Machine learning-based gray-level co-occurrence matrix signature for predicting lymph node metastasis in undifferentiated-type early gastric cancer

Table 1 Patient baseline population and image index characteristic

Variables	Training set				Testing set
	Overall (n = 368)	Yes (n = 62)	No (n = 306)	P value	Overall (n = 158)	Yes (n = 29)	No (n = 129)	P value
Age (median, IQR), yr	51.00 (40.75, 61.00)	52.50 (40.25, 64.50)	51.00 (41.00, 60.00)	0.185	47.00 (37.25, 58.75)	53.00 (39.00, 63.00)	46.00 (37.00, 57.00)	0.215
Sex (%)
Male	266 (72.3)	48 (77.4)	218 (71.2)	0.404	123 (77.8)	21 (72.4)	102 (79.1)	0.594
Female	102 (27.7)	14 (22.6)	88 (28.8)		35 (22.2)	8 (27.6)	27 (20.9)
Site (%)
Nearly 1/3	81 (22.0)	10 (16.1)	71 (23.2)	0.384	37 (23.4)	9 (31.0)	28 (21.7)	0.138
Medium 1/3	73 (19.8)	15 (24.2)	58 (19.0)		23 (14.6)	1 (3.4)	22 (17.1)
Far 1/3	214 (58.2)	37 (59.7)	177 (57.8)		98 (62.0)	19 (65.5)	79 (61.2)
Ulcer (%)
Yes	103 (28.0)	21 (33.9)	82 (26.8)	0.329	45 (28.5)	11 (37.9)	34 (26.4)	0.308
No	265 (72.0)	41 (66.1)	224 (73.2)		113 (71.5)	18 (62.1)	95 (73.6)
Gross_type (%)
Uplift	98 (26.6)	14 (22.6)	84 (27.5)	0.587	56 (35.4)	8 (27.6)	48 (37.2)	0.227
Flat	72 (19.6)	11 (17.7)	61 (19.9)		27 (17.1)	8 (27.6)	19 (14.7)
Sunken	198 (53.8)	37 (59.7)	161 (52.6)		75 (47.5)	13 (44.8)	62 (48.1)
Tumor_size (%)
≤ 2 cm	296 (80.4)	18 (29.0)	278 (90.8)	< 0.001	120 (75.9)	6 (20.7)	114 (88.4)	< 0.001
> 2 cm	72 (19.6)	44 (71.0)	28 (9.2)		38 (24.1)	23 (79.3)	15 (11.6)
Infiltration_depth (%)
Mucosal layer	267 (72.6)	12 (19.4)	255 (83.3)	< 0.001	113 (71.5)	9 (31.0)	104 (80.6)	< 0.001
Submucosa	101 (27.4)	50 (80.6)	51 (16.7)		45 (28.5)	20 (69.0)	25 (19.4)
Vascular_invasion (%)
Yes	124 (33.7)	42 (67.7)	82 (26.8)	< 0.001	56 (35.4)	21 (72.4)	35 (27.1)	< 0.001
No	244 (66.3)	20 (32.3)	224 (73.2)		102 (64.6)	8 (27.6)	94 (72.9)
VTT (%)
Yes	124 (33.7)	45 (72.6)	79 (25.8)	< 0.001	44 (27.8)	23 (79.3)	21 (16.3)	< 0.001
No	244 (66.3)	17 (27.4)	227 (74.2)		114 (72.2)	6 (20.7)	108 (83.7)
TF (median, IQR)	3.78 (3.56, 3.99)	4.13 (3.97, 4.27)	3.70 (3.51, 3.90)	< 0.001	3.79 (3.52, 4.01)	4.16 (4.00, 4.31)	3.70 (3.49, 3.93)	< 0.001
EV (median, IQR)	0.88 (0.64, 1.12)	0.60 (0.49, 0.68)	0.98 (0.72, 1.20)	< 0.001	0.85 (0.65, 1.09)	0.64 (0.54, 0.70)	0.92 (0.72, 1.16)	< 0.001
Entropy (median, IQR)	8.68 (8.37, 8.98)	10.51 (10.07, 10.88)	8.57 (8.33, 8.83)	< 0.001	8.79 (8.43, 9.02)	10.44 (10.16, 10.96)	8.65 (8.38, 8.89)	< 0.001
IG_all (median, IQR)	2.16 (1.76, 2.47)	3.04 (2.64, 3.62)	2.03 (1.69, 2.30)	< 0.001	2.12 (1.75, 2.47)	2.94 (2.70, 3.54)	1.97 (1.64, 2.31)	< 0.001
IG_0 (median, IQR)	2.26 (1.75, 2.66)	3.34 (2.72, 3.80)	2.10 (1.69, 2.48)	< 0.001	2.41 (1.90, 2.79)	3.70 (3.16, 4.18)	2.22 (1.77, 2.62)	< 0.001
IG_45 (median, IQR)	1.88 (1.54, 2.18)	2.85 (2.32, 3.26)	1.78 (1.47, 2.04)	< 0.001	1.85 (1.48, 2.18)	2.73 (2.31, 3.11)	1.74 (1.40, 2.03)	< 0.001
IG_90 (median, IQR)	2.34 (1.85, 2.85)	3.36 (2.89, 3.84)	2.20 (1.75, 2.63)	< 0.001	2.42 (1.94, 2.78)	3.27 (3.03, 3.61)	2.25 (1.75, 2.61)	< 0.001
IV_all (median, IQR)	176.90 (148.98, 207.25)	134.80 (109.30, 163.02)	182.00 (156.00, 210.75)	< 0.001	175.50 (143.25, 200.75)	133.50 (105.80, 155.70)	183.00 (154.00, 206.00)	< 0.001
IV_all_SD (median, IQR)	4584.00 (3148.00, 6602.50)	2166.50 (1340.50, 3535.00)	5025.00 (3747.00, 7011.75)	< 0.001	4940.50 (2987.25, 6682.00)	2849.00 (1841.00, 3428.00)	5618.00 (3813.00, 6897.00)	< 0.001
IV_0 (median, IQR)	149.85 (122.75, 186.75)	96.40 (78.95, 125.82)	163.20 (134.00, 195.65)	< 0.001	146.70 (112.78, 185.57)	74.10 (65.60, 90.60)	158.40 (131.40, 196.20)	< 0.001
IV_45 (median, IQR)	239.55 (201.40, 284.75)	164.40 (123.83, 188.62)	254.30 (220.67, 290.60)	< 0.001	226.25 (201.25, 266.67)	157.40 (133.90, 193.80)	243.90 (214.30, 273.50)	< 0.001
IV_90 (median, IQR)	129.00 (103.00, 154.00)	101.00 (77.75, 119.00)	134.00 (109.25, 159.00)	< 0.001	124.50 (109.00, 150.75)	105.00 (77.00, 118.00)	133.00 (117.00, 156.00)	< 0.001
Haralick_all (median, IQR)	0.10 (0.09, 0.10)	0.12 (0.11, 0.13)	0.09 (0.09, 0.10)	< 0.001	0.10 (0.09, 0.10)	0.12 (0.12, 0.14)	0.09 (0.09, 0.10)	< 0.001
Haralick_30 (median, IQR)	0.10 (0.09, 0.11)	0.14 (0.12, 0.15)	0.10 (0.09, 0.11)	< 0.001	0.10 (0.09, 0.11)	0.14 (0.13, 0.15)	0.10 (0.09, 0.11)	< 0.001
Haralick_45 (median, IQR)	0.09 (0.08, 0.10)	0.11 (0.10, 0.12)	0.09 (0.08, 0.10)	< 0.001	0.09 (0.08, 0.10)	0.11 (0.10, 0.13)	0.09 (0.08, 0.10)	< 0.001
Haralick_90 (median, IQR)	0.11 (0.10, 0.13)	0.14 (0.12, 0.16)	0.11 (0.09, 0.12)	< 0.001	0.12 (0.10, 0.13)	0.15 (0.12, 0.16)	0.11 (0.09, 0.13)	< 0.001
CSV (median, IQR)	106.00 (102.00, 111.00)	108.00 (105.00, 111.00)	106.00 (101.00, 111.00)	0.001	107.00 (102.25, 111.00)	109.00 (105.00, 113.00)	107.00 (102.00, 111.00)	0.007
CP (median, IQR)	65.50 (60.00, 70.00)	68.00 (66.00, 71.00)	64.00 (59.00, 70.00)	< 0.001	64.00 (60.00, 68.00)	67.00 (64.00, 68.00)	63.00 (59.00, 68.00)	0.002

