Animals
All applicable institutional and/or national guidelines for the care and use of animals were followed. Forty-six New Zealand white female rabbits weighing around 3 kg (range 2.68 − 3.55 kg) were included. This is an investigation in the context of an already established experimental study planned for other purposes. All rabbits first underwent basal CECT followed by direct portography.
CECT acquisition protocol
All examinations were performed with a 64-MDCT scanner (Somatom Sensation, Siemens Medical Systems, Erlangen, Germany). Examination parameters were 64 × 0.6 mm collimation, 1.4 mm/s table feed, 2 mm section thickness, 1.5 mm reconstruction interval, 80 kV, 65 mA and 0.5 s rotation time. Rabbits were placed in the supine position and were sedated by intramuscular injection of 100 mg/ml ketamine (Imalgene®, Merial) and 1 mg/ml of medetomidine (Domtor®, Esteve veterinaria). After unenhanced scan acquisition, a contrast-enhanced scan was performed 15 s (arterial phase), 30 s (portal phase), 45 s (venous phase), and 60 s (late venous phase) after intravenous injection of 4 ml of a non-ionic contrast agent (iohexol, 300 mg/ml; Omnipaque, Amersham, Cork, Ireland), followed by 3 ml of saline solution. Resulting axial computed tomography (CT) images were transferred to an external workstation (Leonardo, Siemens Healthcare).
Direct portography protocol
For anaesthesia, each rabbit was given an intramuscular injection of 10 mg/kg ketamine, 0.15 mg/kg of medetomidine, and 2–8 mg/kg of intravenous propofol (Propofol Lipuro®, Braun, Melsungen, Germany). The animals were placed in the supine position and a laryngeal mask size-1 was used. An 8–10 cm midline incision through the skin was made from the epigastrium for a laparotomy. Then an incision was made through the thin subcutaneous tissue to expose the linea alba. Using an inverted number-11 scalpel blade, an incision was made parallel to the linea alba. After exposure of the small bowel, a small branch of the superior mesenteric vein (SMV) was punctured with a 24-gauge needle (Abbocath, Abbott Laboratories, Chicago, IL, USA). After the stylet of the needle was removed, a 0.014-inch guidewire (Transend, Boston Scientific MediTech, Natick, MA, USA) was advanced with fluoroscopic guidance towards the liver. After removal of the cannula, a 4-F coaxial catheter introducer (Micropuncture access set, Cook Medical, USA) was advanced within the SMV. The 2-F introducer was then removed and direct portography was finally performed. Portograms were obtained with injection of 10 ml of contrast (RadialarⓇ 280 mg/ml, Juste SAQF, Madrid, Spain) into the hand. When the procedure was finished and the introducer was removed, the SMV was ligated at the level of the catheterisation. Finally the puncture site was covered with an absorbable haemostatic agent (Surgicel®, Ethicon, Somerville, NJ, USA). After confirming that there was no bleeding, the linea alba was closed in a simple continuous suture pattern with a synthetic absorbable material. The skin was closed using an interrupted pattern. Resulting images were transferred to the ImageJ programme (Rasband WS, ImageJ; National Institute of Health, Bethesda, MD, USA) where the diameter and length of the portal branches were obtained.
Image interpretation and data collection
All CECT examinations were interpreted by two radiologists with more than 10 years of experience in hepatic CT, who selected the acquisition phase in which the original portal vein (OPV) was better demonstrated. The OPV diameter at the level immediately before the origin of the CPV was measured three times in each rabbit by each of the two radiologists. The mean of the six measurements on CECT served to calibrate and set the measurement scale for each animal. The same radiologists measured the diameter and length of the different branches of all the portograms.
Conventional liver rabbit portal anatomy, which was categorised as type 1, consisted of the bifurcation of the MPV into the RPV and LPV, which subsequently divided into MLPV and LLPV (Fig. 1). Any deviation from these branching variations was regarded as variant anatomy. Trifurcation of the LPV was considered type 2 anatomic variation (Fig. 2). The LPV that divides into four smaller branches was classified as type 3 (Fig. 3). Based on the origin of the LIPV, each classification type was divided in three categories: (1) if the origin was from the LPV; (2) if the origin was from the MPV; or (3) if the LIPV was absent. Other different configurations of the portal system were grouped as type 4 (Fig. 4). Because all major (lobar) portal trunks have multiple small branches (smaller than 2 mm), the length of the main trunk was obtained until there was a major bifurcation (branch bigger than 2 mm). The diameters and lengths of the following portal branches were measured: CPV, MPV, RPV, LPV, MLPV and LLPV. The angle of the OPV bifurcation (CPV and MPV) was also measured on each portogram.
The caudate liver lobe and the cranial lobes were manually delineated and their respective volumes were calculated (Fig. 5).