A simplified and accurate CT-volumetry method for the canine liver

Computed tomographic (CT) liver volumetry using the slice addition technique is an accurate, but time-consuming method. Commonly used DICOM-viewing software only allow contouring of one area per image, which can be troublesome in transverse plane as different lobes are separated. In this prospective, experimental, methods comparison study, we aimed to determine if hepatic contouring using sagittal reformatting and a reduced number of images would yield accurate results. CT studies were performed in five canine cadavers and reviewed using sagittal reformatting. For each dog, the number of images that included the liver was used to create four stacks with progressively fewer images in which the liver would be contoured, each with the following median number of images: A: 60, B: 31, C: 16, and D: 9. Liver volume was calculated by three observers using the different stacks of images. After CT examination, the cadavers were dissected, the liver was removed, and its volume determined by water displacement. Single score intraclass correlation coefficient was calculated to assess interobserver agreement. Kruskal-Wallis test was used to compare water displacement and CT- based volumes. There was excellent agreement between observers (ICC= 0.957; 95% CI= 42 0.908-0.982, p<0.0001). No significant difference was found between the volumes obtained by CT-volumetry using each of the stacks and the volumes obtained by water displacement. Using sagittally reformatted images and hepatic contouring in as few as nine images can be an accurate and simple method for CT-volumetry of the canine liver.