International Journal on Magnetic Particle Imaging (IJMPI)
Not a member yet
555 research outputs found
Sort by
An effective reconstruction method for magnetic particle imaging based on deep neural networks constrained by a physical model
No full textMagnetic particle imaging (MPI) visualizes the spatial distribution of magnetic particles based on their nonlinear response signals. Such a process can be implemented by image reconstruction. In recent years, deep learning techniques supported by large amounts of training data have been widely used in various medical image reconstructions. However, the acquisition of MPI data requires a long time of obtaining and preprocessing. This makes it impractical to obtain a sufficient amount of data for training. In this work, we propose an MPI image reconstruction framework that incorporates physical model constraints into deep learning networks to overcome the limitation of dependence on training data. This framework optimizes the network parameters through constraints of the physical model rather than training with paired data. Simulation results show that this is an effective reconstruction strategy and has good reconstruction robustness
2-D magnetic particle imaging based on actively feed-through fundamental frequency recovery
No full textIn the Magnetic particle imaging (MPI) scanner, the fundamental frequency component of the particle signal cannot be accurately detected due to the mutual inductance between the excitation coil and detection coil. In this work, we proposed a method to recover the fundamental frequency component of the particle signal, which uses the hardware circuit to implement the real-time difference between the direct feed-through signal caused by the excitation field and the actively generated feedforward cancelation signal. As a result, we successfully acquired the complete particle signal and reconstructed the 2-D image by the X-Space reconstruction algorithm. The experiment results demonstrated the method could improve the sensitivity of the MPI scanner
Using Negative Bolus in Dynamic MPI
No full textIn Magnetic Particle Imaging, the spatial distribution of a tracer is measured and depicted with a concentration dependent signal intensity for any location that inhibits particles, whereas surrounding tissue does not provide any signal. After tracer injection, the signal over time (positive contrast) can be utilized as a transient response to calculate dynamic diagnostic parameters like perfusion parameter maps. In this work, a bolus of physiological saline solution without any particles (negative contrast) is proposed, where the remaining steady state concentration contributes to the image contrast. This opens up the possibility to stretch the total monitoring time of a patient by utilizing a positive-negative contrast sequence, while keeping the total iron dose constant in the subject. Resulting time responses show that normalized signals from positive and negative boli are concurrent in the phantom experiments, indicating identical diagnostic parameters for in-vivo use
Optimization of the coprecipitation in the synthesis of SPIONs with rigorize control of process conditions
No full textIn this work, superparamagnetic iron oxide nanoparticles were synthesized using an alkaline coprecipitation. Two different bases (NaOH and NH3) were used. Furthermore, the influence of temperature on the properties of the products was investigated. Three different Dextran derivates were used as shell material for the syntheses. The products of the different syntheses were investigated by Magnetic Particle Spectroscopy (MPS) and Photon Correlation Spectroscopy (PCS). The magnetic properties and the hydrodynamic diameter were used as quality criteria of the produced SPIONs. The results of this studies were evaluated and carefully compared
Modeling the Image Reconstruction Problem in MPI
No full textIn this work, we develop a new operator-based approach for MPI image reconstruction that can directly map arbitrary time domain data into images. The method enables efficient, rapid, and accurate image reconstruction using a diversity of transmit and receiver coils while making minimal assumptions regarding the underlying physics.
The system model maps the underlying image, x, to the acquired time domain signal, b, by a sequence of 4 linear matrix operators given by A = GVEH. We then solve for x in a system model b = Ax + n, where n is a noise term, using standard approaches such as regularized least squares. The operator H is derived from the Langevin model and is a combination of convolutional operators, E is a row selector matrix that selects the field accessible point (FFP) at each measured time point, V is related to the FFP velocity at each time point, and G incorporates the system filter of the transmit signal. Each of these operators can be implemented efficiently as matrix-vector products, which makes the inverse problem computationally tractable. The method also enables simultaneous application of physical information such as non-negativity constraints and smoothness.
Initial results show that our approach effectively and efficiently reconstructs 2D images from simulated and experimental data for multiple transmit and receiver coils. This approach provides a pathway for significantly reducing the computational pipeline of future MPI systems, and it will allow for greater flexibility in MPI scanner design.
The current model does not include any nanoparticle relaxation effects and assumes a homogenous B-field. Incorporating these components is planned for future work. 
