Development and Phantom Evaluation of a Closed-Mouth Premaxillary Projection for Odontoid Imaging

Main Article Content

Sang-Hyun Kim, Dong-Hee Hong

Abstract

This study aimed to develop a closed-mouth premaxillary projection for odontoid imaging by determining an appropriate cephalad tube angle from CT-based craniofacial measurements and evaluating the resulting image quality using skull phantoms. Facial bone CT images from 150 patients without a history of spinal disease or acute trauma were retrospectively analyzed. The angles between the orbitomeatal line (OML), infraorbitomeatal line (IOML), external auditory meatus, dens, and acanthion were measured to establish candidate projection angles. Phantom radiographs were then obtained using three skull phantoms with cephalad tube angles ranging from 1° to 11° under standardized exposure conditions. Quantitative image quality was assessed using SNR and CNR, while qualitative image quality was independently evaluated by five radiologic technologists with more than 10 years of clinical experience using a five-point scale. Inter-rater reliability was assessed using ICC. The mean OML-IOML angle was 7.82 ± 1.07°, while the mean dens-to-acanthion angles were 12.4 ± 1.33° for OML-based positioning and 4.57 ± 1.09° for IOML-based positioning. No statistically significant differences in SNR or CNR were detected among the tested projection angles for either OML- or IOML-based positioning. However, the 4° cephalad projection received the highest qualitative image-quality scores for both OML and IOML, with mean scores of 4.02 ± 0.30 and 4.03 ± 0.38, respectively. Inter-rater reliability was acceptable to excellent across the evaluated angles. These findings suggest that a 4° cephalad closed-mouth premaxillary projection may provide diagnostically acceptable visualization of the odontoid process when standard open-mouth positioning is not feasible. Further prospective studies comparing this technique with conventional open-mouth radiography in patients are required before clinical implementation.

Article Details

Section
Articles