To measure magnetic susceptibility in cortical lesions of multiple sclerosis (MS) and compare it to that in white matter lesions using quantitative susceptibility mapping (QSM).Magnetic susceptibility measured with QSM has been proposed as a biomarker for demyelination and inflammation in MS white matter (WM) lesions¹, but a detailed characterization of cortical lesions has yet to be performed. Since normal cortex and cortical lesions have lower degree of myelin and inflammation than normal WM and WM lesions², respectively, we hypothesize that the susceptibility values relative to surrounding parenchyma for cortical lesions may be different from that for WM lesions.Patients: 14 MS patients (13 had relapsing-remitting MS, and 1 had secondary progressive MS; age 40.4 ± 7.9 years, 7.9 ± 7.2 years of disease duration) were recruited for an MRI study. MR Imaging: T1-weighted white matter or CSF nulled magnetization prepared rapid acquisition imaging (WMnMPRGE or CSFnMPRAGE), T2-FLAIR and T2*-weighted imaging were performed on a GE 7T scanner. The T2*-weighted imaging (2D SPGR sequence; TE/TR=17.7/1200ms; FA=60°; FOV=180x180x90mm³; Resolution= 0.47x0.47x1mm3; BW=19.2KHz) was used to compute QSM images using the morphology-enabled dipole inversion (MEDI) method³. Lesion Identification and Segmentation: All anatomical images were used concordantly to identify lesions. Blinded to QSM images, we manually segmented lesions with size ≥ 2mm by drawing regions of interests (ROIs) covering hyperintense voxels on T2*-weighed magnitude images. WM and cortical lesion ROIs were drawn only within WM and cortex, respectively (Figure 1). Reference ROIs were also drawn on adjacent normal-appearing WM (NAWM) and normal-appearing gray matter (NAGM) for WM and cortical lesions, respectively (Figure 1). Blood vessels within each ROI were excluded. Prior 3.0T clinical MRIs were reviewed to estimate the age of white matter lesions (cortical lesions could not be accurately assessed on these clinical scans). Data Analysis: To calculate relative susceptibility value in a lesion, the mean susceptibility value in the lesion ROI was subtracted by the same value in its adjacent normal-appearing ROIs. The relative susceptibility values for the set of cortical lesions and for the set of WM lesions were compared to zero using the one-sample t-test. This was performed to assess the per-lesion-type difference between WM and cortical lesions. The relative susceptibility values for all cortical lesions and for all WM lesions were also averaged in the same subject, respectively, and then the above t-test was repeated to assess the per-lesion-type per-subject difference. The statistical significance threshold was set as p<0.05.A total of 176 lesions were identified, of which 28 (15.9%) were cortical while 148 (84.1%) were WM. Prior clinical 3T MR scans indicated that all WM lesions were older than 3 months. The relative susceptibility value was positive for 133 of the 148 (89.9%) WM lesions, but negative for 25 of the 28 (89.3%) cortical lesions (Figure 2). The relative susceptibility value was significantly higher than zero for WM lesions ( 0.014±0.014ppm, p<10^-23), but significantly lower than zero for cortical lesions ( ‑0.018±0.013ppm, p<10^-7). The patient-averaged relative susceptibility value was again significantly higher than zero ( 0.013±0.010ppm, p<0.0006) for WM lesions but significantly lower than zero (‑0.017±0.012ppm, p<0.006) for cortical lesions. The QSM images illustrate these quantitative difference, with WM lesions being hyperintense and cortical lesions being hypointense relative to their adjacent normal appearing parenchyma (Figure 3).Demyelination (loss of diamagnetic myelin) and accumulation of paramagnetic iron has been interpreted as major contributors to the increased susceptibility in MS lesions¹. However, iron content in chronic inactive WM lesions could be lower than that in NAWM⁴. Since our measured susceptibility in the majority of chronic WM lesions was still positive relative to NAWM, the effect of demyelination may dominate over that of iron loss on the susceptibility changes in chronic WM lesions. On the contrary, the negative relative susceptibility values of cortical lesions, which has been seen in a postmortem QSM study⁵, suggests that iron loss dominates over demyelination in the susceptibility changes of cortical lesions. Substantial iron loss in cortical lesions has been reported in previous histological studies⁶, but the pathology for the iron loss remains unclear. One of possible explanations could be low degree of inflammation in cortical lesions, resulting in a small amount of infiltration of macrophages/microglia, which are the holders of iron released from damaged oligodendrocytes in MS lesions.In summary, high field QSM reveals a negative relative magnetic susceptibility in cortical lesions and a positive one in WM lesions, suggesting that iron loss dominates the susceptibility contrast in cortical lesions. The susceptibility difference between WM and cortical lesions may have implications for MS progression.