Magnetic Particle Imaging

This study investigates the dynamic magnetic properties of Resovist® for magnetic particle imaging (MPI) utilizing static M-H, ac susceptibility (ACS) and magnetic particle spectroscopy (MPS) measurements on a Resovist® suspension and an immobilized sample. Investigating the magnetic moment and anisotropy energy barrier distributions in the sample as well as the relationship between them, we clarified that the MNPs with large magnetic moment (10

− 24

~10

− 23

Wb·m) and small anisotropy energy barrier play an important role in MPI.

Nonlinear Brownian relaxation of magnetic fluids for the case of large excitation field was studied in relation to its biomedical applications. The Fokker-Planck equation, which describes the nonlinear behavior of magnetic fluids, was solved by numerical simulation when ac field was applied. Frequency-dependences of the harmonics were investigated in terms of the effective Brownian relaxation time

τ

eff

, which was empirically obtained from the ac susceptibility of the fundamental component. It was shown that higher harmonics become small even at

ωτ

eff

=1 compared to each quasi-static harmonics amplitudes.

Nanofluids, defined as fluids containing suspended solid nanoparticles, are potential systems for utilization in biomedical applications. Magnetic Particle Imaging (MPI) uses superparamagnetic nanofluids, e.g. a colloidal suspension of iron oxide particles. In this work a new biocompatible nanofluid based on pure and stable ferromagnetic carbon is investigated. Although this material has a relatively small value of coercive magnetic field, it does exhibit a true ferromagnetic behavior up to 300 K. We present results obtained from numerical investigations performed to calculate the impact of a ferromagnetic magnetization to the MPI signal chain. Moreover, by modeling ferromagnetic magnetization we prove here the general suitability of ferromagnetic materials for MPI. Due to the low saturation magnetization, however, MPI for ferromagnetic carbon will be possible only in the near future when realistic concentrations of the nanofluid ferromagnetic carbon will be experimentally obtainable.

Magnetic Particle Imaging (MPI) is a promising new imaging modality, providing 3-dimensional imaging of magnetic nanoparticle tracers with high spatial and temporal resolution. Some recently developed experimental scanners have proven MPI to be feasible for small animal imaging. So far, one assumes that all particles contributing to the MPI signal share the same size distribution. An interesting extension of MPI would be to measure the mobility (or binding affinity) of the particles in the imaging volume. In this scenario, particles in certain regions may be partly immobilized by chemical binding, resulting in a transition from a Brownian to a Néel-dominated magnetization behavior – which is generally assumed for MNP tracers in blood. We propose that using two distinct frequencies, one below and one above the Brownian-Néel transition frequency, the binding state of the particles can be determined and utilized in MPI imaging. In this paper, we describe our MPI system and present simulations of 2-dimensional “Mobility MPI”.

Magnetic particle imaging usually requires image reconstruction in order to obtain the distribution of the superparamagnetic iron oxide particle concentration from the measured signal. An integral part of the reconstruction is the system function, which relates the superparamagnetic iron oxide distribution to the signal it generates in the MPI scanner. Traditionally, the system function is obtained by measuring the signal generated by a known point-like concentration throughout the imaging volume. This however, requires a robot and is very time consuming. In this contribution we demonstrate a method for estimating the system function using focus fields, which avoids the need for a robot.

To investigate various magnetic nanoparticle samples for their suitability as markers for Magnetic Particle Imaging (MPI), various magnetic techniques can be applied. Whereas Magnetic Particle Spectroscopy (MPS) directly provides the harmonic spectrum for a given excitation frequency and ac field amplitude, for the design of optimal markers, it is important to relate the measured harmonic spectra to specific nanoparticle properties. We have applied fluxgate magnetorelaxometry (MRX), ac susceptibility (ACS) up to 1 MHz, multivariate MPS - varying the excitation field amplitude and frequency as well as the magnitude of a superimposed dc field - and static magnetization measurements for the comprehensive characterization of representative magnetic nanoparticle samples. To distinguish between core and hydrodynamic properties, aqueous suspensions as well as immobilized samples were investigated.

Magnetic Particle Spectroscopy (MPS) has been used to estimate the performance of magnetic nanoparticles for Magnetic Particle Imaging. It was demonstrated that one can reconstruct the particles core size distribution from the measured MPS spectrum. However, using MPS as an analytical instrument for the characterization of magnetic nanoparticle samples includes systematic errors. First, previous reports on MPS-based core size estimation only take the Néel process of the relaxation into account. However, other methods, e.g. ac susceptometry, show that for excitation frequencies in the lower kilohertz regime the Brownian magnetization process also plays an important role. We developed an extended MPS setup that enables parameter measurements of the harmonics, depending on the amplitude and frequency of the excitation signal and the amplitude of a static offset field. Based on a dynamic magnetization model, we utilize a multi-variate fitting routine to describe the sample’s core size and hydrodynamic size distribution. Second, existing implementations assume mono-modal log-normal distributions of core sizes, which cannot be guaranteed to match the actual size distribution of the sample. For that reason, we also use a regularized singular value decomposition (SVD) based method for reconstructing arbitrary core size distributions from the harmonic spectrum.

The search for optimal nanoparticlesfor magnetic particle imaging (MPI) has been receiving much attention. Currently, Resovist® an iron oxide nanoparticle-based MR contrast agent is considered as the gold standard for MPS and MPI measurements. In this paper, we evaluate the initial MPS response of iron oxide-based nanoparticle systems with variation in core size and coating. In this context, we synthesized iron oxide nanoparticles with different core sizes (5 and 10 nm) by employing co-preciptiation method. Further, these iron cores were adsorptivelycoated with endogenous fluorophores (flavin analogues: FMN, FAD) to increase their stability and to enhance MR contrast. We have performed initial experiments on these particles in a magnetic particle spectrometer. Their spectrum of higher harmonics was obtained and found to fall off more rapidly than Resovist®. No difference in magnetic behavior was seen between particles with different coatings.

