DIFFERENTIAL MAGNETOMETRY - SELECTIVE MAGNETOMETRY FOR IN VIVO USE
Sebastiaan Waanders is a PhD student in the department Magnetic Detection and Imaging. (Co)Supervisors are prof.dr.ir. B. ten Haken and dr.ir. L. Alic from the faculty of Science & Technology.
The implementation of magnetic detection techniques in clinical practice has long lagged behind developments in society as a whole. Whilst techniques like magnetic resonance imaging and (to a lesser extent) MEG/MCG have been successfully introduced, use of magnetic techniques for diagnostics and treatment has always been limited to well controlled, shielded environments. In this thesis, we describe the development of a sensitive and specific method to use magnetic nanoparticles for diagnostics and localization, without the need for shielded surroundings. This opens up a world of new applications for magnetic detection, ranging from intraoperative detection of sentinel lymph nodes to tracking and tracing of stem cell therapies, to navigation applications.
Differential Magnetometry
In chapter 2, we introduce the concept of Differential Magnetometry. By exciting a sample containing superparamagnetic iron oxide nanoparticles (SPIONs) with an induction coil generating an alternating magnetic field, we magnetize the particles. This magnetization is measured by a pair of detection coils. If we now periodically add a static magnetic field to the excitation, we alter the magnetic response of the particles, because the magnetization dynamics of these particles are strongly nonlinear. By determining the difference between these two magnetization signals, we obtain a specific particle signal, which contrasts with the predominantly linear magnetic properties of tissue and surgical instruments, in this field range. In this chapter, we illustrate how the time derivative of the magnetization changes as the magnetization is pushed towards saturation by the applied DC offset field.
The principle of Differential Magnetometry was implemented in a handheld probe, intended for intraoperative detection of the sentinel lymph node during breast cancer surgery. This application forms a formidable challenge for magnetic techniques, as demands on sensitivity and selectivity are high, and the injected dose needs to be as low as possible to prevent post surgical sideeffects like skin staining and MRI artefacts. Additionally, emitted power of the coils is limited to prevent tissue heatup and other effects like described in the ICNIRP guidelines on non ionizing radiation protection. Whilst conventional magnetometers, explored for this type of application, have issues balancing out the effects of a varying background susceptibility (due to different tissue oxygenation, which is slightly paramagnetic, for example), and metallic surgical instruments, DiffMag is robust against these artifacts, provided the measurement electronics are fast enough to compensate in real time for the changing induced signal. In this chapter, we outlined the clinical requirements for this application, and how they translate into a first prototype probe and electronics combination. The probe consists of coaxial pairs of excitation and detection coils, with a few windings of compensation coil wound around the excitation coil for real-time adjustment of the excitation parameters. The excitation coils are driven by a current controlled power amplifier, which is controlled by MATLAB through a high-speed DAQ system. Following the detection coils, a lownoise preamplifier forms the final element of the probe. An instrumentation amplifier takes the measured signal and feeds it into the processing DAQ. The DiffMag signal analysis and processing are all performed in MATLAB. All characterization measurements were performed us- ing samples of the well known MRI contrast agent Resovist. Probe operation was evaluated by measuring its dose-response curve, lateral and depth sensitivity. Furthermore, we show initial results of the active compensation mode, which strongly suppresses the signal from surgical instruments placed near the detector.
Separation of excitation and detection coils
Then, we look into another challenge in chapter 3: can we downsize the detection system, such that it fits in a standard 6mm trocar opening, for laparoscopic detection of sentinel lymph nodes. The only way we can achieve this, and improve the maximum attainable detection depth, is by separating the excitation coil system from the detection coils. Usually, this is impossible, because the constantly varying mutual induction between the two coil sets is indistinguishable from the signal of interest, that of the particles. But because of the stepwise pulse sequence in DiffMag, we can allow for some measurement time to compensate for this changing mutual induction, as long as the rate of change of the mutual inductance is slow compared to the pulse sequence time.
Building on the DiffMag foundations laid in chapter 2, we elaborate on the phase sensitive detection scheme, and how it is used to calculate the optimal compensation signal which is coupled into the coils of the gradiometer. We then illustrate the physical setup, and the expanded electronics package, as the induction of the excitation coils is now such that two separate power amplifiers are required, one for the AC signal and one for the DC pulses. Following this description, we show the first results of the active, iterative compensation and finish by discussing improvements that are needed before this setup is ready for (pre)clinical evaluation in a controlled setting.
Modelling magnetic nanoparticles
When looking for an optimal combination of magnetic probes and tracers, it is essential to understand the physics underlying the tracer’s behavior in the relevant field regime. These dynamics are primarily governed by the time-dependent evolution of sample’s magnetization due to externally applied magnetic fields. Especially in the magnetic field and frequency range where Néel and Brownian relaxation mechanisms compete, a complete understanding of the physics in- volved is still lacking. In this chapter, we explored a model in which we propose a very basic coupling of both magnetization channels, through the effect that magnetic anisotropy has on the Néel relaxation mechanism. The basic idea is that Brownian relaxation, even in the situation where Néel relaxation dominates, acts to align the magnetic easy axis with the applied field, which shortens the Néel relaxation time. We describe this model in chapter 4.
We explore the feasibility of this model by numerically evaluating the Fokker-Planck equations that describe both relaxation mechanisms, whilst taking into account the particle size and anisotropy distribution. The model was evaluated for three different particles, with overall good agreement between measurement and simulation.
Magnetic measurements were performed using the in house built SuperParamagnetic Quantifier, which is a coaxial magnetometry setup similar to those used in AC magnetometry experiments, but using the DiffMag excitation and detection scheme. In the chapter, we describe experimental parameters used in the experiment, and characterization of the samples, to which we added TEM images for size distribution measurements.
Finally, we combine the discussion points from the individual chapters, and have a short outlook on the potential of DiffMag for clinical applications, and illustrate the steps that need to be taken before (pre)clinical evaluation can take place.
