Preoperative prediction of malignant potential of 2-5 cm gastric gastrointestinal stromal tumors by computerized tomography-based radiomics

Table 1 Patients’ characteristics between low-grade and high-grade malignant potential groups

Characteristics	Low-grade, n = 82	High-grade, n = 21	t or χ2 value	P value
Age in year	58.57 ± 8.90	57.28 ± 10.47	0.570	0.570
Sex, n (%)			0.354	0.552
Male	37 (45.1)	11 (52.4)
Female	45 (59.9)	10 (47.6)
Largest diameter	32.66 ± 8.77	38.76 ± 9.09	2.824	0.006a
Location, n (%)			2.109	0.550
Cardia	2 (2.4)	0 (0)
Fundus	40 (48.8)	12 (57.1)
Body	28 (34.1)	8 (38.1)
Antrum	12 (14.6)	1 (4.8)
Growth patterns, n (%)			2.196	0.334
Endoluminal	39 (47.6)	11 (52.4)
Exophytic	24 (29.3)	3 (14.3)
Mixed	19 (23.2)	7 (33.3)
Contour, n (%)			4.646	0.031a
Regular	56 (68.3)	9 (42.9)
Irregular	26 (31.7)	12 (57.1)
Margin, n (%)			5.645	0.018a
Well-defined	67 (81.7)	12 (57.1)
Poorly	15 (18.3)	9 (42.9)
Necrosis, n (%)			4.268	0.039a
Absent	48 (58.5)	7 (33.3)
Present	34 (41.5)	14 (66.7)
Calcification, n (%)			0.630	0.427
Absent	75 (91.5)	18 (85.7)
Present	7 (8.5)	3 (14.3)
Ulceration, n (%)			7.823	0.005a
Absent	67 (81.7)	11 (52.4)
Present	15 (18.3)	10 (47.6)
Plain CT value	34.65 ± 37.92	31.10 ± 13.23	0.421	0.674
Arterial phase CT value	63.70 ± 36.50	59.81 ± 18.58	0.471	0.639
Venous phase CT value	71.78 ± 35.76	63.43 ± 17.32	1.035	0.303
Delayed phase CT value	73.65 ± 34.96	66.14 ± 14.39	0.960	0.339

^aP < 0.05.

Independent samples t-test was applied in continuous variables. χ² test was applied for categorical variables. CT: Computerized tomography.

Table 2 Patients’ characteristics between the training group and the test group

Characteristics	Training, n = 69	Test, n = 34	t or χ2 value	P value
Age in year	58.30 ± 9.02	58.32 ± 9.71	0.01	0.992
Sex, n (%)			0.004	0.948
Male	32 (53.6)	16 (52.9)
Female	37 (46.4)	18 (47.1)
Ground truth, n (%)			0.001	0.972
Low-grade	55 (79.7)	27 (79.4)
High-grade	14 (20.3)	7 (20.6)

Table 3 The sensitivity, specificity, positive predictive value, and negative predictive value of the prediction by radiological model, radiomics model, and nomogram model with their 95% confidential intervals

Model	AUC	Sensitivity	Specificity	PPV	NPV
Radiological training	0.753 (0.597-0.909)	42.9 (17.7-71.1)	96.4 (87.5-99.6)	75.0 (34.9-96.8)	86.9 (75.8-94.2)
Radiological test	0.642 (0.379-0.870)	71.4 (29.0-96.3)	66.7 (46.0-83.5)	35.7 (12.8-64.9)	90.0 (68.3-98.8)
Radiomic training	0.919 (0.828-1.000)	92.9 (66.1-99.8)	80.0 (67.0-89.6)	54.2 (32.8-74.4)	97.8 (88.2-99.9)
Radiomic test	0.881 (0.772-0.990)	100.0 (59.0-100.0)	66.7 (46.0-83.5)	43.7 (19.8-70.1)	100.0 (81.5-100.0)
Nomogram training	0.916 (0.801-1.000)	85.7 (57.2-98.2)	90.9 (80.0-97.0)	70.6 (44.0-89.7)	96.2 (86.8-99.5)
Nomogram test	0.894 (0.773-1.000)	100.0 (59.0-100.0)	66.7 (46.0-83.5)	43.7 (19.8-70.1)	100.0 (81.5-100.0)

AUC: Area under the curve; NPV: Negative predictive value; PPV: Positive predictive value.