Thoracic duct (TD) ligation has long been the treatment of choice for canine idiopathic chylothorax. Clear identification of all the TD branches at the surgical site is critical to facilitate complete ligation, and this can be difficult due to the highly variable nature of the TD system in number, location and patency of TD branches. Failure to ligate all of the TD branches may result in persistent flow of chyle into the pleural cavity through any missed ducts, and this is the single most common cause of failure with TD ligation. Performing direct positive contrast lymphography with a water soluble contrast medium, administered through a surgically implanted mesenteric lymphatic vessel catheter has been the conventional method used to identify TD branches. This procedure involves invasive surgery to both implant and remove the mesenteric lymphatic catheter, which increases patient risk and discomfort, as well as the diagnostic time and cost. Ultrasound (US) guided percutaneous administration of contrast medium into either a popliteal or mesenteric lymph node (LN) have been proposed as alternatives to mesenteric lymphatic vessel catheterisation, however their comparability with the conventional approach has not been assessed. Computed tomographic (CT) lymphography of the caudal canine TD was performed in seven beagles with contrast medium administered through a mesenteric lymphatic catheter, and by US guided percutaneous injection into a popliteal LN. Images of the TD system were collected using both helical and sequential CT modalities for each contrast medium administration technique. It was found that percutaneous popliteal lymphography had a total diagnostic procedure time just 46% of that found with mesenteric lymphatic vessel administration, and resulted in a time saving of 52 minutes. It also incurred only 29% of the total costs, and patients were assessed to have significantly less discomfort compared to mesenteric lymphatic vessel lymphography. There was no significant difference in the number of TD branches identified by the two contrast medium administration techniques (P = 0.256). However administration of contrast medium into a mesenteric lymphatic vessel did result in the largest TD branch having a significantly greater widest diameter (P < 0.001), cross-sectional area (P < 0.001) and mean Hounsfield unit (HU) value (P < 0.001) than popliteal administration. The significant difference in TD size and contrast medium concentration may help to explain the trend for popliteal administration of contrast medium to detect slightly fewer TD branches (CR = 0.830), however this study could not confirm this trend nor its possible causes. There was no significant difference in the number of TD branches identified by the two CT modalities (P = 0.417). However helical CT did result in the largest TD branch having a significantly greater widest diameter (P < 0.001), cross-sectional area (P < 0.001) and higher mean HU value (P < 0.001). It should be noted however that in this study sequential CT was consistently performed after the helical CT was completed, which could explain the differences seen between the two CT modalities in TD branch size and contrast medium concentration. Despite helical CT having the apparent advantage of a larger TD branch which contained a higher concentration of contrast medium, it was actually found that there was a trend for helical CT to detect slightly fewer TD branches (CR = 0.876). This is possibly due to the positive pressure breath hold that was used to minimise thoracic respiratory movement for the helical CT; while the sequential CT was performed during normal respiration and was therefore not subject to abnormally elevated intrapleural pressure. This study could not confirm the slight superiority of sequential CT in detecting TD branches nor the possible reasons for this apparent difference. / Dissertation (MMedVet)--University of Pretoria, 2010. / Companion Animal Clinical Studies / unrestricted