Magnetic Particle Imaging: Publications

Magnetic particle imaging (MPI) is a young imaging modality using the nonlinear magnetization properties of superparamagnetic iron-oxide nanoparticles to acquire them. It is a highly sensitive and fast method allowing both a quantitative and a qualitative analysis of the measured signal. Since its first publication in 2005, several different scanner types have been presented. Most of them work with permanent magnets and therefore have a small field of view. In 2014, an alternative scanner concept, the traveling wave MPI (TWMPI), was presented, which allows scanning an entire mouse-sized volume at once. In this paper, a detailed description of the main field generator, the dynamic linear gradient array (dLGA), used for TWMPI is given. This paper guides the reader to the first dLGA prototype, deriving relations between the length of the dLGA and the diameter of the segment coils.

Magnetic particle imaging (MPI) is a promising new tomographic imaging method to detect the spatial distribution of superparamagnetic iron-oxide nanoparticles (SPIOs). The aim of this paper was to investigate the potential of MPI to quantify artificial stenoses in vessel phantoms. Custom-made stenosis phantoms (length 40 mm; inner diameter 8 mm) with different degrees of stenosis (0%, 25%, 50%, 75%, and 100%) were scanned in a custom-built MPI scanner (in-plane resolution: ~1–1.5 mm and field of view: 65 × 29 × 29 mm3). Phantoms were filled with diluted Feru-carbotran [SPIO agent, 5 mmol (Fe)/l]. Each measurement (overall acquisition time: 20 ms per image, 400 averages) was repeated ten times to assess reproducibility. The MPI signal was used for semi-automatic stenosis quantification. Two stenosis evaluation approaches were compared based on the signal intensity profile alongside the stenosis phantoms. Using a novel multi-step image evaluation approach, MPI allowed for accurate quantification of different stenosis grades. While low grade stenoses were slightly over-estimated, high grade stenoses were slightly underestimated. In particular, the 0%, 25%, and 50% stenosis phantoms revealed a 6.2% ± 0.8, 25.7% ± 1.0, and 48.0% ± 1.5 stenosis, respectively. The higher grade 75% stenosis phantom revealed a 73.3% ± 2.8 and the 100% stenosis phantom a 95.8% ± 1.9 stenosis. MPI accurately visualized and quantified different stenosis grades in vessel phantoms with high reproducibility demonstrating its great potential for fast and radiation-free preclinical cardiovascular imaging.

Purpose: To investigate the potential of real-time magnetic particle imaging (MPI) to guide percutaneous transluminal angioplasty (PTA) of vascular stenoses in a phantom model. Materials and Methods: Experiments were conducted on a custom-built MPI scanner. Vascular stenosis phantoms consisted of polyvinyl chloride tubes (inner diameter 8 mm) prepared with a centrally aligned cable tie to form ~ 50% stenoses. MPI angiography for visualization of stenoses was performed using the superparamagnetic iron oxide nanoparticle-based contrast agent Ferucarbotran (10 mmol (Fe)/l). Balloon catheters and guidewires for PTA were visualized using custom-made lacquer markers based on Ferucarbotran. Stenosis dilation (n = 3) was performed by manually inflating the PTA balloon with diluted Ferucarbotran. An online reconstruction framework was implemented for real-time imaging with very short latency time. Results: Visualization of stenosis phantoms and guidance of interventional instruments in real-time (4 frames/s, ~ 100 ms latency time) was possible using an online reconstruction algorithm. Labeling of guidewires and balloon catheters allowed for precise visualization of instrument positions. Conclusion: Real-time MPI-guided PTA in a phantom model is feasible.

Die Magnetpartikel-Bildgebung (Magnetic Particle Imaging, MPI) detektiert superparamagnetische Eisenoxid-Nanopartikel hochsensitiv und hochspezifisch auch im Inneren von Körpern. Physikalisch nutzt sie die nichtlineare Magnetisierungsfunktion dieser Partikel. MPI eröffnet der Medizin so einen neuen Zugang zu markerbasierten Anwendungen, der im Gegensatz zu etablierten Verfahren auf Radionuklide verzichtet und damit gerade in der klinischen Anwendung enormes Potenzial besitzt. Mit der Realisierung erster MPI-MRT-Hybridsysteme ist bereits ein wichtiger Schritt in Richtung einer diagnostischen Anwendbarkeit getan. Dabei erfasst MPI das Markersignal und MRT den anatomischen Bildhintergrund. MPI ist auch in den Materialwissenschaften bei der zerstörungsfreien Materialprüfung einsetzbar.