The Magnetic Particle Spectroscopy (MPS)-amplitudes were measured on 7 suspensions of magnetite based magnetic particles (MNP) differing in core size and magnetic anisotropy. The distributions of the effective domain sizes, estimated by means of quasistatic M(H) measurements and Magnetorelaxometry (MRX), matches well the core size distribution for the single core MNP-systems estimated by electron microscopy. Two systems, namely Resovist and M4E clearly exhibit a bimodal domain size distribution. It was shown, that the MPS amplitudes strongly increase with increasing domain size up to 21 nm, the mean value of the larger fraction of Resovist. For M4E with a mean size of the larger fraction of 33 nm the measured MPS-amplitudes became much smaller than those of Resovist, in particular for the higher harmonics. That behaviour was attributed to the mean anisotropy energy of these MNPs, estimated by MRX, exceeding that of Resovist by one order of magnitude. The effect of the MNP’s magnetic anisotropy is also supported by comparison of measured MPS-amplitudes with those which were calculated on the base of M(H)-data.

In Magnetic Particle Imaging, a new medical imaging modality, the relation between measured signals and the spatial distribution of the tracer particles is described by the system function. In existing scanners the system function is measured by moving a delta probe through the field-of-view and the particle response is measured. This procedure is time consuming, caused by the mechanical movement, and makes great demands on the hardware, because of the huge data amount that is measured. To speed up the determination of the system functions other methods have to be considered. In this work the model based and the hybrid system function will be compared to the measured one.

A new microfluidic system was realized by means of a specially designed chip from silicon and glass. Magnetic disks were located above and below this chip. The transportation, mixing, incubation and the chip based separation procedure of the cell-magnetic bead suspension is caused by pumps and the rotation of permanent magnets arranged outside the chip.

The present article reports on the experimental results of applying magnetic separation to DDM128, a magnetic fluid very similar to Resovist ®, with superior MPI performance. Larger particle diameters are assumed to perform better in MPI signal generation. Thus particle size fractionation is expected to result in significant MPI signal enhancement. We separated DDM128 using magnetic separation at different field strengths. In the following we investigated the particle size distribution of the fractions by magnetization measurements and magnetic particle spectroscopy.

Aim of this study was the investigation of the suitability of magnetic multicore nanoparticles (MCNP) for magnetic particle imaging. For this, MCNP of different cluster sizes were investigated. To obtain a set of samples which differ in their cluster sizes the MCNP were classified into fractions of different mean sizes by centrifugation. By the fractionation particles with hydrodynamic diameters from 100 to 800 nm were obtained. Magnetic measurements confirmed a correlation of the hydrodynamic size with the effective magnetic volume of the particles in these fractions – e.g., with increasing particle size the coercivity of the particles varied from 3.5 to 25.8 Oe. Magnetic particle spectrometry investigations showed a clear dependence of the quality of MPS signal on the hydrodynamic diameter of the particles and thus the cluster size. In particular the amplitude ratio of the higher (15...40) harmonics to the 3

rd

harmonic span over one order of magnitude, where the smaller MCNP showed the highest values.

The use of magnetic nanoparticles as contrast agents in biomedical applications requires highly reproducible markers soluble in aqueous media. Here we present a modular approach for the formation of a magnetic contrast agent, which is composed of an inorganic core like iron oxide and a water solubility mediating polymer shell. The shell can also be a carrier for specific affinity molecules like antibodies or peptides, which offer the possibility to selectively target surface receptors upon cells.

Superparamagnetic single-core iron oxide nanoparticles coated with oleic acid have been synthesized via high temperature decomposition of iron-oleate precursor. The particle’s core and hydrodynamic size distributions are characterized utilizing a variety of methods including Fluxgate Magnetorelaxometry (MRX) and Photon Cross-correlation Spectroscopy (PCCS). The harmonic spectra of mobile and immobile samples were measured using our homebuilt Magnetic Particle Spectrometer (MPS) whereby the suitability of the synthesized particles as a tracer for MPI is analyzed. Furthermore, the MPS spectra of the fabricated single-core nanoparticles are compared with the harmonic spectra of Resovist

®

(Bayer Schering Pharma, Berlin) and Ocean NanoTech nanoparticles.

Since its invention, magnetic particle imaging (MPI) has gained increasing attention in academic as well as industrial research and development. It requires the application of an imaging agent, but the structure-efficacy relations are far from being fully understood and no ideal MPI tracer has been found so far. Here, we present a systematic investigation of the size dependence of the MPI spectra of identically composed, but differently sized iron oxide nanoparticles. We furthermore present a small angle X-ray scattering (SAXS) study as a route to assess the particle core structure. To that goal we used FeraSpin

TM

R and the FeraSpin

TM

Series (XS to XXL). We show that FeraSpin R offers an equal whereas FeraSpin L to XXL offer an improved MPI signal as compared to the hitherto “gold-standard” Resovist®. Moreover, the FeraSpin Series constitutes a versatile “toolbox” for MPI tracer research.

Today, a variety of different nanoparticles are used in various applications. In particular, super-paramagnetic iron oxide nanoparticles (or SPIONs) are used

in

vitro

for cell separation and

in

vivo

for hyperthermia or as contrast agent for magnetic resonance imaging (MRI). However, in Magnetic Particle Imaging (MPI), SPIONs play a fundamental role as tracer material. In addition to the overall size of the particles and the particle coating, which are important for medical applications, it is the magnetic core diameter that is relevant for the performance in MPI applications. In general, for

in

vivo

applications, the stability of the particles is of key importance. Therefore, in this paper, the stability of the SPIONs have been analyzed in different particle suspension media.

