Development and Validation of the Random Forest Model via Combining CT-PET Image Features and Demographic Data for Distant Metastases among Lung Cancer Patients

Table 1

Comparison of demographic data, 4 CT image features, and 2 PET image features among the M0 and M1 groups.

Variables	Total (n = 134)	M0 (n = 109)	M1 (n = 25)	Statistic
History of smoking, n (%)				χ^2 = 8.985	0.003
No	46 (34.33)	31 (28.44)	15 (60.00)
Yes	88 (65.67)	78 (71.56)	10 (40.00)
Gender, n (%)				χ^2 = 0.486	0.486
Male	72 (53.73)	57 (52.29)	15 (60.00)
Female	62 (46.27)	52 (47.71)	10 (40.00)
Age, mean ± SD	64.56 ± 9.39	64.80 ± 9.01	63.49 ± 11.03	t = 0.63	0.529
CT
First-order 10 percent, M (Q1, Q3)	136.60 (117.00, 165.00)	134.00 (117.00, 161.00)	140.00 (119.00, 199.00)	Z = 0.768	0.442
First-order robust mean absolute deviation, mean ± SD	357.41 ± 25.63	358.68 ± 24.12	351.87 ± 31.35	t = 1.20	0.233
GLCM joint average, mean ± SD	23.99 ± 2.77	23.86 ± 2.69	24.55 ± 3.10	t = −1.13	0.262
NGTDM strength, M (Q1, Q3)	3.12 (2.63, 3.84)	3.12 (2.60, 3.84)	3.40 (2.76, 3.74)	Z = 0.942	0.346
PET
GLDM low gray-level emphasis, M (Q1, Q3)	0.22 (0.08, 0.32)	0.18 (0.07, 0.32)	0.30 (0.23, 0.33)	Z = 2.547	0.011
NGTDM strength, M (Q1, Q3)	2909.75 (174.97, 12124.24)	2714.84 (179.85, 9674.76)	6176.46 (128.30, 14465.71)	Z = 0.463	0.644

Note. CT: computed tomography; PET: positron emission tomography; GLCM: gray-level co-occurrence matrix; NGTDM: neighborhood gray-tone difference matrix; GLDM: gray-level difference method.