Evaluation of 3D imaging software to study computed tomography images of middle ear defects

Evaluate the contribution of original 3D reconstruction software based on the principle of region growing segmentation of CT images to diagnosis of ossicular malformations of the middle ear. Prospective study on a population of ten patients presenting uni- or bilateral transmission deafness. Twenty ears were studied using multislice CT. All the data were treated using original software segmentation based on the principle of region growing segmentation. The 3D images were compared with the original CT slices to judge the software's potential contribution to the management of these patients. All the images obtained were deemed of sufficient diagnostic quality. The two normal ears on the CT slices were also found to be normal on the 3D images. Typical CT images of otospongiosis found in two patients were not visible on the 3D images. The 18 ossicular malformations were well analyzed on the images obtained, which also provided better visualization of the abnormalities and associated lesions (particularly the anomalies in facial nerve position), thus providing quality surgical planning as judged by the surgeon. The image reconstruction method studied, using region growing segmentation seems to be, in this limited sample, a useful complement to classical CT images in the study of constitutional abnormalities of the middle ear.