X-ray computed tomography is a widely used imaging technique nowadays. Especially in medicine it takes an important role for visualization and diagnostics. Micro-computed tomography (

μ

CT) follows the same principle as conventional medical CT-Scanner. But the objects analyzed are smaller, thus an improvement in spatial resolution down to few micrometers is possible. In the work field of biomedical application of magnetic nanoparticles

μ

CT has been used for the visualization of the nanoparticle accumulations within tumoral regions after magnetic drug targeting. Further on, a calibration procedure has been developed and applied with a

μ

CT-apparatus. The calibration procedure enables a semi-quantification of the nanoparticle content within the tissue samples. The next step stone in visualization process is the observation of the nanoparticle accumulation during the application of magnetic drug targeting in an animal. Thus, we have tested the calibration procedure on a commercial animal scanner. In this paper we compare the semi-quantitative results figured out in two different

μ

CT-equipments.

Detrimental effects of nanoparticles on cell viability, cell growth and morphology have been an ongoing topic in the field of nanoparticle tissue engineering, cell labeling, and drug delivery. Establishing biocompatible nanoparticles is particularly important for stem cell -based therapies and cell-tracking by magnetic particle imaging in regenerative medicine. Recently, magnetic particle imaging (MPI) has been presented as a new method for the measurement of the spatial distribution of superparamagnetic iron oxide nanoparticles (SPIOs). Spatial resolution and signal to noise ratio of MPI depend on the particle quality. Here we developed dextran-coated SPIOs for magnetic particle imaging and analyzed their stability and hydrodynamic diameter by photon cross correlation spectroscopy (PCCS). The uptake of SPIOs and the morphology of labeled human adult stem cells were examined by confocal laser scanning microscopy. Labeled stem cell growth was monitored by the xCELLigence system. We focused on commonly used

in

vitro

assays for estimation of cell viability and cell death. The newly developed SPIOs revealed a low signal to noise ratio. Dextrancoated SPIOs had no significant influence on cell growth and viability of human adult stem cells. Our data support dextran-coated magnetic nanoparticles as a well-tolerated and promising tool for further surface modifications and stem cell -based therapies.

Magnetic particle imaging (MPI) allows quantitative evaluation of the spatial distribution of superparamagnetic iron oxide (SPIO) nanoparticles in the body. With a spatial resolution similar to magnetic resonance imaging (MRI), but superior temporal resolution, MPI has potential for different diagnostic applications. In addition to technical requirements, preclinical and clinical applications of this novel imaging modality require SPIO tracers optimized for MPI. This article discusses the suitability of Resovist as an MPI tracer and challenges and future prospects of tracer development.

Magnetic particle imaging (MPI) is a novel real-time imaging technique visualizing magnetic nanoparticles. Due to its intrinsic features it is especially suited for functional cardiac diagnosis, including angiography, cardiac wall motion assessment and quantitative myocardial perfusion imaging. In addition it may be suitable for cardiac intervention. MPI may reduce the overall diagnostic procedure duration and the complexity of recommended diagnostic pathways, thereby providing medical benefits to patients and economical benefits to hospitals / cardiologists due to the expected increased patient throughput.

Magnetic Particle Imaging (MPI) is a tomographic imaging technique, which relies on the nonlinearity of the magnetization curves of magnetic particles such as iron oxide nanoparticles and the fact that the particle magnetization saturates at some magnetic field strength (1). Sensitivity of MPI highly depends on the magnetic characteristics of used tracer. We have developed colloidal stable iron oxide nanoparticles with different sizes and coatings and optimized their magnetic properties using a variety of fractionation techniques to improve tracer sensitivity. MPI spectra were obtained on the various iron oxide nanoparticles to select the most sensitive tracer for plaque imaging in homozygous mice for the Apoe

tm1Unc

mutation. We conclude that iron oxide nanoparticles with appropriate magnetic properties are useful tracers for MPI.

In a conventional magnetic particle imaging (MPI) method, a reconstructed image is mainly calculated from odd harmonics of the magnetization response. However, the image blurring and artifacts appear due to the interference of the magnetization response generated from magnetic nanoparticles (MNPs) around the field free point (FFP), so that the image resolution is degraded. Therefore, we proposed a new image reconstruction method that was focused on the difference between a waveform of electromotive force generated from inside and outside of an object region. Then, a numerical analysis and a phantom experiment were performed in order to confirm validity of the proposal method, and an experiment system was constructed. As results, it was indicated that image artifacts was reduced by the proposal method. However, since the image blurring was seen on the reconstructed image, the image resolution needs to be improved.

Magnetic particle imaging (MPI) is an emerging medical imaging modality capable of high-sensitivity images with unprecedented contrast and without ionizing radiation [1]. Our laboratory previously developed the x-space theory for MPI, which describes MPI as a scanning process in the spatial domain [2,3]. X-space MPI is particularly critical as it permits real-time image reconstruction, orders of magnitude faster than the traditional harmonic space system matrix reconstruction methods. The x-space theory was derived assuming adiabatic and instantaneous alignment of ultra-small superparamagnetic iron oxide nanoparticles (USPIOs) with the applied magnetic field. However, in reality the magnetization lags behind the applied field due to relaxation. Here, we include relaxation in the x-space MPI theory and show that real-time reconstruction is still feasible even with relaxation effects.

MPI is intrinsically a linear and shift-invariant (LSI) system as described in x-space theory [3,4] assuming perfect signal acquisition and that the imaging equation can be written as a convolution in realspace. However, the received MPI signal is corrupted by a direct feedthrough signal from the excitation source at the drive frequency, which is 10

6

times larger than the nanoparticle signal. Hardware filtering is required to suppress the feedthroughsignalbutalso inevitably removes this frequency component from the particle signal. This signal loss in the received particle signalbreaks the LSI propertiesof x-space MPI. This study will investigate the impact of losing the fundamental frequency component in the image domain and propose an algorithm to recover the lost information. Finally, experimental results demonstrate that LSI propertiescan be restored after applying a simple and robust baseline recovery algorithm.

