Observer ability to detect ex vivo simulated internal or external cervical root resorption

Abstract

Introduction: The objective of this study was to assess observer ability to detect ex vivo simulated internal and external cervical root resorption.

Methods: A total of 90 single-rooted mandibular anterior teeth were split along the coronal plane. By using a 0.5-mm diameter round bur, 50 of the teeth were drilled to simulate resorptive cavities (25 internal and 25 external cervical resorption), and the remaining 40 teeth were left without surface defects. The sectioned teeth were rejoined and placed in the alveolar sockets of a dry human mandible in groups of 6 and imaged with intraoral film by using 3 different angulations and with a cone-beam computed tomography (CBCT) unit. Film and CBCT image sets were evaluated twice by 3 separate observers. Teeth were scored for the presence and location of resorption by using a 5-point scale. Intraobserver and interobserver kappa coefficients and Az values were calculated for each observer and image set. Differences between modalities were compared by using z statistics, with the significance level set at α = 0.05.

Results: Both intraobserver and interobserver agreements were statistically higher (p < .05) for the Iluma CBCT images than for the intraoral images. Az values for CBCT images were also statistically higher (p < .05) than for film images for all observers and readings. In addition, kappa and Az values of external cervical resorption cavities were statistically higher (p < .05) than those of internal cervical resorption cavities for all observers, image types, and readings.

Conclusions: High-resolution Iluma CBCT images performed better than film in the ex vivo detection and localization of simulated internal and external cervical root resorption.