Hepatic warm ischemia-reperfusion injury (WIRI) is considered to be the most common cause of postoperative liver dysfunction, graft rejection, and chronic liver diseases1. Multiparametric functional MR imaging has been successfully applied in the liver and shows high potential in detecting microscopic changes of liver in many diseases2. The purpose of this study was to assess the pathophysiological changes of hepatic WIRI in rabbit models using intravoxel incoherent motion (IVIM), diffusion tensor imaging (DTI) and blood oxygen level dependent (BOLD) MR imaging.The protocols and procedures were approved by the local Animal Experimentation Ethics Committee. Hepatic WIRI was induced in rabbit by occluding hepatic inflow for 30 minutes and reperfusion for 6 hours. 10 hepatic WIRI models and 10 control rabbits were examined using a 3T clinical MR scanner (MAGNETOM Trio, a Tim system, Siemens Healthcare, Erlangen, Germany). Transversal IVIM imaging was acquired using a single-shot echo-planar imaging (ss-EPI) prototype sequence with TE/TR of 57.2/1000 ms, and 11 b values of 0, 20, 40, 60, 80, 100, 150, 200, 400, 600, and 800 s/mm2 on 3 gradient directions. To separate the perfusion and diffusion, a bi-exponential fit was used to calculate the perfusion fraction (PF), Dfast and Dslow3. DTI was performed using the ss-EPI sequence with TE/TR of 86/3800 ms, and 2 b values of 0 and 500 s/mm2 on 12 diffusion directions. BOLD-MR was acquired using a multi-echo gradient echo sequence with TR of 75ms, 9 TE of 2.57, 5.23, 7.52, 9.81, 12.1, 14.39, 16.68, 18.97, and 24.25 ms. Rabbits were sacrificed after MR imaging for biochemical analysis, including alanine transaminase (ALT), aspartate transaminase (AST), lactate dehydrogenase (LDH), malondialdehyde (MDA), myeloperoxidase (MPO) and superoxide dismutase (SOD), and histomorphology. Functional MRI parameters, including IVIM-derived Dfast, Dslow and PF, DTI-derived ADC and fractional anisotropy (FA), BOLD-derived R2*, were compared between the two groups, and the correlations with biochemical parameters were assessed using Spearman correlation test.Fig.1 shows image examples of Dfast, Dslow, PF, ADC, FA and R2* maps in a control rabbit and a hepatic WIRI rabbit model. The statistical analysis is shown in Fig. 2. For IVIM, Dfast, Dslow, and PF were lower in WIRI models than in control rabbits (P <0.01). For DTI, ADC was lower in WIRI models than in control rabbits (P <0.001), but FA showed no statistical difference. For BOLD, R2* was higher in WIRI models than in control rabbits (P <0.001). For biochemical parameters, ALT, AST, LDH, MDA, and MPO were reduced, but SOD was increased in WIRI models (P <0.01). Multiparametric MR findings showed significant correlation with the biochemical parameters (P <0.05 for all). Histopathological analysis showed diffuse hepatocyte swelling, central vein and sinusoids congestion, and acute and chronic inflammatory cell infiltration in WIRI models (Fig. 3).Hepatic WIRI cause sever ischemia and anoxia, oxidation/antioxidation imbalance, and excessive inflammatory response4, which result in liver microcirculatory disorders and histological damage. This study demonstrated that the hepatic diffusivity (including pure molecular diffusion and perfusion-related diffusion), blood perfusion, and oxygenation state decreased obviously in hepatic WIRI. These results could be supported by biochemical parameters and histopathological changes.IVIM, DTI, and BOLD MR imaging are noninvasive and valuable techniques for assessing the pathophysiologic changes of hepatic WIRI in rabbit models, which implies its further application for early detection and dynamic monitoring of hepatic WIRI in clinical application.