Magnetic particle imaging uses a field with a single field-free point for spatial encoding. Such a gradient field can be generated by a Maxwell coil pair consisting of two opposing coils driven by current flowing in converse directions. In order to sample the volume in-between the coils, the field-free point has to be moved through space. When keeping the gradient strength constant the electrical power loss of the Maxwell coil pair setup drastically increases when moving the field-free point off-center. In this paper a coil configuration is proposed, which consists of four coils and considerably reduces the electrical power loss for off-center field-free point generation.

Due to its ability for quantitative 3D real time imaging with high sensitivity and spatial resolution but without ionizing radiation and iodine-based contrast agents, Magnetic Particle Imaging shows great promise for the application to the image guidance of cardiovascular interventions. For this purpose, the blood in the vessels and the instruments would have to be visualized, e.g. using a SPIO-based contrast agent and a SPIO labeling, respectively (SPIO: superparamagnetic iron oxide). In a simulation study of this situation, simple models of a guide wire and a catheter with a coated tip as well as a filled balloon catheter have been examined under a variety of conditions. The appearance of the instruments in the reconstructed images has been shown to be strongly dependent on the imaging parameters (gradient strength), the difference of the SPIO concentrations in adjacent structures as well as the geometric extensions of the instrument and its position inside the vessel (partial volume effect). It has been demonstrated that the visualization of instruments in a vessel may be possible with positive or negative contrast, depending on the individual circumstances.

Magnetic Particle Imaging (MPI), a method that takes advantage of the non-linear magnetization curve of superparamagnetic iron oxide (SPIO) nanoparticles, promises to deliver high spatial and temporal resolution with a sensitivity exceeding that of magnetic resonance imaging (MRI). However, SPIO nanoparticles have a short blood retention time which limits the applicability of such compounds for MPI. We propose the use of red blood cells (RBCs) as carriers of SPIO nanoparticles to realize a blood pool tracer with longer blood retention time. Previously, we described a method of SPIO nanoparticle encapsulation into RBCs. The loading procedure consists of a hypotonic dialysis of cells in the presence of magnetic nanoparticles and successive resealing and reannealing of cells using isotonic solutions. Here, we report for the first time Magnetic Particle Spectroscopy (MPS) and MPI results obtained after intravenous administration of murine Resovist-loaded RBCs in an

in

vivo

MPI experiment.

A new method for selecting sentinel lymph nodes (SNs) in colorectal cancer tissue was investigated in 12 patients. A tracer consisting of superparamagnetic ironoxide (SPIO) nanoparticles was injected in the resected tissue. A handheld magnetic probe was used to select SNs to which the SPIO was drained. Vibrating sample magnetometry was performed on the lymph nodes to quantify the amount of SPIO in the nodes. High-field MRI allowed to depict the distribution of SPIO in the node, and revealed small anatomical structures. One or more SPIO containing nodes were successfully selected with the magnetic probe in all 12 patients.

Breast cancer diagnostic and treatment consists of surgical tumor removal and axillary lymph node resection. Radical axillary lymph node removemend is associated with high morbidity and significant loss of quality of life. The concept of sentinel lymph node biopsy (SNLB) by the use of dye and radio nuclides strongly reduced those side effects. To further reduce the side effects when axillary lymph nodes are removed, super paramagnetic iron oxide nano particles (SPIOs) could replace these marker substances. The magnetic particle imaging (MPI)-procedure will be used to visualize these SPIOs. Intraoperative three-dimensional MPI imaging and distinct localization probably by the use of a MPI hand probe will facilitate the axillary SNL detection and moreover makes it more precise. A mouse model was applied to prove the mentioned principle of SNLB by MPI. We are presenting first results of this approach and, additionally the qualitative and semi-quantitative distribution of SPIOs in lymph-fat tissue is shown for the first time. SPIOs are moving from the injection site through the lymph-fat tissue to the axillary region and finally into the axillary lymph nodes. This was approved by histology and prussian blue iron staining of the slides, electron transmission microscopy and in vivo magnetic resonance imaging. The concept of SNLB by MPI can be applied in principle in all solid tumors.

In order to improve the specificity of chemotherapeutic drugs towards pathological tissue, we investigated minimally invasive delivery methods by simulation of a magnetic targeting system. This system aims at the concentration of superparamagnetic nanoparticles at a tumor site in the body under the influence of external magnetic forces after injection of the particles into the circulatory system. Therefore, the properties of differently synthesized superparamagnetic iron oxides (SPIOs) were analyzed and implemented in a simulation model. FEM simulations were performed using the Navier-Stokes equation of fluid motion, which describes the hydrodynamic forces that act on the SPIOs in blood flow, with an additional magnetic term caused by the interaction between the SPIOs and the external magnetic field. As a result, we could show the feasibility of magnetic targeting by combining the optimization of both the magnetic fields and the SPIOs’ properties.

The unfavorable application-to-tumor-dose-ratio is a drawback of conventional systemic chemotherapy, implying an often insufficient drug dose in the tumor being associated with severe side effects for the patient. The use of chemotherapeutics bound to magnetic nanoparticles offers several advantages. On the one hand it is possible to concentrate the chemotherapeutics in the tumor region by the use of magnetic fields, like it is done in Magnetic Drug Targeting (MDT). On the other hand magnetic particles can serve as contrast agent for magnetic resonance imaging (MRI) that is bound to the therapeutics. Hence, the particles possibly are opening an insight into drug distribution in the tumor region directly after administration.

Another important factor for a successful MDT-application is detailed knowledge about tumor vascularization.

In this pilot study we investigated vascularization and size of an tumor in an experimental

in vivo

tumor model via flat-panel angiography and DYNA-CT before MDT and the particle distribution with MRI after MDT.

We could show that the tumor could be displayed by MRI and DYNA-CT before and after MDT. Flat panel angiography revealed clearly the pathological tumor vascularization before MDT, while MRI imaging afterwards displayed the tumor as well as the particle distribution in the tumor.

