Skip to main content

Table 1 Overview of the baseline characteristics of all study cohort

From: Comparison of deep learning-based image segmentation methods for intravascular ultrasound on retrospective and large image cohort study

 

Training cohort, n = 79

Validation cohort, n = 11

Testing cohort, n = 23

P-value

Patients with CAD, n (%)

78 (98.73)

10 (90.91)

22 (95.65)

0.27

Males, n (%)

45 (56.96)

7(63.64)

15 (65.22)

0.74

Age, mean ± SD

68.46 ± 8.86

73.0 ± 3.56

69.0 ± 4.44

0.40

Age > 60, n (%)

30 (37.98)

3 (27.27)

6 (26.09)

0.50

Involved vessel

 LAD, n (%)

50 (63.29)

8 (72.73)

16 (69.56)

0.74

 LCX, n (%)

7 (8.86)

0 (0)

2 (8.70)

0.59

 RCA, n (%)

13 (16.46)

3 (27.27)

5 (21.74)

0.63

Pullbacks

79

11

23

< 0.001

Images

7808

1073

2189

< 0.001

IVUS image categories

 Calcified plaque, n (%)

2425 (31.06)

393 (36.63)

651 (29.74)

< 0.001

 Bifurcation, n (%)

338 (4.33)

55 (5.13)

145 (6.62)

< 0.001

 Adjacent vessels, n (%)

916 (11.73)

199 (18.55)

197 (9.00)

< 0.001

 Stent, n (%)

171 (2.19)

62 (5.78)

60 (2.74)

< 0.001

 Guidewire artifacts, n (%)

2518 (32.25)

223 (20.78)

945 (43.17)

< 0.001

 Lipid fibrous plaque, n (%)

3962 (50.74)

467 (43.52)

1109 (50.67)

< 0.001

 None, n (%)

1411 (18.07)

203 (18.92)

422 (19.28)

0.39

  1. Analysis of variance (ANOVA) was used as the test for the difference of mean for continuous variables such as Age, Pullbacks and Images. Chi-squared test was used to test for the difference of proportion for categorical variables such as Patient with CAD, Males, Age > 60, Involved vessel and IVUS image category. Multiple categories may exist for the same IVUS image
  2. CAD coronary artery disease, LAD left anterior descending, LCX left circumflex, RCA right coronary artery