Related Experiment Video
Updated: Jul 11, 2026
Reliability of Artificial Intelligence-Based Cone Beam Computed Tomography Integration with Digital Dental Images
Published on: February 23, 2024
Accuracy of linear measurements from imaging plate and lateral cephalometric images derived from cone-beam computed
Mazyar Moshiri1, William C Scarfe, Michael L Hilgers
1School of Dentistry, University of Louisville, Louisville, Ky 40202, USA.
Introduction:
As orthodontic practice moves toward 3-dimensional cephalometric analyses, a solution is required to ensure sustained availability of well-established projected treatment outcomes based on 2-dimensional analyses. This ex-vivo study was conducted to compare the accuracy of linear measurements made on photostimulable phosphor cephalograms with 3 methods for simulating lateral cephalograms with cone-beam computed tomography (CBCT).
Methods:
The linear distances between anatomical landmarks on dentate dry human skulls were measured by observers using digital calipers for S-N, Ba-N, M-N, ANS-N, ANS-PNS, Pog-Go, Go-M, Po-Or, and Go-Co. The skulls were imaged with CBCT with a single 360 degrees rotation, producing 306 basis images and achieving 0.4 mm isotrophic voxel resolution on volumetric reconstruction for making ray-sum reconstructed cephalograms. Two other cephalogram approaches were used with the CBCT system--a single transmission image generated as a scout image designed to check patient positioning before CBCT, and a single-frame lateral basis image. Conventional digital lateral cephalograms (LCs) were acquired with the photostimulable phosphor system. Images were imported into a cephalometric analysis program (Dolphin Imaging Cephalometric and Tracing Software, Chatsworth, Calif) to compute the included linear measurements. Analyses were repeated 3 times and statistically compared with measured anatomic truth with ANOVA (P < or =.05). The intraclass correlation coefficient was determined as an index of intra- and interobserver reliability.
Results:
The intraclass correlation coefficient for the LCs was significantly less than for the measured anatomic truth and for all CBCT-derived images. CBCT images either produced with individual frames or reconstructed from the volumetric data set were accurate for all measurements except Pog-Go and Go-M. CBCT scout images had the second highest accuracy for all measurements except Pog-Go, Go-M, and Go-Co. Conventional LCs had the least accuracy; they were accurate only for Po-Or and ANS-N.
Conclusions:
CBCT-derived 2-dimensional LCs proved to be more accurate than LCs for most linear measurements calculated in the sagittal plane. No advantage was found over single-frame basis images in using ray-sum generated cephalograms from the CBCT volumetric data set.
