Material and Methods
Study was conducted in the Department of Dentomaxillofacial radiology ofXXX University, Faculty of XXX , between January-November 2019, with ethical approval number: 08/03 given by the faculty Clinical Research Ethics committee and with ClinicalTrials.gov ID: NCTXXX . The study has been carried out in accordance with The Code of Ethics of the World Medical Association. Computed tomography scans of 120 patients (59 maxillary deficiency patients and 61 control) were included in the study. CT scan indications for the group with maxillary deficiency was surgical preparation for maxillary advancement with Le Fort I osteotomy. The control group had no maxillary deficiency and CT images were obtained for several other indications such as temporomandibular disease. CT scans of patients with a history of trauma in the midline skull base, and with the presence of sinus diseases such as chronic sinusitis, nasal polyps, and tumors were excluded from the study.
Imaging procedures: CT imaging’s were performed using various CT scanners (GE Lightspeed 16 slice, GE Medical Systems, Milwaukee WI, US; Siemens Somaton Sensation, 16 slice, Siemens Medical Solutions, Erlangen, Germany) at 120 kVp and 25 mA with a display matrix of 256X256. Images were processed and measured using 3D Synapse Software (Fujifilm, Tokyo, Japan). The CT data were transferred as DICOM files to the software for measurements. The volumes (mm3) were measured as semiautomatic segmentation of the right and left maxillary sinuses, and the ethmoid and sphenoid sinus. For segmentations, the thresholding was limited to a minimum of 1024 HU and a maximum of 526 HU. 6 The ethmoid and sphenoid sinuses were sculpted out from these 3D images. Semiautomatic segmentation was performed to include the entire paranasal sinus anatomies (Figure 1, 2 and 3).
For maxillary sinuses, those were clipped according to the surrounding bone structure and the narrowest area of the ostium between the infundibulum and the processus uncinatus using the software’s “edit masks” tool. Then, the connection with the outside air was interrupted by slice using segmentation tools. The “region growing” tool was used to divide the segmentation created by thresholding several objects and remove the floating pixels. The paranasal sinuses then were automatically calculated using the software analysis tool. Presence of Onodi cells were noted.
All measurements were taken twice by two blinded radiologists. The radiologists performed the study twice with an interval of 2 weeks to detect intra-observer variability.
Statistical analysis: The analysis of the data was done in SPSS version 11,5. As descriptive, mean ± standard deviation and median (minimum - maximum) were given for quantitative variables, and number (percentage) was given for qualitative variables. To assess intra-observer reliability, the Wilcoxon matched pairs signed rank test was used for repeat measurements.
In terms of the quantitative variables, the difference between the categories of the qualitative variable with two categories was examined using Student-t test if normal distribution assumptions were provided, and Mann Whitney U test if not. When the difference between two dependent quantitative variables was to be examined, the Wilcoxon Signed Rank test was used, since normal distribution assumptions were not provided. Chi-square test was used to examine the relationship between two qualitative variables. Statistical significance level was taken as 0,05.