Table 1 Diagnostic accuracy of the different image types for detection of simulated implant fractures.

From: A diagnostic accuracy study on an innovative auto-edge detection technique for identifying simulated implant fractures on radiographic images

Image type

AUC

SEa

95% CIb

p value

Sensitivity

Specificity

PPA original

0.792

0.051

0.65–0.89

 < 0.001

0.58

1.00

OPA original

0.792

0.051

0.65–0.89

 < 0.001

0.58

1.00

CBCT MAR (on) original

0.917

0.041

0.80–0.98

 < 0.001

0.92

0.92

CBCT MAR (off) original

0.875

0.045

0.75–0.95

 < 0.001

0.75

1.00

PPA Canny

0.833

0.049

0.70–0.92

 < 0.001

0.67

1.00

OPA Canny

0.875

0.045

0.75–0.95

 < 0.001

0.75

1.00

CBCT MAR (on) Canny

0.958

0.029

0.86–0.99

 < 0.001

1.00

0.92

CBCT MAR (off) Canny

0.875

0.045

0.75–0.95

 < 0.001

0.75

1.00

  1. AUC area under the ROC curve, aDeLong test, bexact binominal test, SE standard error, CI confidence interval.
Back to article page