Comparison of Accuracy and Observer Agreement in the Detection of Simulated External Root Resorption Using Conventional Digital Radiography and Digitally Filtered Radiography

Introduction: External root resorption is a clinical problem that often cannot be detected clinically. Thus, radiography plays a crucial role in its diagnosis. However, optimal radiographic quality with minimal radiation exposure {2.1 [EN] Verify English word/phrase choice} is an important factor in selecting the appropriate radiographic technique. The aim of this study was the comparison of accuracy and observer agreement in the detection of simulated external root resorption using conventional digital radiography and digitally filtered radiography. Materials and methods: The study was performed using 100 single rooted teeth in a dry mandibular jaw. The teeth were divided into 5 groups as follows: (1) without resorption, (2) resorption with 0.25 mm depth on buccal surface, (3) resorption with 0.5 mm depth on buccal surface, (4)resorption with 0.25 mm depth on proximal surfaces, and (5) resorption with 0.5 mm depth on proximal surfaces. Digital radiographic images, both conventional and digitally filtered using 3 filters (diagonal, horizontal, and vertical), were obtained. The 2 groups of images were then evaluated by 4 observers in 2 stages, with an interval of 1 week. Thereafter, sensitivity, accuracy, specificity, and Kappa coefficients were calculated to assess observer agreement. Results: For digitally filtered radiography with diagonal, horizontal, and vertical filters, the accuracy values were 86.5%, 87.2%, and 89.2%; sensitivity values were 93.1%, 92.7%, and 94.3%; and specificity values were 82.5%, 87.5%, and 93.7%, respectively. The accuracy, sensitivity, and specificity values for conventio digital radiography were 86.5%, 91.5%, and 82.5%, respectively. Conclusion: Manipulation of images in digital radiographic systems may not always facilitate diagnosis.