Volumetric and linear assessment of maxillary sinuses using computed tomography

The purpose of this work was to compare linear and volumetric measurements of maxillary sinus (with and without lesion), using 3D (gold standard) and multiplanar reconstructions (MRP) by computed tomography in order to establish the precision and accuracy of these methods, determining which one would allow more effective assessments for this site. Linear measurements were performed in the maxillary sinus of 50 patients and compared with MPR and 3D images. We also analyzed the area and volume of this anatomic structure. Inter- and intra-observer average percentage errors for linear measurements were lower than 4%, the samples with lesion presenting higher values. The average error of accuracy was 4.94% for sinuses with lesion, and 7.41% for those without. We could verify that the antero-posterior and right-left measurements in axial view showed smaller figures of error when compared to sagittal and coronal reconstructions, respectively. The largest diameters observed for the sinuses had, in average, 39.6 mm in the antero-posterior direction, 28.35 mm in the right-left direction, and 36.3 mm in the supero-inferior. For the volume and area assessment, the values of error ranged lower than 9%, the highest values corresponding to the samples with lesion. The precision of measurements was inferior for maxillary sinus with lesion when compared with those without lesion. Measurements performed in the axial view were more precise than those performed in MPR images, and the coronal view measurements were considered to be more precise than the sagittal measurements.