Although regarded as independent processes, recent evidences have suggested that there is a link between diabetes mellitus (DM) and Alzheimer’s disease in terms of underlying pathomechanism (1). Quantitative susceptibility mapping (QSM) can serve as a tool to measure brain iron accumulation in vivo (2, 3). We hypothesized that DM would modulate the iron accumulation in cognitively impaired patients. In this study, we aimed to determine the relationship between the presence of DM and brain iron in cognitive impaired patients, by using QSM.This retrospective study enrolled 66 subjects with cognitive impairments (Mean age, 77.1 ± 8.9 years; 56 women and 10 men; 11 patients with mild-cognitive impairments and 55 Alzheimer's disease; 20 patients with DM and 46 without DM). MR imaging was performed at 3T (GE Signal HDxT). QSM was obtained using gradient echo sequences (SWAN) with following parameters: TR/TE, 37/3.5ms (multi-echo, 8); FA=20º, slice thickness, 2.5mm, matrix 256x256; FOV, 240mm. QSM images were coregistered to 3D-T1 weighted FSPGR images by using SPM. Each region of interest (ROI) for anatomical structures (caudate head, putamen, globus pallidus, thalamus, amygdala and hippocampus) was drawn semi-automatically using MIPAV software application (http://mipav.cit.nih.gov/). White matter hyperintensity (WMH), lacunes and microbleeds were graded using FLAIR and susceptibility-weighted imaging as a measure of vascular risk factor. Gray matter and white matter volumes, which were normalized for total intracranial volume, were measured by using VBM8 toolbox. The difference of measurements and clinical variables between DM (-) and DM (+) patients, chi-square test and standard t-test were used. Analysis of covariance was performed to determine the biologic factors associated with susceptibility changes.Hypertension and hyperlipidemia was more frequently found in DM (+) patients (p = 0.001 & p = 0.022). Vascular risk factors on imaging and normalized brain volumes were not different according to the presence of DM. There were no significant difference in measured susceptibility of all anatomical regions between DM (-) and DM (+) groups except for left thalamus, particularly in pulvinar nucleus (45.02 ± 19.41 versus 33.53 ± 20.61 [ppb], p=0.043). There was a significant effect of DM on the left pulvinar susceptibility (iron accumulation) after controlling for age, hypertension and hyperlipidemia, F (1,61) = 4.148, p = 0.046.We found that the presence of DM is not associated with vacular risk factors on imaging and brain volumes, but associated with decreased susceptibility in left thalamus. Our results indicate that there may be a region-specific alteration of iron accumulation and iron metabolism in patients with both AD and DM.