Skip to main content

Table 5 Diagnostic performances of texture features to predict COD

From: Differential diagnosis of cemento-osseous dysplasia and periapical cyst using texture analysis of CBCT

Texture features (GLCM)

Threshold criterion

TP

FP

FN

TN

Sensitivity (%) (95% CI)

Specificity (%) (95% CI)

Accuracy (95% CI)

AUC (95% CI)

S(4,4) Sum of Average

≥0.610

23

1

2

24

0.92 (0.876, 0.947)

0.96 (0.921, 0.995)

0.94 (0.912, 0.957)

0.9792 (0.965, 0.993)

S(1,0) Sum of Entropy

≥0.270

24

4

1

21

0.96 (0.889, 0.988)

0.84 (0.775, 0.937)

0.9 (0.865, 0.93)

0.9712 (0.956, 0.986)

S(0,1) Sum of Entropy

≥0.300

24

5

1

20

0.96 (0.825, 1.00)

0.8 (0.721, 0.952)

0.88 (0.843, 0.915)

0.96 (0.94, 0.979)

S(1,0) Entropy

≥0.560

23

2

2

23

0.92 (0.851, 0.948)

0.92 (0.884, 0.997)

0.92 (0.889, 0.95)

0.9696 (0.949, 0.99)

S(0,1) Entropy

≥0.500

23

2

2

23

0.92 (0.875, 0.949)

0.92 (0.868, 0.97)

0.92 (0.886, 0.945)

0.9616 (0.937, 0.987)

S(2,0) Difference of Entropy

≥0.460

24

1

1

24

0.96 (0.944, 0.956)

0.96 (0.926, 0.994)

0.96 (0.939, 0.972)

0.9632 (0.932, 0.994)

S(2,2) Difference of Entropy

≥0.530

24

1

1

24

0.96 (0.944, 0.956)

0.96 (0.926, 0.994)

0.96 (0.939, 0.972)

0.9648 (0.935, 0.995)

S(3,0) Difference of Entropy

≥0.430

24

1

1

24

0.96 (0.944, 0.956)

0.96 (0.926, 0.994)

0.96 (0.939, 0.972)

0.9696 (0.944, 0.996)

S(3,3) Difference of Entropy

≥0.580

24

1

1

24

0.96 (0.944, 0.956)

0.96 (0.926, 0.994)

0.96 (0.939, 0.972)

0.9664 (0.938, 0.995)

S(4,0) Difference of Entropy

≥0.570

24

1

1

24

0.96 (0.944, 0.956)

0.96 (0.926, 0.994)

0.96 (0.939, 0.972)

0.9712 (0.947, 0.996)

  1. COD cemento osseous dysplasia, TP true positive, FP false positive, FN false negative, TN true negative, AUC area under the curve, GLCM gray level co-occurrence matrix