Field-swept magnetic particle spectroscopy using a transverse DC field
No full textThe transverse magnetization of magnetic nanoparticles (MNPs) induced by a DC field perpendicular to the AC field is of increasing interest in magnetic particle imaging (MPI) and magnetic particle spectroscopy (MPS). It can be used to improve the image quality in MPI and to enhance the sensitivity of MPS without direct feedthrough interference. In this work, we developed a field-swept MPS to observe the response curve of the transverse magnetization by varying the transverse DC field amplitude, and reported a new metric to characterize the nonlinear properties of MNPs, called DCM (the DC field amplitude at the maximum harmonic amplitude). Two commercial MNPs (synomag®-70 and synomag®-50) were investigated. The experimental results show that the DCM is different for the two MNPs and independent of the concentration of MNPs. In addition, the DCM was positively correlated with the viscosity of the sample solvent, which further indicates a correlation with Brownian relaxation
Dual Contrastive Learning with Adversarial Framework for Magnetic Particle Imaging Deblurring
No full textMagnetic particle imaging (MPI) is an emerging medical imaging technique that has high sensitivity, contrast and excellent depth penetration. In x-space MPI reconstruction, the reconstructed native image can be modeled as a convolution of the magnetic particle concentration with a point-spread function (PSF). The deconvolution is practical and valuable as a post-processing way to deblur the native image. However, to accurately measure or model the PSF used for deconvolution is challenging due to the imperfection of hardware and magnetic particle relaxation. The inaccurate PSF may lead to the loss of the content structure of the MPI image. In this study, we developed a dual adversarial framework with contrastive constraint (DC_GAN) to deblur the MPI image. We evaluate the performance of the proposed DC_GAN model on simulated and real data. Experimental results confirm that our model performs favorably against the deconvolution method that are mainly used for deblurring the MPI image
Deconvolution of direct reconstructions for MPI using Convolutional Neural Network
No full textRecently, an approach was presented that allows a direct and fast image reconstruction without the use of a systemmatrix for Lissajous trajectories of the so called field free point. The method is based on weighting frequencycomponents of the measured voltage signals and additional factors with Chebychev polynomials of the secondkind, resulting in reconstructions of the convolved spatial distribution of magnetic nanoparticles. In order toobtain meaningful images, these reconstructions have to be deconvolved afterwards. For this purpose, differentmethods have already been proposed. In this work, a U-shaped neural network is used for the deconvolution. Thenetwork was trained and tested on simulated data of blood vessel like structures. The proposed model outperformsconventional methods and improves the image quality of the reconstructions
Fast dynamic MPI cytometry
No full textWe developed in vivo fast dynamic MPI “cytometry” to quantify and localize the accumulation of SPIO-labeled stem cells in different organs after intra-arterial injection on a time scale of minutes. Bone marrow-derived mesenchymal stem cells (MSCs) and superoxide dismutase 1 gene-corrected neural precursor cells (NPCs) were labeled with Resovist and injected into Rag2 mice using four separate injections. Whole body standard 2D/3D MPI scans were obtained, quantified and co-registered with CT. Using cell calibration fiducials, cells could be clearly visualized and quantified by MPI in vivo in the brain, liver, and lung. The cytometric ratio of the number of cells in the liver/lung vs. the brain was 1.5 for MSCs and 15.6 for NPCs, respectively, at 24 min post-injection. Fast dynamic MPI cytometry may find applications for optimizing the dose, volume, speed and route of administration when performing interventional cell therapy procedures in real-time
3D Trajectory Analysis for Tomographic Field Free Line Magnetic Particle Imaging
No full textIn magnetic particle imaging (MPI), spatial mapping of the magnetic nanoparticles is achieved by scanning the field-free region (FFR) throughout the field of view (FOV). The scanning trajectory of the FFR has an impact on image quality and scan time, and also is subject to limitations such as hardware requirements and patient safety. Here, we analyzed 3D imaging performance of an open-sided MPI system using simulations for various trajectories and SNR values. A selection field of 0.35 T m?1 gradient with an FFL was created and scanned tomographically. Forlow SNR scenarios, the best imaging results were achieved at the expense of a longer scan time when 3D FOV was densely scanned layer by layer. For more sparse trajectories, the effect of the coherence between the scan angles became prominent with the increasing noise level. Image quality can be improved by assigning non-coherent FFL angles to the consequent layers while taking hardware limitations into account