M.Sc Thesis

M.Sc StudentHalperin Michal
SubjectAccuracy of Cone Beam Computerized Tomography (CBCT) for
the Measurement of Linear Distances in the
Mandible, an Ex-Vivo Study
DepartmentDepartment of Medicine
Supervisors DR. Jacob Horwitz
Full Thesis textFull thesis text - English Version


Cone-Beam Computed Tomography (CBCT) has become a prominent tool for dental implant planning. It provides a three-dimensional image of the alveolar bone architecture and its surrounding anatomical structures, resulting in predictable treatment outcomes. However, it has been suggested that radiographic measurements based on CBCT are not as accurate as was previously thought, which may result in treatment planning errors that might have adverse consequences. Thus, the aim of this study was to compare the accuracy of linear measurements performed on CBCT images to direct linear measurements.

To this end, eleven domestic porcine mandibles were utilized. Six cross-sectional planes were defined in each mandible by Gutta-Percha filled drillings. Mandibles were scanned in a CBCT apparatus and then sectioned in the six marked planes. Four direct measurements (DIR) were recorded in each plane by two examiners. The cross-sectional planes were then identified in the CBCT image and same respective radiographic measurements (RAD) were performed. Differences between RAD and DIR measurements (∆(RAD-DIR)) and their absolute values (∆(RAD-DIR)│) were calculated and expressed as mean?SD. The magnitude and distribution of the differences were compared between measurements characterized as small and large in magnitude as well as between measurements performed in the anterior and posterior areas of the mandible.

A total of 162 pairs of measurements were included. ∆(RAD-DIR)│ was 0.45?0.33 mm (Range: 0 and 1.42 mm). In 35.2% of the cases (95% confidence interval: 27.8%-42.5%), │∆(RAD-DIR)│ was larger than ?0.5 mm and in 7.4% of the cases (95% confidence interval: 3.4%-11.4%), │∆(RAD-DIR)│ exceeded ?1 mm. ∆(RAD-DIR) was -0.17?0.53 mm (range: -1.42 and 1.09 mm), indicating a mean underestimation of RAD relative to DIR measurements. 36% of cases were positive, indicating overestimation of CBCT. In 8% of the cases (95% confidence interval: 3.8%-12.2%), ∆(RAD-DIR) was between 0.5 and 1 mm, and in 1.8% of the cases (95% confidence interval: -0.2%-3.9%), ∆(RAD-DIR) was greater than 1 mm. A statistically significant difference was found in ∆(RAD-DIR)│  and ∆(RAD-DIR) between small and large measurements (P=0.001 and P=0.01, respectively). However, no statistically significant differences were observed between anterior and posterior measurements (P=0.634 and P=0.284, respectively).  

A mean underestimation of RAD compared to DIR linear measurements was found in the present study. However, in an appreciable portion of the cases there was overestimation equal to or greater than 0.5 mm. This indicates a potential error and a need for safety margins, especially when utilizing CBCT for dental implant planning.