Cone-beam computed tomography imaging for detection and visualization of anatomical landmarks and pathological findings of the maxillary sinus : possibilities and limitations

Abstract

Cone-beam computed tomography (CBCT) is a three-dimensional imaging modality widely used in dentistry for research and clinical purposes. The maxillary sinus is a pyramidal-shaped air-cavity superior to the maxillary posterior dentition, and its health/pathology might be associated with the status of teeth in the posterior maxilla. This thesis comprises three parts. The first part aimed to assess the frequency, location, morphology, and dimensions of the accessory maxillary sinus ostia (AMOs) and mucous retention cysts (MRCs) in the maxillary sinus as well as their association with potential influencing factors (demographic data and dental pathologies) using CBCT. Results showed that an AMO/MRC was present in 47.2%/28.6% of the investigated sinuses. Most of AMOs (81.1%) were detected in the nasal fontanelle while MRCs were mainly located on the sinus walls/floor (45.5/45.0%). Sinus pathologies were associated with an increase in AMO size. Significantly more sinuses with associated endodontically treated teeth/periapical lesions were diagnosed with an MRC on the sinus floor. For MRCs on the sinus floor, endodontic status exhibited a significant association with increased MRC volume. The second part aimed to use CBCT data to (1) develop an intelligent algorithm based on convolutional neural network (CNN) for automated detection and segmentation of maxillary sinus pathologies, and to (2) develop bone-graft volume prediction models for sinus augmentation based on augmentation site, elevation height, and sinus width. The developed CNN achieved favorable detection (area under the curves of 0.84-0.93) and segmentation (dice-similarity-coefficients of 0.66-0.79) accuracies on both full-dose and low-dose CBCTs. The developed prediction models explained 89-91% of the variation in the estimated bone-graft volumes, and achieved small absolute percentage deviations between the measured and predicted volumes ranged between 9.78-10.62%. The third part aimed to (1) investigate the retake rate and radiation dose of CBCT examinations for various imaging indications as seen in an Oral-Maxillofacial Radiology unit; to (2) assess their association with potential patient-/imaging-related influencing factors; and to (3) establish institutional indication-oriented diagnostic reference levels (DRLs). Results showed that the retake rate was 4.6% in the investigated unit. The most common reasons for re-exposure were incomplete field-of-view (FOV) coverage (57.5%) and motion artifacts (27.5%). CBCTs of patients<12 years were more likely to be retaken due to motion artifacts while scans of temporomandibular joint were more likely to be retaken due to incomplete FOV coverage. The use of a low-dose protocol greatly affects the CBCT dose distribution, resulting in significantly lower dose in large FOV CBCTs than medium/small FOV scans. CBCTs for endodontic/periodontal/orthodontic/orthognathic evaluation exhibited a significantly higher dose than other indications. Twelve institutional indication-oriented DRLs were established. CBCT is favorable in assessing anatomical variations and pathologies of the maxillary sinus in relation to dentoalveolar pathologies as well as developing image-based intelligent algorithms and prediction models for diagnosis of sinus pathologies and planning of sinus-related surgeries. Information regarding the patient dose from CBCTs and reasons for re-exposure is beneficial to identify the insufficient aspects of the current imaging practice, staff training and education where optimization is required to minimize patient dose and avoid unnecessary radiation exposure.