Skip to main content

Table 5 Results in the validation set for the classification of lesions versus normal tissue and malignant versus benign solid lesions using the full texture feature and the reduced feature set

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

 

Predicted

Actual

Lesions versus normal tissue (n = 105)

Lesions (%)

Normal tissue (%)

Lesions (n = 54)

 Full feature set

50 (92.6)

4 (7.4)

 Reduced feature set

52 (96.3)

2 (3.7)

Normal tissue (n = 51)

 Full feature set

3 (5.9)

48 (94.1)

 Reduced feature set

4 (7.8)

47 (92.2)

Malignant versus benign lesions (n = 54)

Malignant (%)

Benign (%)

Malignant (tot = 27)

 Full feature set

23 (85.2)

4 (14.8)

 Reduced feature set

22 (81.5)

5 (18.5)

Benign (tot = 27)

 Full feature set

1 (3.7)

26 (96.3)

 Reduced feature set

2 (7.4)

25 (92.6)