TABLE 2.

Comparison of models across three metrics for different cohorts trained with augmentation data. The significant p_value should be shown in bold.

PMW-1
		Without aug.	With aug.	$p$	Difference [%]
DSC (Mean ± SD [%])	U-Net	0.824 ± 0.051	0.834 ± 0.047	0.006	1.21
	AFFU	0.833 ± 0.056	0.836 ± 0.050	0.796	NS
RVE (Mean ± SD [%])	U-Net	0.786 ± 0.145	0.780 ± 0.156	0.918	NS
	AFFU	0.801 ± 0.104	0.806 ± 0.125	0.717	NS
HD Mean ± SD [mm]	U-Net	27.2 ± 12.1	22.4 ± 9.0	0.020	17.65
	AFFU	19.7 ± 7.5	17.8 ± 7.0	0.011	9.64
PMW-2
DSC (Mean ± SD [%])	U-Net	0.828 ± 0.073	0.846 ± 0.061	<0.001	2.17
	AFFU	0.833 ± 0.065	0.842 ± 0.060	<0.001	1.08
RVE (Mean ± SD [%])	U-Net	0.880 ± 0.105	0.897 ± 0.094	<0.001	1.93
	AFFU	0.873 ± 0.105	0.887 ± 0.100	<0.001	1.60
HD Mean ± SD [mm]	U-Net	45.0 ± 47.6	36.1 ± 39.9	<0.001	19.8
	AFFU	25.9 ± 27.9	27.0 ± 30.1	0.065	NS
PMW-OB
DSC (Mean ± SD [%])	U-Net	0.857 ± 0.049	0.866 ± 0.046	<0.001	1.05
	AFFU	0.862 ± 0.048	0.864 ± 0.049	0.097	NS
RVE (Mean ± SD [%])	U-Net	0.914 ± 0.072	0.916 ± 0.069	0.472	NS
	AFFU	0.919 ± 0.076	0.913 ± 0.079	0.067	NS
HD Mean ± SD [mm]	U-Net	54.5 ± 58.4	51.6 ± 59.9	0.380	NS
	AFFU	34.8 ± 46.8	36.2 ± 52.9	0.071	NS