Patient-specific variables like gender, date of birth, and date of CBCT were recorded. The date of the CBCT image was further divided into either being in autumn/winter (1 January 2013–19 March 2013; 22 September 2013–31 December 2013) or spring/summer (20 March 2013–21 September 2013). The maxillary sinus was classified into obliterated or nonobliterated. It was also documented if there was a unilateral or bilateral obliteration in the CBCT image. Obliterated cavities were further classified using the following radiographic findings: absence of alteration (0), mucosal thickening (1), sinus polyp (2), complete obliteration (3), mucosal thickening and periapical radiolucency (4), foreign body (5), mucosal thickening and foreign body (6), and nonspecific obliteration (7, partial obliteration, not being defined by the previous criteria). Due to the close relationship between the posterior teeth (premolars, molars) and the maxillary sinus, the teeth starting at the first premolar were recorded as either present or missing, along with the presence or absence of endodontic treatment. Additionally, the number of teeth and roots communicating with the maxillary sinus and any apical radiolucency was documented.

The data was primarily analyzed descriptively. The analysis was performed on two different datasets depending on the main question: either on a sinus level consisting of 128 maxillary sinuses or on a patient level consisting of the respective 64 patients. In cases where sinus-level information was associated with patient-level characteristics (presence of pathology vs. obliterated volume, presence of apical radiolucency vs. obliterated volume, presence of pathology vs. number of communicating roots, dentition status vs. osseus sinus volume), one sinus per patient was randomly chosen for the analysis in order to not violate assumptions of independency for the Wilcoxon rank sum and Kruskal-Wallis tests.

For patient-level analysis, the association between a patient’s age and the presence of obliteration was analyzed using logistic regression and patient’s age vs. the mean osseus sinus volume was assessed using linear regression. The Wilcoxon rank sum test was used to investigate if there is an association between the mean osseus sinus volume and gender. Differences between osseus sinus volumes on the left and right side of a patient were assessed with the Wilcoxon signed-rank test. Fisher’s exact test was used to assess possible associations between bilateral obliteration and the date of the CBCT scan (season of the year) as well as between unilateral obliteration and apical radiolucency. The significance level α was set to 0.05 for all analyses. Calculations were performed using R [3].