An ex vivo study of automated motion artefact correction and the impact on cone beam CT image quality and interpretability.

OBJECTIVES To assess the impact of head motion artefacts and an automated artefact-correction system on cone beam CT (CBCT) image quality and interpretability for simulated diagnostic tasks. METHODS A partially dentate human skull was mounted on a robot simulating four types of head movement (anteroposterior translation, nodding, lateral rotation, and tremor), at three distances (0.75, 1.5, and 3 mm) based on two movement patterns (skull returning/not returning to the initial position). Two diagnostic tasks were simulated: dental implant planning and detection of a periapical lesion. Three CBCT units were used to examine the skull during the movements and no-motion (control): Cranex 3Dx (CRA), Orthophos SL 3D (ORT), and X1 without (X1wo) and with (X1wi) an automated motion artefact-correction system. For each diagnostic task, 88 examinations were performed. Three observers, blinded to unit and movement, scored image quality: presence of stripe artefacts (present/absent), overall unsharpness (present/absent), and image interpretability (interpretable/non-interpretable). κ statistics assessed interobserver agreement, and descriptive statistics summarized the findings. RESULTS Interobserver agreement for image interpretability was good (average κ = 0.68). Regarding dental implant planning, X1wi images were interpretable by all observers, while for the other units mainly the cases with tremor were non-interpretable. Regarding detection of a periapical lesion, besides tremor, most of the 3 mm movements based on the "not returning" pattern were also non-interpretable for CRA, ORT, and X1wo. For X1wi, two observers scored 1.5 mm tremor and one observer scored 3 mm tremor as non-interpretable. CONCLUSIONS The automated motion artefact-correction system significantly enhanced CBCT image quality and interpretability.