IQR: Interquartile range; TF: Total frequency; EV: Energy value; IV_0: Inertia value 0°; IV_45: Inertia value 45°; IV_90: Inertia value 90°; IG_0: Inverse gap 0°; IG_45: Inverse gap 45°; IG_90: Inverse gap 90°; IG_all: Inverse gap full angle; Haralick_30: Haralick 30°; Haralick_45: Haralick 45°; Haralick_90: Haralick 90°; Haralick_all: Haralick full angle; CSV: Cluster shadow value; CP: Cluster prominence.

Table 2 Receiver operating characteristic curve analysis of lymph node metastasis in each mL based model

Model	Training set		Testing set		Variables1
	AUC mean	AUC 95%CI	AUC mean	AUC 95%CI
RFC	0.925	0.378-1.472	0.912	0.355-1.469	Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, IV_45
ANN	0.887	0.340-1.434	0.837	0.280-1.394	Entropy, IG_all, IG_0, IG_45, IG_90, IV_all, IV_all_SD, IV_0, IV_45, Haralick_all, Haralick_30
DT	0.856	0.309-1.403	0.813	0.256-1.370	Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, IV_45
XGboost	0.814	0.267-1.361	0.807	0.250-1.364	Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, IV_45, IG_90
SVM	0.805	0.258-1.352	0.794	0.237-1.351	Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, IV_45
GLM	0.796	0.229-1.362	0.799	0.233-1.365	Entropy, Haralick_all, Haralick_30, IG_all, IG_45, IG_0, IV_45
Radiologist	0.789	0.242-1.336	0.801	0.254-1.348	-

¹Variables are included in the model.

RFC: Random forest classifier; SVM: Support vector machine; DT: Decision tree; ANN: Artificial neural network; XGboost: Extreme gradient boosting; GLM; Generalized linear model; AUC: Area under the receiver operating characteristic curve; 95%CI: 95% confidence interval; IV_0: Inertia value 0°; IV_45: Inertia value 45°; IV_90: Inertia value 90°; IG_0: Inverse gap 0°; IG_45: Inverse gap 45°; IG_90: Inverse gap 90°; IG_all: Inverse gap full angle; Haralick_30: Haralick 30°.