| The electrical properties of biological tissue reflect the structure,function,and pathology of the tissue.At present,many papers have proved that the electrical properties of biological tissues,as imaging markers,have important clinical application value.The pathological process of human body is accompanied by changes in electrical properties of tissues,especially in the process of canceration,the electrical properties change greatly.Therefore,mapping electrical properties distribution within biological tissues through electrical properties imaging technology is expected to provide extremely useful information for the early diagnosis of diseases?especially for cancer?.Magnetic resonance electrical properties tomography technology is an emerging electrical properties imaging technology developed after the beginning of open-ended coaxial probe method and magnetic resonance electrical impedance tomography technology.This technology has attracted extensive attention and research in recent years due to its advantages of non-invasive,no need for external current injection,and simple operation.The MREPT technology needs to be implemented in combination with magnetic resonance radio frequency coil,B1 mapping sequence used to obtain the radio frequency field distribution,and electrical properties reconstruction method.In recent years,research on the MREPT reconstruction method has mainly focused on field strengths of 3T and above.The electrical properties of biological tissues are highly correlated with the frequency of the electromagnetic field.The reconstruction performance of the reconstruction method may be affected by the field strength of the magnetic resonance main magnetic field.At present,the 1.5T magnetic resonance imaging system is widely used in clinic.This article mainly studies the reconstruction performance of the traditional electrical properties reconstruction method and the gradient-based conductivity imaging using only phase method.For the traditional electrical properties reconstruction method,the influence of the transceiver phase assumption error,model invalidity error,finite difference kernel error,voxel size,and noise reduction strategy on the reconstruction result are studied.For the gradient-based conductivity imaging using only phase,the traditional method is compared,and the effects of diffusion terms and boundary conditions on the reconstruction results are analyzed.The work in this paper will provide a certain reference for the next practical clinical experiment,and provide technical support and theoretical guidance for promoting the practical process of electrical properties tomography technology under the 1.5T magnetic resonance system.In order to carry out the above research,the RF coil is simulated on the electromagnetic simulation software——Sim4Life,and the homogemeous phantom,the heterogeneous phantom and the human head model are loaded into the RF coil to obtain the ideal RF field distribution inside different subjects.Reconstruction and analysis are based on the obtained RF field distribution information.The experimental results show that for the traditional electrical properties reconstruction method,the transceiver phase assumption error at 1.5T has little effect on it,and the standard deviation of the absolute phase error is between 0.611°and 1.1241°,and the effect on the reconstruction result is negligible.The model invalidity error has great influence on its reconstruction results,even the slight changes in the electrical properties of biological tissues with the spatial position will cause great errors;the finite difference kernel error experimental results show that the finite difference kernel is very sensitive to noise,when the SNR is reduced from 100dB to 50dB,the(1<30%?proportion of relative errors less than 30%?of the K3and K5 kernels is reduced from 67.33%and 89.99%to 0.32%and 1.33%,respectively,and the noise reduction strategy can effectively improve the reconstruction quality.For the K3 convolution kernel,when SNR=75dB,the noise reduction method can increase(1<30%from 6.16%to 80.05%.In addition,it is found that the large finite difference kernel is more robust,but it will cause more serious boundary artifacts.When SNR=?,the(1<30%of K3 and K5 kernels decrease from 93.25%and 90.07%to 90.02%and83.88%as the voxel size increases.For the gradient-based conductivity imaging using only phase method,in the reconstruction of complex structures,such as brain,this method exhibits superior reconstruction performance compared to traditional methods,and successfully solves the problem of boundary artifacts in traditional methods.The overall reconstruction accuracy is much higher than traditional methods,and it is more robust to noise. |
PDF Full Text Request