RESUMEN
BACKGROUND: Magnetic nanoparticles (MNPs) are used as tracers without ionizing radiation in vascular imaging, molecular imaging, and neuroimaging. The relaxation mechanisms of magnetization in response to excitation magnetic fields are important features of MNPs. The basic relaxation mechanisms include internal rotation (Néel relaxation) and external physical rotation (Brownian relaxation). Accurate measurement of these relaxation times may provide high sensitivity for predicting MNP types and viscosity-based hydrodynamic states. It is challenging to separately measure the Néel and Brownian relaxation components using sinusoidal excitation in conventional MPI. PURPOSE: We developed a multi-exponential relaxation spectral analysis method to separately measure the Néel and Brownian relaxation times in the magnetization recovery process in pulsed vascular MPI. METHODS: Synomag-D samples with different viscosities were excited using pulsed excitation in a trapezoidal-waveform relaxometer. The samples were excited at different field amplitudes ranging from 0.5 to 10 mT at intervals of 0.5 mT. The inverse Laplace transform-based spectral analysis of the relaxation-induced decay signal in the field-flat phase was performed by using PDCO, a primal-dual interior method for convex objectives. Néel and Brownian relaxation peaks were elucidated and measured on samples with various glycerol and gelatin concentrations. The sensitivity of viscosity prediction of the decoupled relaxation times was evaluated. A digital vascular phantom was designed to mimic a plaque with viscous MNPs and a catheter with immobilized MNPs. Spectral imaging of the digital vascular phantom was simulated by combining a field-free point with homogeneous pulsed excitation. The relationship between the Brownian relaxation time from different tissues and the number of periods for signal averages was evaluated for a scan time estimation in the simulation. RESULTS: The relaxation spectra of synomag-D samples with different viscosity levels exhibited two relaxation time peaks. The Brownian relaxation time had a positive linear relationship with the viscosity in the range 0.9 to 3.2 mPa · s. When the viscosity was >3.2 mPa · s, the Brownian relaxation time saturated and did not change with increasing viscosity. The Néel relaxation time decreased slightly with an increase in the viscosity. The Néel relaxation time exhibited a similar saturation effect when the viscosity level was >3.2 mPa · s for all field amplitudes. The sensitivity of the Brownian relaxation time increased with the field amplitude and was maximized at approximately 4.5 mT. The plaque and catheter regions were differentiated from the vessel region in the simulated Brownian relaxation time map. The simulation results show that the Néel relaxation time was 8.33±0.09 µs in the plaque region, 8.30±0.08 µs in the catheter region, and 8.46±0.11 µs in the vessel region. The Brownian relaxation time was 36.60±2.31 µs in the plaque region, 30.17±1.24 µs in the catheter region, and 31.21±1.53 µs in the vessel region. If we used 20 excitation periods for image acquisition in the simulation, the total scan time of the digital phantom was approximately 100 s. CONCLUSION: Quantitative assessment of the Néel and Brownian relaxation times through inverse Laplace transform-based spectral analysis in pulsed excitation, highlighting their potential for use in multi-contrast vascular MPI.
Asunto(s)
Campos Magnéticos , Tomografía Computarizada por Rayos XRESUMEN
PURPOSE: Magnetic particle imaging (MPI) is emerging as a highly promising imaging modality. Magnetic nanoparticles (MNPs) are used as imaging tracers in MPI, and their relaxation behavior provides the foundation for its functional imaging capability. Since MNPs are also utilized in magnetic fluid hyperthermia (MFH) and MPI enables localized MFH, temperature mapping arises as an important application area of MPI. To achieve accurate temperature estimations, however, one must also take into account the confounding effects of viscosity on the MPI signal. In this work, we analyze the effects of temperature and viscosity on MNP relaxation and determine temperature and viscosity sensitivities of relaxation time constant estimations via TAURUS (TAU estimation via Recovery of Underlying mirror Symmetry) at a wide range of operating points to empower simultaneous mapping of these two parameters. METHODS: A total of 15 samples were prepared to reach four target viscosity levels (0.9-3.6 mPa · $\cdot$ s) at five different temperatures (25-45 ∘ $^\circ$ C). Experiments were performed on a magnetic particle spectrometer (MPS) setup at 60 different operating points at drive field amplitudes ranging between 5 and 25 mT and frequencies ranging between 1 and 7 kHz. To enable these extensive experiments, an in-house arbitrary-waveform MPS setup with temperature-controlled heating capability was developed. The operating points were divided into four groups with comparable signal levels to maximize signal gain during rapid signal acquisition. The relaxation time constants were estimated via TAURUS, by restoring the underlying mirror symmetry property of the positive and negative half cycles of the time-domain MNP response. The relative time constants with respect to the drive field period, τ Ì $\hat{\tau }$ , were computed to enable quantitative comparison across different operating points. At each operating point, a linear fit was performed to τ Ì $\hat{\tau }$ as a function of each functional parameter (i.e., temperature or viscosity). The slopes of these linear fits were utilized to compute the temperature and viscosity sensitivities of TAURUS. RESULTS: Except for outlier behaviors at 1 kHz, the following global trends were observed: τ Ì $\hat{\tau }$ decreases with drive field amplitude, increases with drive field frequency, decreases with temperature, and increases with viscosity. The temperature sensitivity varies slowly across the operating points and reaches a maximum value of 1.18%/ ∘ $^\circ$ C. In contrast, viscosity sensitivity is high at low frequencies around 1 kHz with a maximum value of 13.4%/(mPa · $\cdot$ s) but rapidly falls after 3 kHz. These results suggest that the simultaneous estimation of temperature and viscosity can be achieved by performing measurements at two different drive field settings that provide complementary temperature/viscosity sensitivities. Alternatively, temperature estimation alone can be achieved with a single measurement at drive field frequencies above 3 kHz, where viscosity sensitivity is minimized. CONCLUSIONS: This work demonstrates highly promising temperature and viscosity sensitivities for TAURUS, highlighting its potential for simultaneous estimation of these two environmental parameters via MNP relaxation. The findings of this work reveal the potential of a hybrid MPI-MFH system for real-time monitored and localized thermal ablation treatment of cancer.