Comparative clinical evaluation of atlas and deep-learning-based auto-segmentation of organ structures in liver cancer

Table 6 Comparison of relative volume difference (RVD) for atlas-based segmentation against deep-learning-based segmentation with four organs (heart, liver, kidney, stomach). Averages and standard deviations are listed for ten test cases

Test Case	Heart		Liver		Right kidney		Left kidney		Stomach
Test Case	C_atlas	C_deep	C_atlas	C_deep	C_atlas	C_deep	C_atlas	C_deep	C_atlas	C_deep
# 1	0.78	1.10	3.22	2.44	4.63	0.36	14.9	2.67	84.38	14.56
# 2	7.65	9.33	12.10	0.89	3.57	1.77	6.27	0.12	21.33	18.52
# 3	3.08	8.51	0.70	3.81	9.17	3.39	20.25	1.44	48.75	25.55
# 4	7.40	5.13	3.30	0.89	1.62	0.51	0.60	1.26	59.02	19.12
# 5	15.48	0.62	2.89	1.16	1.18	1.58	1.56	7.57	63.49	10.94
# 6	24.82	8.20	12.20	0.35	10.58	1.77	6.16	1.55	70.37	43.71
# 7	23.57	1.77	13.21	0.81	2.11	2.35	3.94	2.69	25.68	30.79
# 8	11.80	7.51	0.11	0.83	4.52	9.72	12.15	3.51	21.67	14.32
# 9	8.31	2.74	4.66	3.64	42.26	12.9	12.29	2.49	89.19	20.80
# 10	26.12	6.75	3.18	3.81	17.87	10.93	24.21	1.21	16.73	14.30
Avg SD	12.90	5.17	5.56	1.86	9.75	4.53	10.23	2.45	50.06	21.26
Avg SD	8.72	3.17	4.72	1.34	11.89	4.49	7.52	1.94	25.92	9.35

ISSN: 1748-717X