Validation of periapical radiographs to detect orthodontically induced root resorption

Objectives: To test whether apical root resorption during orthodontic tooth movement is detectable on periapical radiographs. Methods: In a standardized experimental protocol, 29 premolars in 16 subjects were orthodontically moved in a buccal direction during 8 weeks, applying 1 N force. 19 contralateral premolars not subjected to orthodontic tooth movement served as controls. Standardized periapical radiographs were taken before and after the experimental period (Rx-method). These teeth were extracted and scanned in a micro-CT scanner with a resolution of 9 μm. Two calibrated examiners assessed blindly the presence or absence of apical root resorption on the digitized radiographs, and evaluated the presence or absence of apical root resorption on the 3D reconstruction of the scans. To compare the differences between the experimental and control group and the two methods (Rx vs. CT-scanner), the Pearson's Chi-Square test was performed. By taking the scanner images as the gold-standard, the validity of the Rx-method was assessed. The specificity and sensitivity of the Rx-method were calculated. Results: Significant differences were detected between orthodontically moved teeth and controls. 86 % of orthodontically moved teeth showed apical root resorption when using the CT-scanner as a diagnostic tool, compared to 5 % of control teeth. When evaluating the results with the Rx-method, the presence of apical root resorption was visible radiographically in 55 % of the orthodontically moved teeth and in 5 % of the control teeth. The Rx-method showed a high specificity of 90% and a relatively low sensitivity of 53 %, which means that only half of the cases with apical root resorption identified by the CT-scanner were also identified by periapical radiographs. Conclusion: Nearly all the orthodontically moved teeth showed apical root resorption. Apical root resorption may be underestimated when evaluated by periapical radiographs. Supported by the Swiss National Science Foundation (Grant No. 3200B0-112296).