Giles E. Santyr

Comparison of a reference region model with direct measurement of an AIF in the analysis of DCE-MRI data

Models have been developed for analyzing dynamic contrast‐enhanced (DCE)‐MRI data that do not require measurements of the arterial input function (AIF). In this study, experimental results obtained from a reference region (RR) analysis are compared with results of an AIF analysis in the same set of five animals (four imaged twice, yielding nine data sets), returning estimates of the volume transfer constant (Ktrans) and the extravascular extracellular volume fraction (ve). Students t‐test values for comparisons of Ktrans and ve between the two models were 0.14 (P = 0.88) and 0.85 (P > 0.4), respectively (where the high P‐values indicate no significant difference between values derived from the two models). Linear regression analysis indicated there was a correlation between Ktrans extracted by the two methods: r2 = 0.80, P = 0.001 (where the low P‐value indicates a significant linear correlation). For ve there was no such correlation (r2 = 0.02). The mean (absolute) percent difference between the models was 22.0% for Ktrans and 28.1% for ve. However, the RR parameter values were much less precise than the AIF method. The mean SDs for Ktrans and ve for the RR analysis were 0.024 min–1 and 0.06, respectively, vs. 0.002 min–1 and 0.03 for AIF analysis. Magn Reson Med 57:353–361, 2007.

Theoretical signal‐to‐noise ratio and spatial resolution dependence on the magnetic field strength for hyperpolarized noble gas magnetic resonance imaging of human lungs

In hyperpolarized noble gas (HNG) magnetic resonance (MR) imaging, the available polarization is independent of magnetic field strength and for large radiofrequency (rf) coils, such as those used for chest imaging, the body noise becomes the primary noise source making signal-to-noise ratio (SNR) largely frequency independent at intermediate field strengths (0.1-0.5 T). Furthermore, the reduction in the transverse relaxation time, T2, of HNG in lungs with increasing field strength, results in a decrease in the achievable SNR at higher fields. In this work, the optimum field strength for HNG MR imaging was theoretically calculated in terms of both SNR and spatial resolution. SNR calculations used the principle of reciprocity and included contributions to the noise arising from both coil and sample losses in a chest-sized coil for lung imaging. The effects of susceptibility differences, transverse relaxation time, and diffusion were considered in the resolution calculations. The calculations show that the optimum field strength for HNG MR imaging of human lungs is between 0.1 and 0.6 T depending on gas type (helium or xenon) and sample size. At the field strengths currently used by conventional clinical proton MR imaging systems (1-3 T), the predicted SNR are 10%-50% lower than at the optimum field with only slightly worse spatial resolution (10%-20%). At higher fields (>3 T), however, the SNR degrades considerably reducing the achievable spatial resolution. Although HNG of the lung is still feasible at very low field strengths (<50 mT), the available SNR is much lower than at optimum fields and this reduces the achievable spatial resolution. These findings suggest that HNG imaging may be optimally performed at much lower field strengths (0.1-0.6 T) than conventional clinical proton MR imaging systems. This could considerably decrease cost, improve patient access, and reduce chemical shift and susceptibility artifacts and rf heating.

Measurement of xenon diffusing capacity in the rat lung by hyperpolarized 129Xe MRI and dynamic spectroscopy in a single breath-hold

We used the dual capability of hyperpolarized 129Xe for spectroscopy and imaging to develop new measures of xenon diffusing capacity in the rat lung that (analogously to the diffusing capacity of carbon monoxide or DLCO) are calculated as a product of total lung volume and gas transfer rate constants divided by the pressure gradient. Under conditions of known constant pressure breath‐hold, the volume is measured by hyperpolarized 129Xe MRI, and the transfer rate is measured by dynamic spectroscopy. The new quantities (xenon diffusing capacity in lung parenchyma (DLXeLP)), xenon diffusing capacity in RBCs (DLXeRBC), and total lung xenon diffusing capacity (DLXe)) were measured in six normal rats and six rats with lung inflammation induced by instillation of fungal spores of Stachybotrys chartarum. DLXeLP, DLXeRBC, and DLXe were 56 ± 10 ml/min/mmHg, 64 ± 35 ml/min/mmHg, and 29 ± 9 ml/min/mmHg, respectively, for normal rats, and 27 ± 9 ml/min/mmHg, 42 ± 27 ml/min/mmHg, and 16 ± 7 ml/min/mmHg, respectively, for diseased rats. Lung volumes and gas transfer times for LP (TtrLP) were 16 ± 2 ml and 22 ± 3 ms, respectively, for normal rats and 12 ± 2 ml and 35 ± 8 ms, respectively, for diseased rats. Xenon diffusing capacities may be useful for measuring changes in gas exchange associated with inflammation and other lung diseases. Magn Reson Med, 2006.

Improvement in Breast Lesion Characterization With Dynamic Contrast-Enhanced MRI Using Pharmacokinetic Modeling and Bookend T 1 Measurements

Dynamic contrast‐enhanced breast MR imaging was performed on 14 patients (five cancerous lesions, nine benign) with slice‐selective spoiled gradient‐recalled echo (2D SPGR) imaging. Adiabatic saturation recovery T1 measurements were performed before (T1pre) and after (T1post) 2D SPGR imaging. These two “bookend” T1 measurements were used to calibrate the equations which were employed to convert the time course of the 2D SPGR signal strength to T1‐vs.‐time, which in turn was used to compute the gadolinium concentration‐vs.‐time ([C](t)) in the lesion. The extraction‐flow product (EF) was computed for each lesion by pharmacokinetic modeling of [C](t). For this study, EF provided a sensitivity and specificity for cancer of 100% and 78%, respectively. When only T1pre was used to estimate [C](t) (which assumes a priori knowledge of the shape and amplitude of the slice profile), the sensitivity and specificity fell to 80% and 56%, respectively. This is presumably due to unexpected variations in the shape and/or amplitude of the slice profile, which could be caused by factors such as patient‐to‐patient variations in breast geometry or inconsistently set transmit gains. Therefore, both T1pre and T1post measurements are necessary for optimum sensitivity and specificity using pharmacokinetic analysis. Magn Reson Med 51:1066–1070, 2004.