Cancers of the gastrointestinal tract are one of the leading causes of death in affluent countries. Although they are usually curable if detected at an early enough stage, the compliance to undergo routine screenings for gastrointestinal cancers in a presymptomatic stage is low, mostly due to the discomfort which current diagnostic measures, such as gastroscopy, colonoscopy, digital rectal exams or feces examinations pose to the patient. Magnetic particle imaging (MPI) has the potential to overcome this hurdle as the procedure is noninvasive, radiation-free and assumed to be more sensitive than most other imaging modalities used today. Yet, MPI bears a catch as it requires special magnetizable contrast agents in the nanoscale range. These contrast agents have to meet a number of requirements. They must be detectable in the magnetic field, be manufacturable in a cost-effective manner, display a certain shelf life, be administrable in a compliant way, be stable in vivo, be specific and sensitive in labeling the neoplastic tissue and must not be harmful to the individual who incorporates them. Due to this complex requirement profile MPI contrast agent engineering lags behind imaging hardware development. In order to close in on this field it is important to identify, rate and define the relevant players which ingested MPI contrast agents have to deal with en route to their destination in the body. Potential players are the anatomy and microanatomy of the gastrointestinal tract, the chemical and biological composition of the luminal constituents as well as potential degradation and scavenger mechanisms. Once the possible impact of these factors has been established contrast agents can be designed to master these challenges. Yet, reaching the destination without causing harm is a necessary but not sufficient requirement. In order to fully exploit the potential of MPI and to qualify it for mass screening the method must be highly sensitive and specific. This may be achieved by targeting systems that direct the contrast agent exclusively and efficiently to the neoplastic tissue in question. In this paper we provide an overview on our current knowledge of the fate of magnetizable contrast agents in the gut, the players which determine this fate, measures to properly endow contrast agents for this task and means to equip them with the necessary specifity and sensitivity.

Due to the possibility of high temporal and spatial resolution, high sensitivity and three-dimensional imaging, Magnetic Particle Imaging is a promising new imaging approach for interventional cardiovascular procedures. In this contribution we present first MPI-images of a specifically labeled, though commercially available device for cardiovascular intervention. Furthermore, different approaches to label those instruments are discussed.

Screening for early ovarian cancer is critical because it would increase survival dramatically. Ovarian cancer is cured at very high rates when found early but is almost never found until it has metastasized and the survival rate is very low. Ovarian cancer screening might be possible using immunologically targeted magnetic nanoparticles (mNP). mNP injected into the peritoneum are delivered to the cancer by phagocytic cells using the bodies own immune system. Magnetic detection of the mNP does not require ionizing radiation and can be made at low mNP concentrations. Screening only requires the number of nanoparticles in each ovary so minimal localization is required: only two pixels. We present simulation results showing that spatial encoding for two pixels can be achieved using many sets of harmonics ranging from only two harmonics to essentially all the harmonics. However, the highest SNR measurements are achieved using only two harmonics that approximate a wavelet basis.

The use of a field free line in magnetic particle imaging promises a considerable increase in sensitivity of this new imaging technique. Furthermore, efficient reconstruction algorithms emerge for line acquisition in MPI, since it is possible to transform the data into Radon space. This allows for the use of well-known and powerful reconstruction algorithms like for instance the filtered backprojection, which will substantially speed up reconstruction time. In this work, a simulation study is presented, which analyzes the influence of magnetic field optimization on the image quality achieved for efficient Radon-based reconstruction.

M

agnetic

P

article

I

maging is based on the nonlinear response of ferro- and superparamagnetic particles to magnetic fields [1]. For imaging, a field free point (FFP) within a string magnetic gradient on the order of 1–5 T/m is moved through the sample. A new gradient system design allows performing dynamic imaging in a linear sampling scheme by using a traveling wave approach [2]. We present an extension for doing 3D imaging using a traveling wave in combination with frequency mixing [3] and a sliced field of view (FoV). This approach provides the possibility of an arbitrarily large FoV in one direction without increasing the specific absorption rate (SAR) and allows the spatial encoding in the additional 2 dimensions.

Magnetic Particle Imaging (MPI) is a very recent medical imaging technique providing tomographic data avoiding use of ionizing radiation. The first MPI scanner presented by Gleich and Weizenecker has a closed geometry which has to fit the object of interest [1]. In order to be able to examine larger objects, Sattel et al. developed a new coil configuration, the single-sided MPI scanner geometry [2]. The detection of axillary sentinel lymph nodes is one medical application scenario. MPI improves the surgical procedure by real-time 3D image guidance and may contribute towards reducing cost and the time needs per patient. This contribution presents improvements of various coil topologies. Furthermore, the cooling system is optimized and the send and receive chain will be improved.

Tomographic imaging using a shifted and rotated field free line (FFL) with filtered backprojection image reconstruction can approach an order of magnitude SNR improvement over a field free point (FFP) given equal scan time. In this paper, we demonstrate a projection reconstruction x-space imager. The imager consists of a 2.4 T/m permanent magnet FFL gradient, a Helmholtz pair of off-the-shelf electromagnets, a solenoidal transmit coil and a gradiometer receive coil. A motor driven rotary table rotates the sample and the system acquires multiple projection images at evenly spaced angles between zero degrees and 180 degrees. Filtered back-projection is used to reconstruct a three-dimensional tomographic image stack. Sample rotation, which is sometimes employed in commercial mouse CT scanners, has been used to test this method. Later systems may rotate the gradient similar to a human-sized CT gantry or may generate an electronically rotated FFL gradient. In previous work, we have shown an MPI capable FFL scanner. Here, we show 3D experimental results of our PR-MPI scanner using acrylic USPIO imaging phantoms and post-mortem mice.

