Skip to main content
. 2017 Dec;7(6):668–677. doi: 10.21037/qims.2017.11.02
Critical structure CT dataset utilised Approximate time required for segmentation Comments
Liver body Axial, portal venous phase image data 4 hours Manual segmentation
Error in manual segmentation was minimised by (I) selecting the portal venous phase dataset featuring normal enhancement of liver parenchyma; and (II) applying a median filter to better define liver boundaries
Tumour and associated arterial vascular supply Axial, arterial phase image data 1 hour Semi-automatic segmentation involving (I) rough manual segmentation to distinguish the region including the tumour and hepatic arteries; followed by (II) application of a suitable threshold to separate the enhancing structures from non-enhancing intrahepatic tissues
Portal and hepatic veins (and associated venous branches) Axial, segmented liver body originating from the portal venous image data 1 hour Combination of thresholding and object separator tools utilised
Success of thresholding was dependent on the quality of contrast enhancement
The segmented liver volume data was used as unwanted abdominal structures of similar attenuation were excluded, allowing for the thresholding to be more effective

CT, computed tomography.