Measurement of glioma electrical properties for data-driven electrical properties tomography

Introduction Electrical properties (EPs) are widely used to assign values to electromagnetic simulation models, providing the comprehensive datasets needed for data-driven electrical properties tomography (EPT) reconstruction. However, the key limitations in current research on glioma EPs are the small sample sizes and the lack of EPs at certain Larmor frequencies. Therefore, this study aims to provide accurate glioma EPs in the full frequency range required for EPT. Methods We measured 128 samples, including 81 gliomas and 47 adjacent surrounding tissues, in the frequency range of 50 MHz to 4 GHz using an open-ended coaxial probe (OCP) method. The accuracy of our measurements was verified through uncertainty analysis. Using the Cole–Cole model, we obtained the EPs across the full frequency spectrum. Subsequently, we developed a human simulation model to evaluate the impact of different EPs on the simulated B1 field. At frequencies of 64 MHz, 128 MHz, 213 MHz, and 298 MHz, the model was assigned either the EPs measured in this study or those reported in the literature, and the relative differences in the simulated fields were calculated. Results We first analyzed the discrepancies between our measured EPs and those from previous studies. The results suggested that small sample sizes may introduce bias, potentially compromising the reliability of the findings. Moreover, interpolated data may fail to reflect the frequency-dependent behavior of EPs, leading to reduced accuracy. Simulation results indicate that differences in EPs can lead to measurable deviations in magnetic field simulations. This variation in the <mml:math xmlns:mml="http://www.w3.org/1998/Math/MathML" id="m1"> <mml:mrow> <mml:msubsup> <mml:mi>B</mml:mi> <mml:mn>1</mml:mn> <mml:mo>+</mml:mo> </mml:msubsup> </mml:mrow> </mml:math> field in response to EPs is critical for enabling accurate EPs reconstruction. Discussion These findings underscore the importance of precise EPs measurements, as errors in the training dataset may compromise the performance of models. Therefore, the reliable EPs of glioma provided in this study have the potential to enhance the accuracy of data-driven EPT, thereby assisting clinical diagnosis and treatment.