Magnetic Particle Imaging evolves rapidly and human scanners are conceivable, already. However, the growing scanner size and therefore the increasing data within the field of view give rise to several unsolved problems. The reconstruction process, solving an inverse problem with the measured signal and the system function, is a storage consuming procedure for high resolution 3D imaging. Additionally, the size of the field of view strongly depends on the used gradient field and field amplitudes. Due to technical as well as medical limitations, such as specific absorption rates and peripheral nerve stimulation, the conventional procedures will not be sufficient to image large regions of interest. This paper compares and discusses approaches enlarging the field of view that might be used to reduce the reconstruction process and/or enlarge the field of view despite limited technical properties.

The imaging volume that is rapidly encoded by drive fields in 3D magnetic particle imaging is limited by power dissipation and nerve stimulation thresholds. Additional coils have been implemented to generate so-called focus fields that operate at lower frequencies and extend the accessible imaging range. This contribution presents the possibility of sweeping the rapidly encoded imaging volume along an arbitrary 3D path using continuous focus field variations. This technique can be useful for following a tracer bolus, for tracking devices, or for dynamically moving the image focus to different regions of interest.

Here we describe the construction of our third generation x-space MPI scanner, the fifth MPI scanner built at UC Berkeley. The scanner has two goals, (1) High-resolution native MPI resolution using x-space reconstruction, and (2) extended FOV suitable for mice and rats. In this paper we describe our design criteria, and we show the initial characterizations of the 7 T/m gradient field.

Here we describe the construction and images of the first projection x-space MPI scanner. The scanner is a side-access quadrupole design, and generates a 2.35 T/m main field gradient. The system excites the sample and receives signal in one axis, and reconstructs full-body images using x-space reconstruction. The resulting images are of high quality, and we demonstrate linear and shift invariance of the imaging system by imaging a resolution phantom and mice injected with Resovist.

A new measurement principle based on the frequency mixing technique for investigating the shape of the magnetization curve of soft non-hysteretic magnetic materials is introduced. Based on Taylor expansion of the magnetization curve and spectral investigation of an inductively detected signal, a mathematical model for the reconstruction of M(H) is proposed, [7]. Here, the model is experimentally verified using a nanocrystalline soft magnetic material with defined properties. It is shown that the magnetization curve can be reconstructed very accurately and the influence of an additional parameter, i.e. strain, can be investigated in detail as well.

Using microscopically models and Green’s function techniques we demonstrate how one can get information of ferroelectric nanoparticles. The approach can be extended to multiferroic systems which are defined as materials possessing two or more ferroic orders in a single phase. In detail we show that the unexpected ferromagnetic properties of BaTiO

3

(BTO) observed recently at room temperatures are due to oxygen vacancies at the surface of the nanocrystalline materials. Such vacancies lead to the appearance of Ti

3 + 

or Ti

2 + 

ions with nonzero net spin. The resulting different valence offers a nonzero magnetization which decreases with increasing particle size. The system shows a multiferroic behavior below a critical size of the nanoparticles and the related polarization tends to a saturation value when the particle size is enhanced.

Starting from the basic principles of a Magnetic Particle Spectrometer (MPS), this paper explains the benefits and limitations of conventional spectral representation of the magnetization behavior of magnetic particles. After motivating the advantages of direct

m

(

H

) representation for particle analysis, it is shown how this curve, or at least its derivative, which is related to the point spread function, can be derived from the measured data. To illustrate, MPS results for Resovist® are presented, that show experimental evidence of hysteresis in dynamically acquired

m

(

H

) curves.

For small excitation fields in the microtesla (

μ

T) range, the dependency of the magnetic moment of magnetic iron oxide nanoparticles (MNP) on the external field can be regarded as linear. Sensitive superconducting quantum interference devices (SQUIDs) enable the detection of the response of MNP in biological tissue in the pT range. The co-registration of the excitation field is reduced by appropriate geometrical configuration of excitation coil and sensor coil. MNPs in a wide range of mean diameter and distribution parameters can be used for signal generation. The spatial distribution of MNP is reconstructed using data from a parallel multi-sensor and sequential multi-coil arrangement and applying linear estimation techniques. The time delayed response of MNP due to Brownian and Néel relaxation processes represents a specific signal not being influenced by the diamagnetic contribution of water in the tissue. We present the theoretical background and measurement data from different setups that will exemplify the concept.

Multi-channel magnetorelaxometry (MRX) is demonstrated to locate and quantify several foci of a magnetic nanoparticle (MNP) distribution after magnetic drug targeting (MDT) in an in-vivo rabbit carcinoma model. By this non-invasively technique MNP accumulations of lateral extensions up to 20 x 30 cm

2

can easily be accessed in biological tissue. The total measurement duration of about 20 min including preparation of the rabbit and measurement of two regions of interests (ROI), tumor and thorax (liver, lung and spleen), enables a high throughput of animals. Modelling the MNP distribution either by magnetic point dipoles (ROI tumor) or by an extended, homogenously magnetized body (ROI thorax) of simple geometry, e.g. cuboid, the total magnetic moment and location of each focus is determined by minimum norm estimation. Simultaneously, vital functions like respiration and heart activity can be monitored directly and non-invasively during the MRX measurements. Thus, our MRX procedure gains valuable information for monitoring of MDT procedures in large scale animal models.

