Table 9 Comparison of descriptive statistics from models trained with segments, segments with vessel (color channel) and segments with spleen on portal vein embolization cohort C_PVE.

From: Fully automated deep learning based auto-contouring of liver segments and spleen on contrast-enhanced CT images

	Seg 1, N = 20				Seg 2, N = 20¹				Seg 3, N = 20¹				Seg 4, N = 20¹				Seg 5–8, N = 20¹				Spleen, N = 20¹
	(IntraMD^2a = 0.88)				(IntraMD^2a = 0.88)				(IntraMD^2a = 0.94)				(IntraMD^2a = 0.92)				(IntraMD^2a = 0.99)
	(InterMD^2b = 0.82)				(InterMD^2b = 0.85)				(InterMD^2b = 0.91)				(InterMD^2b = 0.88)				(InterMD^2b = 0.96)
	M_paU-Net¹	M_nnUnet¹	M_seg+spleen¹	M_vess¹	M_paU-Net¹	M_nnUnet¹	M_seg+spleen¹	M_vess¹	M_paU-Net¹	M_nnUnet¹	M_seg+spleen¹	M_vess¹	M_paU-Net¹	M_nnUnet¹	M_seg+spleen¹	M_vess¹	M_paU-Net¹	M_nnUnet¹	M_seg+spleen¹	M_vess¹	M_seg+spleen¹
Dice	0.90	0.90	0.91	0.90	0.81	0.82	0.82	0.82	0.87	0.88	0.88	0.88	0.85	0.89	0.89	0.89	0.92	0.93	0.93	0.94	0.97
	0.88, 0.05	0.88, 0.06	0.88, 0.06	0.89,0.05	0.77, 0.19	0.80, 0.10	0.80, 0.10	0.80, 0.11	0.86, 0.05	0.87, 0.05	0.87, 0.05	0.87, 0.05	0.81, 0.10	0.88, 0.05	0.88, 0.05	0.87, 0.07	0.91, 0.06	0.92, 0.04	0.93, 0.04	0.92, 0.05	0.97, 0.01
	0.95, 0.75	0.94, 0.69	0.94, 0.67	0.94, 0.75	0.89, 0.00	0.89, 0.42	0.89, 0.43	0.89, 0.41	0.92, 0.69	0.93, 0.68	0.92, 0.71	0.93, 0.70	0.91, 0.46	0.94, 0.77	0.95, 0.75	0.94, 0.64	0.96, 0.72	0.98, 0.82	0.98, 0.84	0.98, 0.80	0.98, 0.96
\({{\text{HD}}}_{95}^{3}\)	4	4.3	4	4	13	10.9	11	11	12	9.3	10	9	15	7.8	8	9	15	14.9	11	10	2
	5, 4	5.5, 4.3	5, 4	5, 3	16, 13	13.5, 10.3	12, 6	14, 11	17, 14	10.2, 4.3	11, 8	12, 8	18, 8	9.5, 5.1	9, 6	11, 7	18, 10	16.1, 15.9	12, 7	12, 9	2, 0
	16, 2	21.7, 2.5	21, 2	17, 2	66, 5	51.7, 5.8	27, 6	55, 5	56, 5	22.0, 5.8	40, 5	42, 5	36, 6	21.9, 3.4	22, 3	27, 3	47, 5	77.2, 2.5	27, 2	41, 3	3, 1
Avg.\({{\text{HD}}}_{{\text{A}}}^{3}\)	0.21	0.19	0.19	0.21	0.80	0.72	0.70	0.71	0.53	0.43	0.46	0.45	0.89	0.36	0.37	0.43	0.28	0.30	0.26	0.23	0.03
	0.33,0.32	0.35,0.47	0.35,0.49	0.31, 0.31	2.35,6.49	0.96, 0.84	0.96, 0.78	1.14, 1.43	0.91, 0.87	0.60, 0.60	0.65, 0.77	0.68, 0.84	1.15,0.92	0.49, 0.38	0.50, 0.45	0.62, 0.65	0.61, 0.94	0.42, 0.36	0.33, 0.30	0.41, 0.51	0.03, 0.01
	1.26,0.06	2.15, 0.09	2.26, 0.08	1.43, 0.08	29.81,0.23	3.71, 0.27	3.15, 0.25	6.47, 0.26	3.56,0.20	2.92, 0.21	3.73, 0.20	4.11, 0.20	3.41, 0.24	1.49, 0.10	1.74, 0.07	2.83, 0.09	4.19, 0.09	1.25, 0.03	1.16, 0.03	2.26, 0.04	0.08, 0.02
PDV	7	6	6	5	19	15	15	20	7	5	6	6	13	6	7	6	3	4	4	4	1
	7, 5	10, 18	10, 20	8, 13	34, 68	18, 13	18, 13	22, 16	9, 6	7, 6	7, 7	8, 6	14, 14	8, 7	9, 8	10, 12	6, 5	6, 6	6, 5	6, 6	1, 1
	19, 0	85, 0	92, 0	61, 1	320, 3	53, 1	53, 5	56, 0	23, 2	22, 0	25, 1	26, 1	62, 1	26, 0	32, 1	48, 0	16, 0	24, 1	22, 0	22, 0	3, 0

¹Data in each cell is organized as row 1 = Median, row 2 = Mean, Standard deviation, row 3 = Max, Min; ¹M_nnUnet = model trained with segments only, M_seg+spleen = Model trained with Segments and Spleen as labels, M_vess = Model trained with segments as label and vessel as color channel; ^2aIntraMD = Intra-observer mean dice,^2bIntrerMD = Inter-observer mean dice, ³HD = Hausdorff distance (95 = 95th percentile and A = Average) in mm; PDV = percent difference in volume.

Back to article page

Table 9 Comparison of descriptive statistics from models trained with segments, segments with vessel (color channel) and segments with spleen on portal vein embolization cohort C_PVE.

Search

Quick links