Paranasal sinuses of 52 patients with sinusitis, tumours or fibrous dysplasia were scanned. Axially acquired spiral data were obtained and reconstructed coronally. Four radiologists compared the two sets of images. They evaluated the diagnostic quality for visualization of the ostiomeatal unit, infundibulum, infraorbital canal, inflammatory disease, fine osseous lamellae and presence of amalgam or step artefacts. Two test phantoms were scanned for both techniques and lens dose was measured. Statistical significant differences in the diagnostic quality of the representation of the fine osseous structures in the paranasal sinuses, attributable to step artefacts were found in the coronally reconstructed images (p < 0.001). However, there was no amalgam and almost no motion artefacts in the reconstructed images. Interobserver correlation was r = 0.953 versus 0.956 for inflammatory disease, r = 0.816 versus 0.852 for artefacts, and r = 0.596 versus 0.547 for fine osseous lamellae in coronally acquired or axially acquired and reconstructed images, respectively. Lens dose was measured between 11.8 mGy and 13.8 mGy for axially acquired and reconstructed images. The advantage of axially acquired, coronally reconstructed images is the absence of artefacts attributable to amalgam and fewer motion artefacts. Axially acquired, coronally reconstructed images are inferior to coronal helical CT images, because of step artefacts, when it comes to evaluating the resolution of fine osseous structures. Nevertheless, reconstructed images are suitable as a investigatory procedure for patients with inflammatory disease who cannot maintain the prone position.