual	Single	Dual	Single	Dual
Song et al.	CT	0.712 ± 0.140	0.734 ± 0.089	0.597 ± 0.257	0.638 ± 0.165	2.574 ± 2.447	1.938 ± 1.132
	PET	0.735 ± 0.084	0.740 ± 0.074	0.629 ± 0.158	0.643 ± 0.141	2.901 ± 2.607	2.656 ± 2.131
Ju et al.	CT	0.778 ± 0.082	0.765 ± 0.098	0.765 ± 0.093	0.759 ± 0.100	1.484 ± 0.745	1.650 ± 0.823
	PET	0.817 ± 0.064	0.820 ± 0.061	0.776 ± 0.106	0.782 ± 0.099	1.526 ± 1.068	1.473 ± 0.970
Zhong et al.	CT	0.798 ± 0.058	0.809 ± 0.058	0.766 ± 0.095	0.783 ± 0.095	1.186 ± 0.828	1.080 ± 0.714
	PET	0.774 ± 0.065	0.816 ± 0.054	0.711 ± 0.123	0.778 ± 0.086	2.102 ± 2.002	1.502 ± 1.083
3D-UNet (CELoss)	CT	0.812 ± 0.115	—	0.780 ± 0.185	—	1.667 ± 1.863	—
	PET	0.846 ± 0.084	—	0.811 ± 0.133	—	1.127 ± 0.718	—
DFCN-CoSeg (CELoss)	CT	—	0.850 ± 0.061	—	0.836 ± 0.095	—	0.895 ± 0.661
	PET	—	0.848 ± 0.064	—	0.823 ± 0.086	—	1.066 ± 0.660
3D-UNet (DICELoss)	CT	0.839 ± 0.085	—	0.811 ± 0.151	—	1.291 ± 1.313	—
	PET	0.832 ± 0.075	—	0.794 ± 0.111	—	1.229 ± 0.587	—
DFCN-CoSeg (DICELoss)	CT	—	0.865 ± 0.034	—	0.861 ± 0.037	—	0.806 ± 0.605
	PET	—	0.853 ± 0.063	—	0.828 ± 0.087	—	1.079 ± 0.761