A Deep Learning Radiomics Analysis for Survival Prediction in Esophageal Cancer

Table 1

Clinical characteristics of esophageal cancer patients.

	Training cohort (n = 116)			Validation cohort (n = 38)
Characteristic	Nonsurvival (81)	Survival (35)		Nonsurvival (24)	Survival (14)
Gender
Male: female	51 : 30	23 : 12	0.836	12 : 12	6 : 8	0.745
Age	67.96 ± 9.41	68.83 ± 8.80	0.644	66.54 ± 9.65	67.36 ± 7.09	0.785
BMI	21.70 ± 3.14	23.69 ± 4.25	0.006	23.04 ± 3.98	22.55 ± 3.24	0.702
T-stage
T1	1 (1.2%)	2 (5.7%)	0.042	1 (4.2%)	0 (0.0%)	0.938
T2	20 (24.7%)	16 (45.7%)		9 (37.5%)	7 (50.0%)
T3	43 (53.1%)	13 (37.1%)		10 (41.7%)	5 (35.7%)
T4	17 (21.0%)	4 (11.4%)		4 (16.7%)	2 (14.3%)
N
N0	20 (25.0%)	14 (40.0%)	0.123	7 (29.2%)	6 (42.9%)	0.486
N1	60 (75.0%)	21 (60.0%)		17 (70.8%)	8 (57.1%)
M
M0	69 (85.2%)	35 (100.0%)	0.017	22 (91.7%)	14 (100.0%)	0.522
M1	12 (14.8%)	0 (0.0%)		2 (8.3%)	0 (0.0%)
TNM
I	0 (0.0%)	2 (5.7%)	0.002	0 (0.0%)	0 (0.0%)	0.766
II	27 (33.3%)	19 (54.3%)		10 (41.7%)	8 (57.1%)
III	42 (51.9%)	14 (40.0%)		12 (50.0%)	5 (35.7%)
IV	12 (14.8%)	0 (0.0%)		2 (8.3%)	1 (7.1%)
PTV	379.62 ± 158.34	335.22 ± 179.33	0.186	382.56 ± 134.39	362.53 ± 166.27	0.687
HCR_signature	−0.50 ± 1.03	0.68 ± 1.12	<0.001	−0.23 ± 1.48	0.80 ± 1.26	0.035
DLR_signature	−2.22 ± 0.75	1.10 ± 0.75	<0.001	−1.90 ± 1.12	0.15 ± 1.11	<0.001