Skip to main content

Table 3 Statistically different features between diabetic and non-diabetic plaques

From: The relationship between HbA1c and ultrasound plaque textures in atherosclerotic patients

Features

Non-diabetes (n = 70)

Diabetes (n = 66)

Fisher

MI

t

p

WavEnLH_s-4

756.21 ± 418.35

522.48 ± 385.45

0.342

0.109

3.383

0.001

Vertl_Fraction

0.883 ± 0.056

0.847 ± 0.096

0.213

0.036

2.630

0.010

135dr_ShrtREmp

0.921 ± 0.037

0.893 ± 0.0716

0.25

0.066

2.866

0.005

Vertl_ShrtREmp

0.915 ± 0.040

0.886 ± 0.072

0.243

0.046

2.810

0.006

WavEnLL_s-2

16,327 ± 4554

13,092 ± 7191

0.229

0.055

2.2733

0.007

WavEnLL_s-1

16,327 ± 4559

13,474 ± 7337

0.224

0.041

2.705

0.008

WavEnLH_s-3

465.39 ± 275.86

343.30 ± 247.33

0.220

0.039

2.712

0.008

135dr_Fraction

0.892 ± 0.052

0.858 ± 0.091

0.213

0.040

2.627

0.010

45dgr_ShrtREmp

0.918 ± 0.038

0.892 ± 0.073

0.210

0.038

2.606

0.011

WavEnLL_s-3

15,479 ± 4811

12,802 ± 7299

0.193

0.069

2.510

0.014

Horzl_ShrtREmp

0.816 ± 0.095

0.757 ± 0.169

0.191

0.048

2.491

0.014

45dgr_Fraction

0.887 ± 0.053

0.854 ± 0.095

0.184

0.040

2.444

0.016

MDF

0.033 ± 0.050

−0.035 ± 0.052

7.712

<0.001

  1. WavEnLH_s-3 and WavEnLH_s-4 are energy of wavelet coefficients in vertical high-frequency subbands at scale 3 and 4; WavEnLL_s-1, WavEnLL_s-2 and WavEnLH_s-3 are energy of wavelet coefficients in low-frequency subbands at scale 1, 2 and 3; 45dgr_Fraction, Vertl_Fraction and 135dr_Fraction are fraction of image in runs computed for 45°, vertical and 135° orientation; Vertl_ShrtREmp, Horzl_ShrtREmp, 45dgr_ShrtREmp, Vertl_ShrtREmp and 135dr_ShrtREmp are short run emphasis computed for horizontal, 45° vertical and 135° orientation
  2. MDF most discriminating feature, MI mutual information