However , this will only induce a 5% difference in the perfusion estimate, and therefore would not affect the significant differences measured between metastases and liver tissue. the between-session repeatability coefficient (RCBS) was 29%. Finally, the HASL method was applied to a mouse model of liver metastasis, in which significantly lower mean perfusion (1. 1 0. 5 mL/g/min, n= 6) was measured within the tumours, as seen by fluorescence histology. These data indicate that precise and accurate liver perfusion estimates SFN can be achieved using BEC HCl ASL techniques, and provide a platform intended for future studies investigating hepatic perfusion in mouse models of disease. Copyright 2014 Steve BEC HCl Wiley & Sons, Ltd. Keywords: liver, ASL, perfusion, mouse, repeatability, metastasis, variability, preclinical == Introduction == The application of arterial spin labelling (ASL) to the liver has not been undertaken extensively, possibly because of the liver’s complex blood supply and respiratory-induced artefacts, and continues to be restricted to a limited number of clinical investigations (12). Alternative methods, such as contrast-enhanced computed tomography, Doppler ultrasound and radioactive microspheres (3), have used perfusion to predict the onset of hepatocyte dysfunction (4), monitor tumour therapy BEC HCl (5) and inform on post-transplant success (6). Moreover, MRI assessment of liver disease pathology is well suited to the longitudinal evaluation of disease progression and therapeutic response in experimental models (7). A non-invasive and contrast agent-free method that could robustly measure liver perfusion would benefit researchers investigating a range of liver conditions, including cirrhosis (4) and either primary or metastatic tumours (8). Currently, dynamic contrast-enhanced (DCE)-MRI is the most common MRI measure of liver perfusion in clinical research, in which the pharmacokinetics of a chelated gadolinium contrast agent are modelled (9). The vasculature from the liver BEC HCl is unique, in that the portal vein delivers approximately 75% from the blood (10), whilst the remaining 25% is drawn from the hepatic artery. This dual supply means that the quantification of liver blood flow can be challenging, and requires careful consideration of acquisition and modelling methods. For dual input quantification, arterial and venous phases in the signal enhancement curve must be separated, which can be challenging in small animal models of disease, due to limited temporal resolution (1112), although recent clinical studies have adopted advanced purchase strategies to improve this (13). Moreover, the liver is significantly affected by respiratory motion, and the management of the interaction of respiratory gating with all the passage of a bolus of contrast agent during arterial and portal phases can be particularly challenging. However , some success continues to be reported in the measurement from the ratio from the arterial and venous contributions via the hepatic perfusion index, which has been shown to be informative for a number of liver diseases (14). Conversely, ASL-MRI is not reliant on external contrast agent administration, thereby offering a key advantage over DCE-MRI, and the approach taken in this study was to estimation the total (both portal and arterial) regional delivery to the liver. ASL-MRI has been developed preclinically and has been utilised to measure perfusion in the brain, heart and kidneys (1517). Despite the potential to fulfil the need for robust and reliable non-invasive liver perfusion measurement, ASL has not yet been reported in the liver of small animals, although a few human studies have been released (12), (1819). Here, we report the feasibility of a flow-sensitive alternating inversion recovery (FAIR) (20) sequence in the liver, which is a form of pulsed ASL that estimates perfusion from twoT1measurements following slice-selective and global inversion pulses, centred around the imaging slice. For image acquisition during inversion recovery, a segmented BEC HCl LookLocker sampling technique was implemented (21), because of the signal-to-noise ratio (SNR) efficiency from the sequence as well as previous demonstration for the measurement.