Magnetic Particle Imaging (MPI) is a quite new imaging technique, based on the non-linear magnetization behavior of superparamagnetic iron oxide nanoparticles (SPIOs). These SPIOs are applied to a patient as tracer material. An important key aspect for successful and reliable imaging is tracer development and characterization. Hence, a Magnetic Particle Spectrometer (MPS) is needed to perform extensive studies about the SPIOs dynamic magnetization behavior. The MPS hardware build-up consists mainly of a drive coil, a receive coil, a cooling circuit, amplifiers and a control unit. This paper proposes a new self-contained control unit hardware system based on a System-On-Chip (SoC) Field-Programmable- Gate-Array (FPGA). The control unit controls, collects, and maintains data acquisition, coil excitations and the behavior of the cooling circuit in realtime. The SoC is also able to pre-process measurement data (i.e. filtering, averaging, up- and down-sampling etc.). The generation of the magnetic field in the drive coil is controlled by a 14 bit Digital/Analog Converter (DAC) coupled with an external amplifier. The receive coil signal is amplified and digitalized via a 14 bit Analog/Digital Converter (ADC). To prevent system overheating, also temperature is supervised and kept under certain limits. For that purpose four Resistive Thermal Devices (RTDs) in combination with a cooling unit are utilized. The transmission of measurement data is realized via USB 2.0 with up to 40 MByte/s. The USB link can also be used to setup the control unit and as power supply for the self-containing SoC.

Magnetic Particle Imaging in general applies two different types of magnetic field geometries in order to obtain information about the spatial distribution of magnetic nanoparticles. First, a static gradient field called selection field is used for spatial encoding. It either provides a field free point or a field free line. Second, additional uniform drive fields are superimposed to steer the field free region through the field of view. Most efficiently, the latter fields are orientated perpendicularly to each other. For field generation, current carrying coils are used, while the static field generation may be supported by strong permanent magnets. In this contribution, discs carrying circular current distributions are investigated for generating both types of field geometries. These distributions are optimized with respect to the achieved field quality as well as to the total power loss.

The signal-to-noise ratio in magnetic particle imaging can be limited by distortion interference arising in the imaging system. In this work, we investigate the contribution of resonant transmit capacitors to system interference. Feedthrough interference spectra obtained using four capacitors with varying voltage ratings in a custom MPI interference testbed show a 20dB reduction in distortion interference with higherrated capacitors. Finally, we discuss the applicability of the interference testbed to treat other interference mechanisms in magnetic particle imaging.

Current small-animal-sized MPI scanners operate at 1-25 kHz frequency range with 0.1-20 mT peak amplitude, neither of which is optimized. SAR and especially dB/dt safety limits will determine the optimum excitation field, and will impact the optimum scanning speed, field-of-view (FOV) and signal-to-noise ratio (SNR). In this work, we describe the first human-subject safety limit experiments for MPI. Our results indicate that the magnetostimulation threshold monotonically decreases with increasing frequency and is inversely correlated to the body-part size.

Magnetic particle imaging (MPI) is a new imaging modality, which allows for the determination of the distribution of super paramagnetic nanoparticles in vivo with excellent contrast, penetration and high temporal resolution. So far, real-time imaging in a mouse has been realized using a scanning-system with a field free point (FFP). Recently, an alternative encoding scheme has been developed promising faster scanning times and a higher sensitivity. This can be handled by extending the FFP to a field free line (FFL). Preliminary scans of phantoms showed the feasibility of the FFL in practice, based on projection x-space MPI. To ensure the safety of imaging switching from phantom to

in vivo

scans, a specific protocol has to be provided for scanning live animals. This paper describes the construction and testing of a mouse bed for heating as well as delivery and recovery of anesthesia gases. The mouse bed is constructed with non-magnetic materials and sized to the specific scanner. The size of the bed is limited by the diameter of the bore, and a larger bed (and bore) would be required for larger animals.

We designed a mouse bed fulfilling all mentioned requirements by engineering a specialized water warming system and a modified connection for the anesthesia system. The whole bed is moveable and rotatable in and around the longitudinal axis of the bore by jointing it to a robot. Rotation is critical for performing volumetric 3D MPI with projection reconstruction (or Radon) computerized tomography.

Due to forthcoming use of MPI on humans there is an urgent need for a thorough research on possible adverse effects of this technique on patients’ health. However, the health impact of exposure to time-varying magnetic fields in a frequency range between 10 kHz and 100 MHz, such as the MPI drive field, are still poorly investigated.

The current paper intends to give an overview on an in-silico approach to investigation of stimulating effects that could be caused by the MPI drive field. For this purpose, cell models of myocardiocyte, myocyte and neurocyte, as well as a suitable setup for the simulation of the exposure to time-varying magnetic fields have been developed. The evaluation of performed simulations was carried out on the basis of transmembrane voltage elevation and induced current densities.

This paper studies the performance of a modular class-D switching amplifier to supply the drive field coils of a magnetic particle imaging scanner with a high quality sinusoidal voltage. While a class-A or class-AB amplifier is capable of delivering such a sinusoidal voltage, its low efficiency does not qualify for an economic solution. Therefore, a class-D amplifier has been used. The modular amplifier consists of commercially available class-D amplifier chips, which are connected in parallel, to achieve the required output power. This paper describes the simulation model and a hardware setup of the modular amplifier which consists of 10 chips connected in parallel.

In this paper a hybrid filter topology for the reduction of the total harmonic distortion in AC sources is introduced. The hybrid filter contains active and passive components. The structure of the digital filter influencing the signals gain and phase will be explained. Furthermore an analytical consideration of optimal filter coefficients is shown. For the validation of the concept a matlab/plecs simulation is presented.

Magnetic Particle Imaging is a promising new imaging technique using magnetic fields to image magnetic tracer material in the body. As with MRI systems, time varying magnetic fields raise some safety issues. The stimulation of peripheral nerves and tissues is one of them. In the paper, the stimulation thresholds are explained and an evaluation of the stimulation generated by a pre-clinical scanner is calculated. It appears clearly that, even if driving fields of high amplitude are used, cardiac arrhythmias are unlikely to be induced. However, it is yet unclear whether some peripheral nerve stimulation may be induced.

Motivation:

Systemic transplantation of human mesenchymal stem- and progenitor cells (MSC) is a promising approach in regenerative medicine. Magnetic resonance imaging (MRI) of transplanted MSC labeled with superparamagnetic iron oxide nanoparticles (SPION) is an excellent tool for

in vivo

cell tracking because it offers a highspatio-temporal, sensitive and non-invasive cell detection. However, cell labeling with commercial SPION is still inefficient and requires the use of potentially toxic transfection agents. New results are reported here about magnetite@citrate ferrofluids that allow highly efficient magnetic MSC labeling without transfection agents.

