Reliability and Reproducibility of Manual and Automated Volumetric Measurements of Periapical Lesions

Introduction: The present study aimed to test the reliability and reproducibility of 2 methods: manual and automated segmentation (using a threshold-based region growing algorithm) for measuring the volume of periapical lesions. Methods: A total of 43 cone-beam computed tomographic scans (14 men and 29 women, mean age of 54.6 +/- 8.5 years) were analyzed by 3 observers. Intraobserver reproducibility and inter-observer reliability were assessed using the intraclass correlation coefficient. Parametric correlation between manual and automated volumetric measurements was performed. In addition, the Student t test was also used to compare the mean time required for manual and automated volumetric measurements. Results: Automated segmentation showed slightly higher intraclass correlation coefficient values for all observers. A strong significant correlation was found between manual and automated volumetric measurements performed by the 3 observers. A significant difference in the mean procedure time was also found between both methods (P = .001). Conclusions: Within the limitations of this study, the present results suggest that automated segmentation with a region growing algorithm is faster and slightly more reliable to calculate the volume of periapical lesions.