Successive measurement errors of consecutive computed tomography for airway-related craniofacial dimensional measurements

Background/purpose: The use of computed tomography (CT) for craniofacial measurements is common in medical imaging, but concerns about accuracy and reliability persist, especially with different CT technologies. This study assessed the accuracy of twenty-six common measurements on consecutive CT images from the same patients, using multidetector CT (MDCT) and cone-beam CT (CBCT) with two software programs (Amira and Dolphin). Materials and methods: Ten adult subjects with consecutive CBCT scans within one year were randomly selected. Another ten subjects with consecutive MDCT scans were paired with the CBCT group based on age, gender, race, occlusion, and craniofacial pattern. All digital imaging and communications in medicine (DICOM) files were randomly coded and analyzed using the two software programs. Intra-examiner reliability was assessed using the intraclass correlation coefficient. Successive measurement errors from consecutive scans for both imaging modalities and software programs were compared. Results: For most skeletal linear and angular measurements, Dolphin showed greater successive measurement errors compared to Amira. Eight of the 26 common measurements had errors greater than one unit (millimeter or degree). Despite almost perfect intra-examiner reliability for upper airway analysis, average successive measurement errors were notably high, particularly for intraoral and oropharyngeal airway volumes. The successive Dolphin measurement error for oropharyngeal airway volume on CBCT images was over three times that on MDCT images. Conclusion: Given the substantial successive measurement errors observed during consecutive CT scanning for the upper airway, this study does not support the quantitative use of CT for analyzing changes in airway dimensions for research purposes. (c) 2024 Association for Dental Sciences of the Republic of China. Publishing services by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).