Superparamagnetic iron oxide nanoparticles (NPs) with appropriate surface coating are widely used for numerous

in vivo

applications and in particular for MRI contrast enhancement. To improve the contrast enhancement and targeting properties as well as the biodistribution of functionalized iron oxide NPs, challenges have to be overcome such as:

(i)

the design of an organic coating favouring ideal biodistribution, ensuring multifunctionalization (targeting, optical imaging…) and preserving a small size distribution of coated NPs in physiological media (<50 nm),

(ii)

the synthesis of iron oxide NPs with good magnetic properties and

(iii)

the development of strong anchors at the NPs surface avoiding desorption of molecules in blood. Indeed, the coating design as well as the interaction nature between the organic shell and the nanocrystal surface are more and more key points to address.

Head and neck cancer includes the squamous cell carcinomas of the oral cavity, pharynx and larynx and is one of the most common solid neoplasms worldwide. The malignancy is an important public health problem worldwide with more than 500000 new cases diagnosed each year. The change in survival over the last 20 years remains minimal and despite recent attention, the mortality rates are still high due to local tumor invasion and to a high predilection for the development of relapses and metastases. The cellular and molecular mechanisms responsible for tumor aggressiveness and its response to chemo- and radiation therapies remain mostly unknown.

The detection of neuronal currents (NC) may foster the understanding of the flow of information in the brain. Existing methods of NC detection like electro- and magnetoencephalography (EEG/MEG) are limited by the ambiguity of the inverse problem, while imaging methods like fMRI monitor secondary effects (e.g., blood oxygenation changes). On the other hand, attempts to directly detect NC in high field NMR techniques by a local shift of the

1

H resonance frequency have yielded controversial results. In magnetic fields around 1 microTesla, the relative contribution of the biomagnetic field generated by neuronal currents is orders of magnitude higher than in conventional high fields. This may make the situation much more favourable.

In the endeavor to perform in vivo magnetic resonance imaging in very low fields, we developed a dedicated SQUID based NMR/MRI measurement system. The low noise performance of < 30 fT/

$\surd$

Hz above 1 Hz enables the measurement of nuclear magnetic precession at magnetic fields well below 50

μ

T down to 100 nT. The system is operated inside a heavily magnetically shielding - the Berlin Magnetically Shielded Room BMSR-2.

Each measurement starts with a prepolarization of the sample in a field of up to 5 mT. After this preparation, the free precession decay in the much weaker detection field is measured by the SQUID.

Magnetic particle imaging (MPI) is an emerging magnetic nanoparticle detection technique that has great potential as a novel biomedical imaging procedure. Particularly, MPI offers a safer real-time option over conventional x-ray angiography procedures since it uses safe magnetic fields (no ionizing radiation) and biocompatible superparamagnetic magnetite (Fe

3

O

4

) nanoparticle tracers, which are the source of the signal and play a significant role in spatial resolution. Current tracer formulations such as Resovist® offer poor spatial resolution, and thus, inadequate performance for high-quality angiographies. Alternatively, our superparamagnetic magnetite (SuperMag) tracers show 30% improvement in spatial resolution compared to Resovist®. However, an ideal MPI tracer consists of a balance between an optimized magnetic core and a biocompatible shell that enhances circulation times combined with appropriate functionalization necessary to enhance the tracer’s bioavailability. For angiographies, tracer availability in the vasculature is of utmost importance to determine the most effective method of administration and ensure sufficient time for the imaging procedure. In this preliminary study we report pharmacokinetics and biodistribution characteristics of SuperMag tracers in an animal model. SuperMag tracers were formulated with variations in the polymeric shell and subsequently tested in CD-1 mice. Dose-dependent biodistribution was studied using MR-imaging and post-mortem histology analysis. Implications of

in vivo

circulation characteristics on MPI angiography procedures are discussed.

Magnetite nanoparticle (MNP) tracers can be optimized for Magnetic Particle Imaging (MPI) by tuning their magnetic core size and size distribution. In previous work, our group showed that ~15 nm MNPs were optimal at 250 khz by synthesizing a series of tracers with tuned size and narrow size distributions. This optimization approach is general, and here we present experimental results that demonstrate that the same principles apply at 25 khz excitation. The key result of this work is that MPI spatial resolution can be improved by the appropriate selection of MNPs.

For several medical applications of magnetic nanoparticles (MNP) it is desired to know the quantity and characteristics of the particles in the tissue of interest. That can either be necessary to determine how successful a procedure was of how it will be. Therefore a system is built, that is suitable to analyze small intact biological samples at room temperature. The magnetometer is used for the analysis and selection of sentinel lymph nodes in colorectal cancer. In this clinical procedure MNPs are administered in the resected part of the colon to determine the sentinel lymph node. The magnetometer is based on copper wound coils and comprises of two detection coils in series opposition, enclosed by two separately driven excitation coils.

Here we present an aqueous route for the synthesis of uniform magnetite nanoparticles with sizes around the monodomain diameter (20-100 nm). The method is based on the precipitation of an Fe(II) salt in a mild oxidant in hydroalcoholic solutions, and leads to highly uniform and crystalline magnetic nanoparticles in a single step. Colloidal suspensions of these particles were directly obtained by simple ultrasonic treatment of the powders thanks to the presence of sulphate anions at the particle surface. All magnetisation curves saturate at much lower magnetic fields and show larger saturation magnetization than samples prepared by coprecipitation. Saturation magnetisation values vary between 83 and 92 emu g-1, close to the theoretical values reported for bulk magnetite at room temperature. Those magnetic particles have shown the maximum heat efficiency for magnetic hyperthermia and they could show great potential in MPI imaging.