Research On Post-processing Optimization Technology Based On Cardiac Diffusion Tensor Imaging

Cardiac diffusion tensor imaging(cDTI)is a unique technology that can assess the microstructure of the in-vivo myocardium non-invasively without the need for contrast agents.Unlike traditional histological detection,cDTI performs imaging based on the diffusion characteristics of water molecules in the human body in tissues,uses sequence technology that applies a diffusion-sensitive weighted gradient for image acquisition,and establishes a Gaussian motion index model to model signal attenuation.Finally,the tensor is used to characterize the anisotropy of the diffusion,and a series of characteristic parameters are derived to analyze the changes in the myocardial structure quantitatively.Studies have shown that cDTI technology can describe the microstructure of the dynamic heart,which can be used to assess myocardial regeneration and the differential diagnosis of cardiomyopathy and characterize changes at the cellular level in myocardial infarction,hypertrophic cardiomyopathy,and dilated cardiomyopathy.It has great potential and clinical application prospects.At present,because the image is very susceptible to noise in practical applications,the accuracy and stability of the parameter are limited.Simultaneously,the motion of the heart causes serious loss of magnetic resonance signals,which cannot be used for subsequent parameter analysis.Besides,the acquisition time is too long due to the breathing and heart motion,and the technology has not been applied in clinical settings as a conventional technical means.This article explores and solves the limitations in the development of cDTI technology generally.Firstly,it starts from the principle of magnetic resonance imaging(MRI),then summarizes the excitation and encoding,acquisition and image reconstruction of signals in MRI,and elaborates the theory of cardiac diffusion imaging,that is,the use of diffusionsensitive gradients causes the phase accumulation of moving molecules,resulting in signal attenuation.It establishes a tensor model to characterize the diffusion process.At the same time,it contrasts the commonly used imaging sequence technology,including STEAM and the steady-state free precession(BSSFP)sequence technology with diffusion preparation,analyzes and compares their advantages and disadvantages,and introduces the establishment of tensor analysis model and the main parameters and physiological significance in the quantitative analysis model in detail.To solve the inherently low signal-to-noise ratio of the diffuse image,the three different denoising algorithms in the denoising preprocessing of image data are evaluated on the effect of image quality and the accuracy of parameter results based on in-vivo and simulation experiments.SNR and RMSE are used as evaluation criteria,and multiple sets of data are statistically analyzed,and it is concluded that LPCA is suitable for DWI image denoising.It is determined to use LPCA as a denosing algorithm to improve the accuracy of parameter analysis.A cDTI in-vivo data acquisition program is designed.In response to the signal loss caused by involuntary heart motion and short T2 relaxation during in-vivo acquisition,the signal acquisition scheme of the diffusion-weighted spin-echo plane echo(DWI-SE-EPI)sequence technology dual-triggered with breathing navigation and ECG gating is used.The second-order motion compensation(M2C)gradient is implemented on the sequence to reduce the signal loss caused by heart motion during the acquisition process.It summarizes the optimal imaging parameters in the acquisition of volunteer heart,establishes a complete in-vivo data analysis program,and verifies that the use of M2 C sequence in diastole and multiple repetitions can effectively improve imaging stability and quality.Simultaneously,volunteer experiments are carried out for imaging parameters such as delay time,repetition times,and directions during the acquisition process.Control variables and comparative analysis are conducted with the existing published results,and the optimal parameters for in-vivo acquisition are finally determined.Since there is no standardized processing program globally,and there is no mature processing software to support cDTI image analysis in China,this thesis designs a complete data processing flow and develop cDTI data processing software to achieve batch processing of DWI from mainstream manufacturers.The software is easy,friendly,functional,simple,and scalable.It provides clinicians and scientific researchers with an efficient tool software for cDTI.