Skip to main content

Table 4 Textural features comparing alveolar hemorrhages (A), PJ-pneumonias (PJP) (B) and. HSV1 = herpes simplex virus-1 pneumonias (C). A value of p < 0.025 was considered significant. The scale was selected by tuning the fine filter parameter (fine texture features of 4 pixels in width)

From: Potential role of CT-textural features for differentiation between viral interstitial pneumonias, pneumocystis jirovecii pneumonia and diffuse alveolar hemorrhage in early stages of disease: a proof of principle

 

p

alveolar haemorrhages vs. PJ-pneumonias

p

PJ-pneumonias vs. herpes simplex virus-1 pneumonias

p

hemorrhages vs.herpes simplex virus-1 pneumonias

Medium filter

Mean

Entropy

Uni-formity

Mean

Entropy

Uni-formity

Mean

Entropy

Uni-formity

Heterogeneity

0.015*

0.568

0.991

0.420

0.091

0.054

0.155

0.212

0.144

Intensity

0.485

0.969

0.502

0.260

0.130

0.097

0.671

0.210

0.096

Average

0.450

0.527

0.879

0.224

0.029

0.052

0.653

0.159

0.067

Deviation

0.046

0.065

0.128

0.036

0.033

0.061

0.803

0.382

0.173

Skewness

0.617

0.060

0.006*

0.095

0.862

0.287

0.543

0.133

0.169

Entropy (Co-occurrence Matrix)

0.432

0.722

0.592

0.133

0.089

0.085

0.533

0.120

0.183

Difference Variance (Co-occurrence Matrix)

0.584

0.621

0.601

0.858

0.098

0.095

0.469

0.345

0.113

Number non-uniformity (NGLDM)

0.583

0.699

0.385

0.722

0.034

0.047

0.410

0.036

0.143

Entropy (NGLDM)

0.043

0.886

0.530

0.261

0.143

0.076

0.610

0.119

0.087

Contrast (NGTDM)

0.635

0.836

0.287

0.219

0.040

0.095

0.526

0.123

0.248

  1. Abrreviations: NGLDM Neighboring Grey-Level Dependence Matrix.