In the last decade research in the Magnetic Particle Imaging community focused mainly on hardware development as well as the optimization of reconstruction. These steps were essential for improving MPI technology and advance it to pre-clinical applications. However, another important part is the software package working behind MPI systems. This software has to be able to keep up with modern scanner features and should be able to support features such as adjustable scanning trajectories or multiple reconstruction methods. Especially a real-time reconstruction feature is important to use the full potential of the fast imaging capabilities of modern MPI scanners. This could be pivotal in applications such as guided vascular interventions. In this work, a software package is presented, which improves reconstructing MPI data with different reconstruction methods in real-time and with low latency. 

The first two manuscripts in this second issue of the second volume of the International Journal on Magnetic Particle Imaging are covering novel reconstruction techniques. While there are already established methods available the very young field of magnetic particle imaging (MPI) still offers much room for new concepts and algorithms to improve image quality, quantification, or processing speed. The remaining manuscripts are dealing with measurements obtaining additional information on top of tracer localization and the study of magnetic nanoparticle processing in organisms to help establish MPI as a clinical diagnostic tool.

Magnetic Particle Imaging is a tomographic imaging technique offering high sensitivity and temporal resolution and is a promising tool for pre-clinical applications. The state of the art Traveling Wave Magnetic Particle Imaging (TWMPI) scanners offer a mouse-sized FOV, which can be scanned at once. Since the first introduction of TWMPI, several sequences have been introduced to scan the FOV in 2D as well as 3D. So far, the proposed 3D sequences do not provide real-time capability, which is an important feature for future clinical applications such as guided vascular interventions. In this study a modified sequence for TWMPI scanners is presented, which allows scanning an entire 3D volume on a short time scale. Furthermore, for real-time capability a novel reconstruction method using an image-based approach is used providing a 3D visualization in real-time.

Introduction: Common MPI scanners have difficulties to correctly visualize areas with a broad signal intensity spectrum due to their low signal dynamic capability. The aim of this study was to develop a selective signal suppression technique using an advanced trajectory design to circumvent this issue. Materials and Methods: Simulations where performed for a traveling wave (TW) MPI scanner. In the so called slice-scanning mode the field free points (FFPs) travel through the scanner on a sinusoidal trajectory given by the frequencies of the dynamic linear gradient array and a perpendicular saddle coil system. The signal can be selectively suppressed by locally reducing the speed of the FFPs. Series of measurements with two point samples containing different SPIO concentrations with and without selective signal suppression technique were simulated. Results: When imaging samples with two markedly different SPIO concentrations in the slice-scanning mode the signal of the lower concentration cannot be properly reconstructed. Applying the signal suppression technique to the area of the high concentration sample selectively reduces this signal and demarcates samples with lower concentrations. Conclusion: A selective signal suppression technique for TWMPI is introduced which allows to focus on low SPIO concentrations in samples with a large range of tracer concentrations. This can facilitate preclinical functional imaging in the proximity of organs with high unspecific SPIO uptake such as the liver. 

Magnetic particle imaging (MPI) is a non-invasive imaging modality for direct detection of superparamagnetic iron-oxide nanoparticles based on the nonlinear magnetization response of magnetic materials to alternating magnetic fields. This highly sensitive and rapid method allows both a quantitative and a qualitative analysis of the measured signal. Since the first publication of MPI in 2005 several different scanner concepts have been presented and in 2009 the first in vivo imaging results of a beating mouse heart were shown. However, since the field of view (FOV) of the first MPI-scanner only covers a small region several approaches and hardware enhancements were presented to overcome this issue and could increase the FOV on cost of acquisition speed. In 2014 an alternative scanner concept, the traveling wave MPI (TWMPI), was presented, which allows scanning an entire mouse-sized volume at once. In this paper the first in vivo imaging results using the TWMPI system are presented. By optimizing the trajectory the temporal resolution is sufficiently high to resolve the dynamic of a beating mouse heart.

