01.12.2016  Original investigation  Ausgabe 1/2016 Open Access
The relationship between HbA_{1c} and ultrasound plaque textures in atherosclerotic patients
 Zeitschrift:
 Cardiovascular Diabetology > Ausgabe 1/2016
Background
Subjects and methods
Subjects
Carotid ultrasonography
Texture analysis
Histogram (1) mean, (2) variance, (3) skewness, (4) kurtosis and (5) percentiles 1, 10, 50, 90 and 99 % 
Absolute gradient (1) mean, (2) variance, (3) skewness, (4) kurtosis and (5) percentage of pixels with nonzero gradient 
Runlength matrix (1) runlength nonuniformity, (2) greylevel nonuniformity, (3) longrun emphasis, (4) short run emphasis and (5) fraction of image in runs. Parameters computed for horizontal, 45° vertical and 135° orientation 
Cooccurrence matrix (1) angular second moment, (2) contrast, (3) correlation, (4)sum of squares, (5) inverse difference moment, (6) sum average, (7) sum variance, (8) sum entropy, (9) entropy, (10) difference variance and (11) difference entropy. Parameters are computed for 4 orientations: (a, 0), (0, a), (a, a), (a, −a) and 5 distances: a = 1, 2, 3, 4, 5; between image pixels 
Autoregressive model (1) model parameter vector includes 4 parameters and (2) standard deviation of the driving noise 
Wavelet (1) Energy of wavelet coefficients in lowfrequency subbands, (2) horizontal highfrequency subbands, (3) vertical highfrequency subbands and (4) diagonal highfrequency subbands at successive scales 
Texture feature selection and reduction
Fisher coefficient
Mutual information measure
Linear discriminant analysis
Statistical analysis
Results
Baseline characteristics of study participants
Characteristics  Diabetes (n = 66)  Nondiabetes (n = 70) 
t

p


Age, mean (SD)  67.8 ± 10.0  69.4 ± 9.8  −0.972  0.333 
Male gender, n (%)  43 (65)  46 (66)  0.068  0.946 
Total number of plaques  2.1 ± 1.2  2.0 ± 1.1  0.533  0.104 
Hypertension, n (%)  43 (65)  44 (63)  0.277  0.783 
CHD, n (%)  18 (27.3)  22 (31.4)  −0.528  0.598 
SBP (mm Hg)  136.9 ± 19.0  139.1 ± 22.0  −0.642  0.522 
DBP (mm Hg)  75.8 ± 9.2  76.7 ± 12.8  0.465  0.643 
BMI (kg/m^{2})  23.9 ± 3.1  22.0 ± 2.8  3.437  0.001 
Total cholesterol (mmol/L)  4.78 ± 1.20  4.46 ± 1.17  −1.548  0.116 
Triglyceride (mmol/L)  1.90 ± 1.57  1.31 ± 0.89  2.677  0.009 
HDLcholesterol (mmol/L)  1.05 ± 0.28  1.13 ± 0.34  −1.484  0.140 
LDLcholesterol (mmol/L)  3.04 ± 0.97  2.86 ± 1.06  1.023  0.308 
Lipoprotein A (mg/L)  211.9 ± 219.2  284.9 ± 288.7  −1.652  0.101 
Apolipoprotein A1(g/L)  1.29 ± 0.31  1.31 ± 0.26  −0.362  0.718 
Apolipoprotein B100 (g/L)  1.20 ± 0.47  1.17 ± 1.23  0.163  0.871 
HbA_{1c} (%)  8.87 ± 2.25  5.50 ± 0.41  11.940  <0.001 
FPG (mmol/L)  10.76 ± 4.26  5.14 ± 0.70  10.611  <0.001 
PlaqueIMTmax (mm)  2.4 ± 0.8  2.5 ± 0.8  −0.635  0.527 
Plaques selection
Texture selection and reduction
Features  Nondiabetes (n = 70)  Diabetes (n = 66)  Fisher  MI 
t

p


WavEnLH_s4  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_s2  16,327 ± 4554  13,092 ± 7191  0.229  0.055  2.2733  0.007 
WavEnLL_s1  16,327 ± 4559  13,474 ± 7337  0.224  0.041  2.705  0.008 
WavEnLH_s3  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_s3  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 
Pearson correlation of the study variables with HbA_{1c}
Age (year)  BMI (kg/m^{2})  TNP  SBP (mmHg)  DPB (mmHg)  PlaqueIMTmax (mm)  MDF  

HbA_{1c} (%)  Pearson  −0.123  0.182  −0.022  −0.071  0.084  −0.097  −0.348 
p

^{0.155}

^{0.034}

^{0.795}

^{0.408}

^{0.333}
 0.262 
^{<0.001}

Linear regression model for HbA_{1c}
Model  Variables/constant  Coefficients 
P
 95.0 % confidence interval  

Lower bound  Upper bound  
HbA_{1c} (r = 0.392, p < 0.001)  Constant  7.136  <0.001  6.766  7.506 
MDF  −13.297  <0.001  −19.407  −7.186 