What Is Magnetic Particle Imaging?

Magnetic particle imaging (MPI) is a tomographic imaging modality that directly detects the three-dimensional spatial distribution of superparamagnetic iron oxide nanoparticle tracers administered to a subject, without using ionizing radiation. Unlike X-ray computed tomography or positron emission tomography, MPI generates signal only from the tracer particles themselves rather than from tissue attenuation or radioisotope decay. This eliminates background signal from surrounding anatomy and makes the measured signal quantitatively proportional to local tracer concentration.

The technique was proposed in 2005 by Bernhard Gleich and Jürgen Weizenecker at Philips Research Laboratories in Hamburg, whose original Nature paper on tomographic imaging using the nonlinear response of magnetic particles established the feasibility of the approach. It draws on principles from magnetic resonance imaging, signal processing, and nanoparticle physics.

Tracer Physics and Signal Generation

MPI exploits the nonlinear magnetization curve of superparamagnetic iron oxide nanoparticles (SPIONs). When a sinusoidal excitation field is applied, SPIONs in an unsaturated region respond with a nonlinear magnetization that generates harmonic frequencies beyond the fundamental drive frequency. These harmonics are detectable by a receive coil tuned away from the drive frequency, effectively rejecting the much larger excitation signal.

Spatial localization is achieved with a gradient field that creates a field-free point (or field-free line) at the center of the imaging volume, where the net static field is zero. Only particles at that point remain unsaturated and contribute harmonics; particles elsewhere are magnetically saturated and produce no detectable signal. Mechanically or electronically scanning the field-free point through the volume maps the tracer distribution. The tracers are referred to as such rather than as contrast agents because the particles, typically Resovist or purpose-designed SPIONs, constitute the only source of image signal.

System Architecture

An MPI scanner consists of three main field subsystems: a static gradient magnet that establishes the field-free point, a drive coil generating the oscillating excitation field (typically at 25 kHz), and a receive coil chain with narrow-band filtering to extract the harmonic response. A transmit-receive switch or physical separation between coils prevents the large excitation signal from saturating the receive chain.

Current preclinical systems achieve spatial resolution of approximately 1 mm with temporal resolution up to 40 volumes per second, enabling real-time imaging of tracer boluses moving through the vasculature. Scaling these specifications to human-sized scanners requires gradient fields strong enough to confine the field-free point to millimeter dimensions over a larger bore, which remains the primary engineering challenge. A 2015 review of magnetic particle imaging developments and future directions in the International Journal of Nanomedicine documents early system designs and the transition from bench-top demonstration toward clinical scale.

Image Reconstruction

Converting the received harmonic signals to a spatial tracer map requires solving an inverse problem. The standard approach uses a measured system function: a known point source of tracer is scanned through every voxel position to record how it contributes to each harmonic at each drive-field phase. This system matrix is then used with regularized inversion algorithms such as Kaczmarz iteration or Tikhonov regularization to reconstruct the tracer image from an unknown distribution. The system function approach captures real-world imperfections in particle response and field geometry, but acquiring it is time-consuming. Model-based reconstruction methods that parameterize the Langevin magnetization curve are an active alternative, as surveyed in research on signal encoding and system function properties in MPI.

Applications

Magnetic particle imaging has applications in a range of fields, including:

• Real-time cardiovascular imaging and angiography
• Interventional guidance for catheter placement
• Cell tracking to monitor stem cell or immune cell therapies
• Brain perfusion imaging and stroke assessment
• Cancer detection through accumulation of targeted tracers
• Preclinical research in small animal models