Evaluation of Radiographic Orthodontic Records Image Quality Derived from CBCT

A study was performed to determine the use of a specific reconstruction algorithm (RA), adaptive-steepest-descent-projection-onto-convex-sets (ASD-POCS) on the image quality of orthodontic radiographic images constructed from two cone beam computed tomography (CBCT) scans. Four different image reconstructions were prepared using ASD-POCS. Reconstructed images varied among each other in the number of basis images. These were RA 300, same as the i-CAT image (control) and reducing the number of basis images to RA 150, RA 79, and RA 33. All the CBCT images with different numbers of basis images were used to reconstruct the multiplanar images of the panoramic view, lateral cephalogram, and posterior-anterior cephalogram manually using Dolphin 3D software. Seventy three examiners evaluated the reconstructed 2D radiographs (panoramic, lateral, and posterior-anterior cephalogram) for the purpose of general screening, diagnosis, and treatment planning from the reconstructed CBCT images. The examiners were orthodontists, orthodontic residents, oral surgeons, oral surgery residents and oral and maxillofacial radiologists. Using the preferences (Yes/No) and 10 point rating scale the examiners evaluated if the images were diagnostic for screening, diagnosis, and treatment planning by identification of specified landmarks. Non-parametric tests were used to determine statistical significant differences. Preferences of all the images among participants were statistically significant (p<0.05). Mostly, for all 2D reconstructed images, RA 300 was preferred over i-CAT (control) and had the highest mean rank. For the panoramic radiograph and posterior-anterior radiograph, it was found that the quality of the images with the full number of basis images RA 300 and RA 150 were not significantly different from the i-CAT reconstructed images. For the lateral cephalogram, it was found that the images generated from RA 300, RA 150, and RA 79 number of basis images were not significantly different, resulting in usable images with up to one half and one fourth reduction in number of projections. This shows that reconstruction algorithm ASD-POCS could be used to reduce the number of basis images while maintaining the image quality for the purpose of 2D orthodontic records.