Browsing by Author "Ghoncheh, Zahra"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Accuracy of Digitally Enhanced Images Compared with Unprocessed Digital Images in The Detection of External Root Resorption
    (2017) Ghoncheh, Zahra; Afkhami, Farzaneh; Fard, Mohammad Javad Kharazi; Sorkhabi, Rasa Ebrahimi; Aydin, Ulkem; https://orcid.org/0000-0002-2130-2812
    This study was for comparing the accuracy of processed digital images (reverse-contrast and colorization) with that of unprocessed digital images in detection of external root resorption (ERR). Eighty single-rooted human teeth were selected for this study. Mild, moderate, and severe ERR were simulated on 20 teeth each, and 20 were left untreated. Digital images using the paralleling technique were made, and three types of images were finally produced: unprocessed, reverse-contrast, and colorized. Three experienced dentists examined the images. The Wilson confidence intervals were calculated to analyze the diagnostic data. The kappa statistic was used to determine interobserver agreement. For unprocessed images, the rate of correct classification of mild and moderate to severe ERR was 88.3 and 80.0 %, respectively. The corresponding rate for reverse-contrast images was 81.7 and 80.0 %, and that for colorized images was 93.3 and 80.0 %, respectively. The sensitivity of unprocessed images in the detection of mild and moderate to severe ERR was 0.93 and 0.84, respectively. The corresponding sensitivity for reverse-contrast images was 0.83 and 0.84, and that for colorized images was 0.93 and 0.84, respectively. The specificity of unprocessed, reverse-contrast, and colorized images was 0.90, 0.92, and 1.00, respectively. The kappa coefficient for interobserver agreement was 0.86 for unprocessed images, 0.88 for reverse-contrast images, and 0.89 for colorized images. The difference between the sensitivity and specificity of unprocessed, reverse-contrast, and colorized images was not statistically significant (p > 0.05). The three techniques were of similar and desirable accuracy in detection of ERR.