A simplified CT‐volumetry method for the canine liver

Abstract

Computed tomographic (CT) liver volumetry using the slice addition technique is an accurate, but a time‐consuming method. Commonly used DICOM‐viewing software only allows contouring of one area per image, which can be troublesome in the 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. Computed tomographic 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 was 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 (intraclass correlation coefficient = 0.957; 95% confidence interval, 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.

Computed tomographic features of the normal spleen in rabbits (Oryctolagus cuniculus domesticus)

“…These findings supported those reported in another study on canine liver CT volumetry, using the same, widely available software. 19 Limitations of the current study include the lack of histopathological confirmation of splenic normalcy and the inability to record splenic weight, due to the retrospective nature of the study and the study population of live patients. Some of these limitations were partially mitigated by performing the pilot studies.…”

Pilot Study: The Effects of Slice Parameters and the Interobserver Measurement Variability in Computed Tomographic Hepatic Volumetry in Dogs without Hepatic Disease

“…In the previous publication evaluating interobserver variability of pulmonary nodule diameter measurement, a CV of 16% was considered a low interobserver variability [ 21 ]. A previous manual CT hepatic volumetry publication used an intraclass correlation coefficient to evaluate interobserver variability, and an intraclass correlation coefficient of 0.957 with 95% confidence interval of 0.908–0.982 was considered excellent [ 22 ]. In the present study, CV of CT hepatic volumetry among three observers was 2.5 (±1.4)%.…”

“…There has been a publication on the effect of slice numbers used for CT hepatic volumetry in dogs using sagittal reformatted images [ 22 ]. This study suggested using more than nine slices of sagittal reformatted images to accurately perform CT hepatic volumetry [ 22 ].…”

“…There has been a publication on the effect of slice numbers used for CT hepatic volumetry in dogs using sagittal reformatted images [ 22 ]. This study suggested using more than nine slices of sagittal reformatted images to accurately perform CT hepatic volumetry [ 22 ]. However, in the previous study, the calculated hepatic volumes using a different number of slices (9 to 58 through 65 slices depending on the size of the dogs) showed wide variations in each dog with at least more than 10% volume difference from the mean calculated hepatic volume in all five dogs, and the differences of the largest volume and the smallest volume using different slice numbers in each dog were 11.1% to 37.9% [ 22 ].…”

“…This study suggested using more than nine slices of sagittal reformatted images to accurately perform CT hepatic volumetry [ 22 ]. However, in the previous study, the calculated hepatic volumes using a different number of slices (9 to 58 through 65 slices depending on the size of the dogs) showed wide variations in each dog with at least more than 10% volume difference from the mean calculated hepatic volume in all five dogs, and the differences of the largest volume and the smallest volume using different slice numbers in each dog were 11.1% to 37.9% [ 22 ]. In the present study, we evaluated the greatest difference in calculated hepatic volume compared to the full study and used 5% as a cut-off to reduce the error caused by using a smaller number of slices.…”