Gas-phase nuclear magnetic relaxation in 129Xe revisited

In this contribution gas-phase 129Xe spin-lattice relaxation time measurements are extended to conditions (pressure, temperature, magnetic-field strength, isotope composition) not previously used. It is shown that wall effects become apparent at densities below ∼20 amagat, and that these become dominant below ∼3 amagat. A significant new discovery from field-dependent studies is that, in addition to the previously identified field-independent spin–rotation relaxation operative in the bulk gas, there is a contribution from the modulation of the chemical shift that depends on the square of the applied magnetic-field strength. The weak temperature dependence of the relaxation times can be understood in terms of the opposite temperature coefficients of the field-independent and field-dependent contributions to the relaxation. The spin–rotation contribution and its temperature dependence are calculated and found to be in good agreement with the experimental data. The low field, low density limit for the relaxa...

Coatings for optical pumping cells and short-term storage of hyperpolarized xenon

For a number of years now, siloxanes have been the materials of choice for coating vessels used in the production and short-term storage of hyperpolarized xenon. The methods used to apply this material, however, often vary from one research group to another and it is commonly reported that it is difficult to obtain cells with consistently long spin-lattice relaxation times (T1) and high-polarization levels. In a series of controlled experiments individual production variables were altered and optimized, leading to improved protocols for the reliable production of high-quality siloxane-coated cells. During these studies we discovered that the surface-induced relaxation rates in bare and coated Pyrex cells differ profoundly. This information on Xe relaxation helps to define the limits on the way pumping cells can be improved and suggests the need for further fundamental work on relaxation mechanisms.

Coatings for optical pumping cells and extending the lifetime of hyperpolarized xenon

We propose evaporative induction heating as a method for the reliable production of coatings for glass cells suitable for the optical pumping and storage of hyperpolarized xenon. The long spin-lattice relaxation times of hyperpolarized xenon-129 contained in cells coated with polyethylene or dotriacontane showed that high quality coatings could be prepared this way. Measurements on cells coated with deuteriated versions of these compounds showed that the expected improvement in performance with isotopic substitution did not occur. This leads to some questions about the level of understanding of wall-induced relaxation in polarization cells.

Laser-polarized 129Xe NMR at 1.88 T and 8.5 mT: a signal-to-noise ratio comparison

The signal-to-noise ratio of nuclear magnetic resonance signals from laser-polarized 129Xe gas was investigated at 8.5 mT and compared to that of signals acquired at 1.88 T. A dedicated 8.5 mT resistive magnet was constructed and used to acquire the signals. The SNR for 1 atm of xenon gas with a polarization of 1% was measured to be 1900 at a field of 1.88 T. Under identical acquisition conditions, the SNR at 8.5 mT was about 60 (or 32 times lower). After measuring and including all of the electrical factors of the detection systems at each field strength, theory indicates the SNR value measured at 8.5m T should be about 36 times lower. Considering the widely differing frequencies and completely different detection systems the agreement is quite good and indicates that extrapolating the frequency dependence of the SNR down to very low fields does work as long as the detection system parameters are carefully accounted for. This work suggests that magnetic resonance (MR) imaging is achievable on ideal gas samples at 8.5 mT using laser-polarized 129Xe gas down to the practical resolution limit of about 0.5mm, although the SNR will be very low (approximately 1.4). The feasibility of imaging small animals at 8.5 mT is discussed and it is suggested that a field of about 50 mT is required.

Increasing the spin-lattice relaxation time of hyperpolarized xenon ice at 4.2 K

After cryogenic trapping of hyperpolarized xenon produced by optical pumping, significant increases in the spin-lattice relaxation time (T1) of 129Xe in solid xenon at 4.2 K can be achieved by annealing the solid at an appropriate temperature. Thus, T1 at 4.2 K in a field of 180 G increased from 20.4 to 35 h on warming a sealed sample initially condensed at 77 K in an isopentane bath at 113 K for 10 min. This provides further confirmation that the primary relaxation mechanism for 129Xe at low temperature is cross relaxation to 131Xe and demonstrates that long term storage of hyperpolarized xenon produced using flow polarizers is feasible.

Two-compartment radial diffusive exchange analysis of the NMR lineshape of Xe129 dissolved in a perfluorooctyl bromide emulsion

Hyperpolarized (129)Xe (xenon) gas dissolved in a perfluorooctyl bromide (PFOB) emulsion stabilized with egg yolk phospholipid (EYP) is a possible contrast agent for quantitative blood flow measurements using magnetic resonance imaging. The NMR line shape of xenon dissolved in PFOB emulsion depends strongly on the exchange of spins between PFOB and water. The exchange in this system depends on three factors: the geometrical factors (i.e., droplet size and surrounding water volume), the permeability of the EYP monolayer surrounding the droplet, and the diffusion coefficients of xenon in the two media. A theoretical model which predicts the line shape of xenon in the emulsion based on the Bloch-Torrey equations is presented. Fitting the full width at half maximum (FWHM) of the theoretical line shapes with the FWHM of the experimental spectra obtained from emulsions with different water dilutions allows estimation of the volume-weighted average diameter of the PFOB droplets (3.5+/-0.8) microm and the permeability of the EYP membrane surrounding the droplet (58+/-14) microm / s.