Electric Properties Tomography (EPT) is a recently developed technique which estimates electrical conductivity (σ) of tissues noninvasively and in vivo, from the phase images of an MRI scan¹. Preliminary studies have suggested its potential usefulness in grading glioma and differentiation between meningioma and lymphoma^2,3. Nevertheless, the former was performed on a small number of glioma patients and included patients who were already undergoing treatment². Regarding validity of noninvasive σ measurement by EPT, tests of validation have been successfully done with normal saline and chicken breast phantoms^3,4; but validation of σ of tumor measured in vivo by EPT with that by ex vivo measurement using a dielectric probe has not yet been tested. The purposes of this study were two folds: (1) to determine the electrical conductivity characteristics of glioma in treatment-naïve patients, and (2) to test correlation between σ of glioma measured in vivo using EPT and that measured ex vivo using a dielectric probe.This study included 24 treatment-naïve glioma patients with confirmed diagnosis (8 Grade II, 8 Grade III, and 6 Grade IV). EPT was performed with a Steady-State Free Precession (SSFP) sequence with non-selective RF pulse (TR/TE/α/NEX=3.5 ms/ 1.7 ms/ 25°/2, Voxel size=1.3 x 1.2 x 1 mm³), using a 3T scanner (Achieva TX, Philips Medical Solutions, Best, the Netherlands). Axial pre- and post-contrast-enhanced T1WI, T2WI, and FLAIR imaging were also performed. σ maps were constructed from the phase images of EPT, using the formula σ= (2μ₀ω)^-1ΔΦ, where μ₀=mean magnetic permeability of the body, ω=Larmour frequency, Δ=Laplacian operator, and Φ=transceive phase. In each patient, the contrast-enhanced (CE) and non-contrast-enhanced (NCE) tumor components, and normal-appearing brain parenchyma (N) were segmented semiautomatically, from the pre- and post-contrast-enhanced T1WI, T2WI, and FLAIR images. These components were then superimposed onto the σ maps coregistered with the other images. The major histogram metrics were derived, and compared among the components and tumor grades. Friedman with posthoc Wilcoxon tests or one way ANOVA and posthoc LSD tests were used to determine significance. Receiver-operating characteristic (ROC) analysis was also performed to determine accuracy in grading glioma. Correlation between the histogram metrics of each component and tumor proliferation index (Ki67) was also tested. Correlation between σ of glioma measured in vivo using EPT and that measured ex vivo using a dielectric probe (N1500A, Keysight Technologies,USA) was tested in 11 patients, in whom tissue sample of the tumor was available for ex vivo measurement. For correlation analyses, Pearson’s product-moment correlation analysis was used to determine significance.For all tumor grades, the mean σ was significantly different among the CE and NCE tumor components and N (P<0.05). The CE component mostly had the highest mean σ, followed by NCE component. Mode of σ of CE and NCE tumor components of Grade III and Grade IV gliomas was significantly higher than N (P<0.05) (Figure 1). Grade IV gliomas had significantly higher mode of σ of the NCE component than the Grade III gliomas (P<0.05), and tended to have higher value than Grade II gliomas (Figure 2). ROC analysis revealed that, with a cut-off value of 914.41 mS/m, mode of σ of the NCE component would distinguish Grade IV gliomas from the other grades with 80% sensitivity and 86% specificity. The normalized mode of σ of NCE tumor component showed significant strong positive correlation with Ki67 (r=0.65, P=0.01)(Figure 3). Significant strong positive correlation was also observed between the normalized mode of σ of NCE tumor component and σ measured by dielectric probe (r=0.73, P=0.01)(Figure 4).The observed σ difference between glioma and N is thought to be due to destruction of cell membranes that results in loss of insulating barriers and increase in sodium ion concentration and/ or content in the tumor environment^5,6. The σ difference among the tumor components and N may imply that EPT can be used to identify tumor and allows tumor segmentation. The σ of NCE component may distinguish Grade IV gliomas from the lower grade gliomas which have better prognosis, with high sensitivity and specificity. Increase in σ with increase in Ki67 is thought to be due to high sodium concentration within the rapidly proliferating tumor cells⁶. This is the first report which validates the noninvasive σ measurement of gliomas. Failure to observe significant correlation between σ of CE component and that measured by dielectric probe may be due to limited sample size (50% of Grade II gliomas did not enhance).This study determined the electrical conductivity characteristics of glioma noninvasively and in vivo, and validated noninvasive σ measurement of glioma by EPT.