CBCT and MRI Demonstrate Comparable Soft Palate Measurements

Objectives: This study compared soft palate volume, length, thickness, and angle measurements using cone beam computed tomography (CBCT) and 3-Tesla (3T) magnetic resonance imaging (MRI) scans.
Methods: This retrospective study involved twelve human subjects, all of whom had a whole head CBCT scan (NewTom 5G, AFP Imaging, USA) and a whole head MRI scan (Siemens Medical Solutions, DE) performed within a two-week time period. CBCT images were taken with a field of view of 18x16 inches and resolution of 0.3x0.3x0.3mm, over an exposure time of 5.4 seconds. MR images were acquired using a 3.0T imaging system with a 12-channel head array coil and a resolution of 1.0x1.0x1.0mm, over an exposure of time of 4-5 minutes. Patients were imaged in a supine position for both CBCT and MRI scans. Simpleware ScanIP software was used for analysis of the resultant DICOM images. After registration and superimposition in three planes of orientation, the soft palate was segmented and measured. Statistical analysis was performed using Intraclass Correlation Coefficient (ICC) analysis and Bland-Altman plots.
Results: Excellent agreement was observed between CBCT and MRI for volume, length, thickness, and angle measurements of the soft palate. Mean Intra-observer ICC was 0.986-.998 for CBCT and 0.949-0.998 for MRI measurements. Dimensional ICC between CBCT and MRI ranged from 0.743 (angle) - 0.998 (volume). Assessment of bias ranged from 0.0243 (thickness) - 41.9 (volume). No statistically significant systematic bias (p>0.05) was observed when comparing CBCT and MRI measurements.
Conclusions: Soft palate measurements made on CBCT and MRI demonstrate comparable volume, length, thickness and angle measurements. In addition, measurement bias was observed to be smaller with larger soft palate measurements. These results indicate that further studies are warranted to characterize the interchangeability of CBCT and MRI for the purpose of soft palate analysis.