Skip to main content

Table 2 Texture features that showed significantly different mean values comparing lesions (benign and malignant) versus normal tissue (fat and fibroglandular) and corresponding area under the curve (AUC)

From: Diagnostic performance of machine learning applied to texture analysis-derived features for breast lesion characterisation at automated breast ultrasound: a pilot study

Feature Lesions (mean ± standard deviation) Normal tissue (mean ± standard deviation) p value AUC (95% confidence interval)
Entropy 5.48 ± 0.35 5.65 ± 0.24 < 0.00001 for all 0.67 (0.63–0.71)
Variance 153.53 ± 23.59 137.75 ± 20.12 0.70 (0.66–0.73)
Contrast 24.10 ± 9.87 40.10 ± 12.5 0.84 (0.81–0.87)
Correlation 0.88 ± 0.06 0.81 ± 0.06 0.80 (0.65–0.83)
Energy 3.4 × 10−3 ± 1.5 × 10−3 2.2 × 10−3 ± 0.7 × 10−3 0.83 (0.80–0.86)
Homogeneity 0.36 ± 0.05 0.33 ± 0.04 0.72 (0.69–0.76)
Contrast 26.43 ± 12.38 40.94 ± 13.41 0.80 (0.76–0.83)
Correlation 0.87 ± 0.06 0.81 ± 0.07 0.78 (0.74–0.82)
Energy 3.9 × 10−3 ± 1.4 × 10−3 2.6 × 10−3 ± 0.7 × 10−3 0.86 (0.82–0.88)
Homogeneity (GLCM) 0.36 ± 0.05 0.33 ± 0.04 0.71 (0.66–0.74)
GLN (GLCM) 78.42 ± 77.69 156.28 ± 104.76 0.79 (0.76–0.83)
RLN (GLCM) 1,870.43 ± 1,792.01 4,085.00 ± 2,628.83 0.82 (0.79–0.85)
LRHGE (GLCM) 1,706.53 ± 209.23 1,635.50 ± 197.26 0.61 (0.57–0.65)
SRE (GLRLM) 0.91 ± 0.02 0.92 ± 0.04 0.69 (0.65–0.73)
LRE (GLRLM) 1.47 ± 0.17 1.39 ± 0.12 0.65 (0.61–0.69)
GLN (GLRLM) 78.53 ± 77.67 156.36 ± 104.73 0.79 (0.76–0.83)
RLN (GLRLM) 1,871.42 ± 1,786.51 4,080.90 ± 2,622.01 0.82 (0.79–0.85)
LRHGE (GLRLM) 1,677.30 ± 196.74 1,598.70 ± 168.52 0.61 (0.57–0.65)
LZE 5.38 ± 3.34 4.07 ± 2.38 0.66 (0.62–0.70)
HGZE 1,221.08 ± 45.52 1,191.37 ± 86.03 0.67 (0.63–0.71)
  1. GLCM Grey-level co-occurrence matrix, GLN Grey-level non-uniformity, RLN Run length non-uniformity, LRHGE Long-run high grey-level emphasis, GLRLM Grey-level run length matrix, SRE Short-run emphasis, LRE Long-run emphasis, LZE Large zone emphasis, HGZE High grey-level zone emphasis