Different types of scanners have been presented since the first publication of Magnetic Particle Imaging (MPI) in 2005. As a result, there are different types of reconstruction methods available, which can be separated into two basic concepts: reconstruction using a system matrix and reconstruction using a direct deconvolution. Both methods have their merits and drawbacks. For the first approach hardware parameters like sampling rate and frequencies have to be chosen carefully to fit the parameter selection process required for the system matrix. For the other approach the temporal and spatial homogeneity of the magnetic field gradient over the entire FOV has to be high to perform an accurate reconstruction, which results in smaller FOVs. In this paper a novel reconstruction method is presented, which combines the advantages of both reconstruction methods to be more flexible during the entire reconstruction process. Furthermore, it enables the possibility of performing a dynamic patch reconstruction, which allows to select arbitrary areas of the FOV for higher resolution reducing reconstruction time significantly. In addition, this new reconstruction improved the image quality of a Traveling Wave MPI scanner substantially.

Due to its broad range of applications in biology and medicine, detection of magnetic particles gains increasing importance. The established method of magnetic particle imaging (MPI) bases on the non-linear magnetization behavior of superparamagnetic nanoparticles [1]. As an alternative approach, rotational drift spectroscopy (RDS) [2,3] aims at detecting the properties of magnetic particles in liquid suspensions based on their motional behavior inside an external rotating magnetic field. Serving as basis for the physical understanding and further development of RDS, the rotational behavior of particles with and without motional restriction is studied by numerical simulation. For this purpose, two different approaches were implemented: one using an explicit Euler algorithm and an improved version using a semi-analytical two-step procedure in order to overcome numerical instabilities of the original code. The results show independently of motional restriction two characteristic types of rotational behavior. The particles either follow the external field with a locked phase lag or exhibit a rotational drift.

Magnetic particle imaging (MPI) was first presented in 2005. It is based on the nonlinear response of ferromagnetic material and the fact that the magnetization saturates at sufficiently high magnetic fields. In contrast to magnetic resonance imaging (MRI), MPI directly detects the concentration and distribution of superparamagnetic iron-oxide nanoparticles without any background of any tissue. To overcome this issue, a traveling wave MPI (TWMPI) device was combined with a low field MRI scanner to demonstrate the feasibility of a hybrid scanner, which contains both imaging modalities in a single device. The hardware of both separate approaches should be improved and optimized to reach higher fields and a higher resolution, especially for the MRI measurement. Therefore, the dynamic linear gradient array from the TWMPI scanner was modified in a way to produce also a homogenous magnetic field, which can be used for MRI.

Since the introduction of magnetic particle imaging in 2005, several different types of scanner were presented. One of them is the traveling wave magnetic particle imaging (TWMPI) scanner. It uses for imaging a dynamic linear gradient array for a dynamic generation of a strong gradient and field free point. An unresolved issue of the TWMPI approach so far is the non-isotropic spatial resolution. As an alternative in this paper, a rotating slice scanning mode for TWMPI is presented to overcome this issue. This approach rotates the scanning-slices around the scanner axis and uses a projection reconstruction method to get a 3-D volume with a high isotropic resolution.

Magnetic particles have become a core ingredient for many applications in chemistry, biology, and medical diagnostics, e.g., as a basis for bioanalytical methods or as tracer material for medical imaging. This paper presents a new method called rotational drift spectroscopy (RDS) which uses rotating magnetic fields for measuring the properties of magnetic nanoparticles (MNPs) in liquid suspensions. The RDS signal is based on the nonlinear rotational drift behavior of MNPs in rotating magnetic fields, which is highly dependent on the properties of the MNPs as well as their interaction with the environment. This dependency allows detecting the binding of functionalized MNPs with, e.g., proteins, viruses, or cells with potentially very high sensitivity. This paper presents first experiments demonstrating rotational drift behavior on aggregated magnetic particle ensembles and the corresponding experimental setup.

Magnetic particles have become a core ingredient for many applications in chemistry, biology, and medical diagnostics, e.g., as basis for bioanalytical methods or tracer material for medical imaging. This paper discusses the theory and presents numerical simulations of a new method called rotational drift spectroscopy (RDS), which uses rotating magnetic fields for measuring the properties of magnetic nanoparticles (MNPs) in liquid suspensions. The RDS signal is based on the nonlinear rotational drift behavior of MNPs in rotating magnetic fields, which is highly dependent on the properties of the MNPs as well as their interaction with the environment. This dependence makes it a highly promising tool for detecting the binding of functionalized MNPs with, e.g., proteins, viruses, or cells. This paper presents the theory and numerical simulations of this method.

