Noninvasive model for predicting future ischemic strokes in patients with silent lacunar infarction using radiomics

Table 1 Statistical analysis of the clinical characteristics of the training and validation data sets

Characteristics	Training cohort (n = 74)	Validation cohort (n = 74)	P-value
Sex			0.49
Male	40 (54.1)	46 (62.2)
Female	34 (45.9)	28 (37.8)
Age, years	75 [68–82]	77 [72–82]	0.10
< 76 years	39 (52.7)	28 (37.8)
≥ 76 years	35 (47.3)	46 (62.2)
Multiple infarctions			0.61
Yes	31 (41.9)	27 (36.5)
No	43 (58.1)	47 (63.5)
Smoking status			0.68
Yes	14 (18.9)	16 (21.6)
No	60 (81.1)	58 (78.4)
Alcohol abuse			0.27
Yes	2 (2.7)	6 (8.1)
No	72 (97.3)	68 (91.9)
Hypertension			0.84
Yes	57 (77.0)	58 (78.4)
No	17 (23.0)	16 (21.6)
Diabetes mellitus			0.99
Yes	28 (37.8)	33 (44.6)
No	46 (62.2)	41 (55.4)
Dyslipidemia			0.29
Yes	57 (77.0)	63 (85.1)
No	17 (23.0)	11 (14.9)
Internal carotid plaque^a			0.79
Yes	59 (86.8)	61 (89.7)
No	9 (13.2)	7 (10.3)
Internal carotid stenosis^a			0.36
Yes	4 (5.9)	8 (11.8)
No	64 (94.1)	60 (88.2)
Using statin strategies			0.80
Yes	10 (13.5)	9 (12.2)
No	64 (86.5)	65 (87.8)
Using antithrombotic therapies			0.77
Yes	11 (7.4)	10 (6.7)
No	63 (42.6)	64 (43.3)

Values are presented as n (%) or median (range)
Internal carotid plaque^a and internal carotid stenosis were only collected from 136 patients who had initial carotid artery ultrasound reviews
The differences in the patients’ ages were evaluated using the Mann–Whitney U test
The distribution of the other characteristics was assessed using the χ² test