Cerebral venous oxygen saturation and the cerebral metabolic rate of oxygen (CMRO₂) are very important parameters for the evaluation of brain function and tissue viability. Haemodialysis (HD) is a way to eliminate uremic toxins and excessive water from the blood. However, there are some central neurologic complications in HD patients because of the inability to remove all uremic toxins, including cerebrovascular disease, cognitive dysfunction and cerebral dysfunction. In previous studies, researchers used near-infrared spectroscopy (NIRS) to examine cerebral oxygen saturation changes in HD patients^{1, 2}. However, the spatial resolution of NIRS is quite low. Recently, susceptibility mapping (SWIM) for the evaluation of cerebral venous oxygen saturation was introduced³. In this study, we aim to observe the cerebral venous oxygen saturation in patients with long-term HD using SWIM and investigate the changes of the cerebrovascular metabolism.86 HD patients and 64 age-and gender-matched healthy controls participated in this study. All the patients and healthy controls were right-handed and had normal venous shape and distribution. MRI data were collected on a MAGNETOM Trio Tim 3T MR scanner (Siemens Healthcare, Erlangen, Germany) including routine head scan and susceptibility-weighted imaging (SWI). The parameters of SWI were: TR/TE=27/20ms; field of view=230×200mm²; receiver bandwidth=120 Hz/pixel; flip angle=15°; number of slices=56; acquisition time=179 sec; voxel resolution=0.5×0.5×2mm³. SWIM was reconstructed using both the phase and magnitude data using SMART (Susceptibility Mapping and Phase Artifacts Removal Toolbox, Detroit, Michigan, USA) software. The susceptibility of bilateral cerebral veins (cortical veins, thalamostriate veins, septum pellucidum veins, internal veins, basal veins) and straight sinus were measured using SPIN (signal processing in nuclear magnetic resonance, Detroit, Michigan, USA) software (Fig. 1). A threshold of 90 ppb was used to eliminate the susceptibility of background tissue and only preserve the susceptibility of cerebral veins. The global venous oxygen saturation and CMRO₂ were calculated using the susceptibility of straight sinus in the healthy controls according to the equations [∆χ_vein-tissue=κ∆χ_do.Hct (1-Y) and CMRO₂=C_aCBF(S_aO₂-S_vO₂)] as reported in previous literatures³. Two independent sample t-test was used to explore the cerebral venous oxygen saturation changes of HD patients compared to those of healthy controls. Stepwise regression analysis was used to explore the correlation between cerebral venous oxygen saturation of patients with HD and clinical parameters. Spearman correlation analysis was used to explore the correlation between cerebral venous oxygen saturation and mini mental state examination (MMSE). The global venous oxygen saturation and CMRO₂ in healthy people were 60.0±5.5% and 153.8±21.7 umol/100g/min, which is consistent with the values reported by PET and MRI studies. The susceptibility of straight sinus, bilateral cortical veins, bilateral thalamostriate veins, bilateral septum pellucidum veins and left basal vein in HD patients was lower than that in normal groups (p<0.05) (Table 1). The susceptibility of cerebral veins positively correlated with these clinical parameters including age, gender, red blood cell (RBC), hemoglobin, triglyceride, serum potassium, and low-density lipoprotein cholesterol. On the other hand, it also negatively correlated with glucose, post-dialysis systolic blood pressure and pulse pressure, and serum sodium (p<0.05) (Table 2). The susceptibility of the left cerebral cortical vein showed a weakly positive correlation with MMSE scores (r=0.281, P<0.05) (Fig. 2).Compared to the healthy controls, the global CMRO₂ in HD patients was decreased, and venous oxygen saturation of straight sinus, bilateral cortical veins, bilateral thalamostriate veins, bilateral septum pellucidum veins and left basal vein was increased. The reason may be that haemodialysis resulted in cerebrovascular impairment, which further caused low CMRO₂ and high venous oxygen saturation. The cerebral venous oxygen saturation was affected by some clinical parameters including age, gender, RBC, hemoglobin, triglyceride, serum potassium, low-density lipoprotein cholesterol, glucose, post-dialysis systolic blood pressure and pulse pressure, and serum sodium. The increased oxygen saturation of the left cerebral cortical vein mildly correlated with neurocognitive impairment. The reason may be that the increased venous oxygen saturation indicated that the decreased CMRO₂ and the decreased CMRO₂ caused the brain function disorder.Our study suggests that SWIM is a feasible and reliable method to measure the venous oxygen saturation and can show the decrease of CMRO₂ in HD patients.