Clinicians and scientists interested in MRF, parameter mapping and MR contrast agents.MR contrast agents are typically imaged with either T₁ or T₂ weighted sequences. However, valuable information can be lost when only one parameter of relaxation is assessed. In addition, by assessing both T₁ and T₂, it is possible under certain circumstances to image two contrast agents together and resolve their individual concentrations. Acquiring sequential T₁ and T₂ maps, however, is time consuming and not aligned with the ever-increasing pressure to reduce scan time. In addition, dynamic phenomena such as perfusion and metabolism are not suited to time-incoherent sequential imaging. To address this we aimed here to use Magnetic Resonance Fingerprinting (MRF) [1] to acquire multiple MR-parameter maps in a single rapid acquisition. The primary goal was to determine whether MRF would allow mixed concentrations of Gadolinium and Ferumoxytol to be accurately quantified. In addition, we aimed to determine whether the approach would be suitable for quantifying the bound and unbound state of targeted contrast agents.

Two contrast agents (CAs) were initially used; the gadolinium-based contrast agent Gd-DTPA (Magnevist) and the iron-oxide contrast agent ferumoxytol (Feraheme). Three conditions were tested: Magnevist only, Feraheme only, and a mixture of both agents. Magnevist concentrations varied from 0.2 to 1.0 mM, and ferumoxytol concentrations from 5 to 20 μg/ml. All phantoms were diluted in phosphate-buffered saline (PBS). The samples were first mapped using conventional mapping techniques. The relaxivities of Magnevist and Feraheme at 3T were derived from these maps, and in turn used to calculate the CAs’ concentrations. Next, MRF scans were performed using a Fast Imaging with Steady State Precession (FISP) sequence [2], and T₁ and T₂ maps were derived using dictionaries tailored to the expected relaxivities. The concentrations of Magnevist and Feraheme were derived from the MRF derived T₁ and T₂ maps using a priori calculated values of their relaxivities. Gold-truth concentrations of the agents were derived using inductively plasma coupled mass spectroscopy (ICP-MS).
In addition, the ability of MRF to measure the relaxivities of the albumin bound Gd chelate, MS-325, was tested. Two sets of samples containing concentrations of MS-325 were created with concentrations varying from 0.1 to 1.0 mM, the first set was diluted in PBS and the second set was diluted in PBS + 4.5% human serum albumin (HSA). Using MRF the longitudinal and transverse relaxivities were calculated for both the unbound as the bound (to HSA) state, results were analyzed and compared to values from previous research [3].

The created parametric maps that were used for calculating the Magnevist and Feraheme relaxivity data are displayed in figure 1 and 2. Relaxivity data showed good linear fits (R-square >0.99) for both the conventional and the MRF method, although all relaxivity results showed slightly higher values with MRF. This was more pronounced for transverse relaxivity r₂ MRF data. Results for the quantification of the CAs in mixture were obtained and compared to the known concentrations. Calculations using the conventionally obtained data resulted in measurements of 96%, 100%, 102% and 112% for Magnevist and 99%, 98%, 99% and 98% for Feraheme, compared to the known concentrations. Data acquired using MRF resulted in concentration calculations of 103%, 106%, 115% and 97% for Magnevist and 96%, 100%, 88% and 105% for Feraheme. Relaxivity calculations for the contrast agent MS-325 showed good linear fits (R-square >0.98) for both longitudinal and transversal relaxivity, resulting in r₁ = 7.6 ± 0.7 mM^-1s^-1 and r₂ = 10.0 ± 0.8 mM^-1s^-1 for the unbound state and r₁ = 9.7 ± 0.4 mM^-1s^-1 and r₂ = 60.2 ± 6.6 mM^-1s^-1 for the bound state. Previously conducted research indicated accurate measurements [3].This research indicates the feasibility of quantifying multiple contrast agents in a mixture using MRF. Although the relaxivity values for MRF did not completely agree with conventional findings, the quantification of the CAs was in most cases >90% accurate. In addition, MRF could also be useful for quantifying the bound and unbound state of certain CAs in near real-time, which raises numerous possibilities for in-vivo applications.