Since the first publication in 2005, several different scanner types for magnetic particle imaging (MPI) have been presented. One of these scanner concepts is traveling wave MPI (TWMPI). It uses a dynamic linear gradient array, which generates and moves a field free point with a strong gradient, which is necessary for scanning the sample in 3-D. Due to the linear properties of the TWMPI device, very fast 2-D imaging with frame rates higher than 1500 frames/s is possible (superspeed mode). Using the superspeed mode different high speed measurements are conceivable, e.g., fluid-dynamic investigations.

Magnetic particle imaging (MPI) is a novel imaging method for the depiction of superparamagnetic materials. Since the first publication several MPI scanner were presented, which work at gradient strength of ~2-7 T/m. This is reasonable for in vivo imaging and provides a resolution of ~1 mm. In this paper, several approaches for a micro-MPI (μMPI) device for very small samples are presented, which works at a very high magnetic gradient strength of ~85 T/m. This results in a theoretical resolution <100 μm. This μMPI device can be operated at several modes, which have different advantages. In preliminary tests, the feasibility of the μMPI device has been proven.

While magnetic particle imaging (MPI) constitutes a novel biomedical imaging technique for tracking superpara magnetic nanoparticles in vivo, unlike magnetic resonance imaging (MRI), it cannot provide anatomical background information. Until now these two modalities have been performed in separate scanners and image co-registration has been hampered by the need to reposition the sample in both systems as similarly as possible. This paper presents a bimodal MPI-MRI-tomograph that combines both modalities in a single system. MPI and MRI images can thus be acquired without moving the sample or replacing any parts in the setup. The images acquired with the presented setup show excellent agreement between the localization of the nano particles in MPI and the MRI background data. A combination of two highly complementary imaging modalities has been achieved.

Most 3-D magnetic particle imaging (MPI) scanners currently use permanent magnets to create the strong gradient field required for high resolution MPI. However, using permanent magnets limits the field of view (FOV) due to the large amount of energy required to move the field free point (FFP) from the center of the scanner. To address this issue, an alternative approach called “Traveling Wave MPI” is here presented. This approach employs a novel gradient system, the dynamic linear gradient array, to cover a large FOV while dynamically creating a strong magnetic gradient. The proposed design also enables the use of a so-called line-scanning mode, which simplifies the FFP trajectory to a linear path through the 3-D volume. This results in simplified mathematics, which facilitates the image reconstruction.

Determining the composition of solid materials is of high interest in areas such as material research or quality assurance. There are several modalities at disposal with which various parameters of the material can be observed, but of those only magnetic resonance imaging (MRI) or computer tomography (CT) offer a non-destructive determination of material distribution in 3D. A novel non-destructive imaging method is Magnetic Particle Imaging (MPI), which uses dynamic magnetic fields for a direct determination of the distribution of magnetic materials in 3D. With this approach, it is possible to determine and differentiate magnetic and nonmagnetic behaviour. In this paper, the first proof-of-principle measurements of magnetic properties in solid environments are presented using a home-built traveling wave magnetic particle imaging scanner.

Current simulations of the signal in magnetic particle imaging (MPI) are either based on the Langevin function or on directly measuring the system function. The former completely ignores the influence of finite relaxation times of magnetic particles, and the latter requires time-consuming reference scans with an existing MPI scanner. Therefore, the resulting system function only applies for a given tracer type and the properties of the applied scanning trajectory. It requires separate reference scans for different trajectories and does not allow simulating theoretical magnetic particle suspensions. The most accessible and accurate way for including relaxation effects in the signal simulation would be using the Langevin equation. However, this is a very time-consuming approach because it calculates the stochastic dynamics of the individual particles and averages over large particle ensembles. In the current article, a numerically efficient way for approximating the averaged Langevin equation is proposed, which is much faster than the approach based on the Langevin equation because it is directly calculating the averaged time evolution of the magnetization. The proposed simulation yields promising results. Except for the case of small orthogonal offset fields, a high agreement with the full but significantly slower simulation could be shown.