Highly Sensitive and High-Throughput Magnetic Resonance Thermometry of Fluids Using Superparamagnetic Nanoparticles

Magnetic resonance imaging (MRI) enables noninvasive three-dimensional thermometry, which has potential applications in biological tissues and engineering systems. In biological tissues, where MRI is routinely used to monitor temperature during thermal therapies, $T_1$ or $T_2$ contrast in water are relatively insensitive to temperature, and techniques with greater temperature sensitivity, such as chemical shift or diffusion imaging, suffer from motional artifacts and long scan times. MR thermometry is not well developed for nonbiological or engineering systems. We describe an approach for highly sensitive and high-throughput MR thermometry that is not susceptible to motional artifacts and could be applied to various biological systems and engineering fluids. We use superparamagnetic iron-oxide nanoparticles (SPIONs) to spoil $T_2$ of water protons. Motional narrowing results in proportionality between $T_2$ and the diffusion constant, dependent only on the temperature in a specific environment. Our results show, for pure water, the nuclear magnetic resonance linewidth and $T_2$ follow the same temperature dependence as the self-diffusion constant of water. Thus, a $T_2$ mapping is a diffusion mapping in the presence of SPIONs, and $T_2$ is a thermometer. For pure water, a $T_2$ mapping of a 64 × 64 image (voxel size = 0.5 mm × 0.5 mm × 3 mm) in a 9.4 T MRI scanner resulted in a temperature resolution of 0.5 K for a scan time of 2 min. This indicates a highly sensitive and high-throughput MR thermometry technique that potentially has a range of applications from thermal management fluids to biological tissues.

Figure 1

(a) Temperature dependence of NMR linewidths (and consequently T₂*) of protons in water in the presence of SPIONs and its comparison to the temperature dependence of the reciprocal of the diffusion coefficient of proton in water (blue solid line). The linewidths show the same temperature dependence as the reciprocal of the diffusion constant, until the concentration of the nanoparticles is low and the linewidth due to improper shimming becomes significant (at around 50 ppm nanoparticles by volume). (b) Comparison of NMR linewidth of protons in water with the volume fraction of SPOINs at 278 K (black squares), 318 K (open black triangles), and 348 K (black circles). The linewidths scale linearly with the volume fraction at all temperatures except for low-volume fraction, where shimming imperfections contribute significantly to the linewidth.

Figure 2

Temperature dependence of relaxation rates 1/T₂ (red) and 1/T₂*(black) at SPION volume fractions of 110 ppm (open circles), 65 ppm (filled circles), and 10 ppm (open triangles). The temperature dependence of the relaxation rates is compared with the temperature dependence of the reciprocal of the diffusion constant in water (blue solid line). The measurements of relaxation rates for SPION volume fractions of 110 and 65 ppm are carried out in a Unity/Inova 14.1 T NMR spectrometer while the measurements for the volume fraction of 10 ppm is carried out in Bruker BioSpec 9.4 T preclinical MRI scanner. For lower volume fractions (<65 ppm) and high temperature, T₂* has a significant contribution from field inhomogeneities due to imperfect shimming while T₂ is not affected and remains a good thermometer.

Figure 3

T₂ maps obtained from cross-section (coronal) images of water in a cylindrical container at 23 (a), 25 (b), 29  (c), and 31 °C (d) in the presence of SPIONs (10 ppm by volume). The T₂ maps are obtained from 12 MSME images with equal echo spacing of 5 ms. The gray scale represents T₂ values with black corresponding to 0 ms and white corresponding to 8 ms. The average echo times determined from the T₂ maps are 6.25, 6.57, 7.19, and 7.42 ms respectively.