PURPOSE

Some X-ray contrast agents that are approved for human use contain exchangeable protons that may give rise to exchange-based effects on MRI, including chemical exchange saturation transfer (CEST). However, both the amide and hydroxyl protons in such agents are not in slow exchange with water, so that high powers are required to effectively saturate these groups, resulting in non-specific direct saturation and semi-solid magnetization transfer effects. We have previously shown that spin-lock sequences at high field are also sensitive to chemical exchange¹, and can be more suitable for detecting intermediate and fast exchanging groups than CEST. In addition, R_1p, the spin-lock relaxation rate obtained by comparing signals at a fixed locking power but at different locking times, is well estimated by linearly adding individual exchange contributions, whereas CEST signals from different exchanging pools have mutual interactions². Moreover, the variation of R_1p with locking power (the R_1p dispersion) can be used to estimate exchange parameters and emphasize the effects of protons of a specific exchange rate. Here we demonstrate the ability of spin-lock imaging to detect iohexol in tumor bearing rat brains.

METHODS

We adopt the theoretical model of Chopra et al.³ in which $$$R_{1p}=(R_2+\frac{R^\infty_{1p}\omega^2_1}{S^2_p})/(1+\frac{\omega^2_1}{S^2_p})$$$, and S_p, R₂, and $$$R^\infty_{1p}$$$ may be regarded as parameters in the fitting of R_1p vs. the spin-lock amplitude (w₁). In realistic cases,$$$S^2_p \approx k^2_{sw}+\triangle\omega^2_s$$$, where k_sw is the exchange rate of protons with chemical shift $$$\triangle\omega_s$$$. An R_1p contrast is defined then as $$$\triangle R_{1p}=R_{1p}(low)-R_{1p}(high)$$$ where R_1p(low) and R_1p(high) are the R_1p values acquired with low and high w₁. A series of iohexol samples served as test solutions to evaluate the dependence of $$$\triangle R_{1p}$$$ and S_p on pH and concentration. Four samples were made with 30 mM iohexol and pH was titrated to 6.6, 7.0, 7.2, and 7.4, respectively. Three samples were made with 15, 30, and 45 mM iohexol (pH = 7.0), respectively. Four rats bearing 9L tumors were also studied before and after administration of the agent. Spin-lock signals were acquired with locking amplitudes from 100 to 3162 Hz on a 9.4 T Varian small animal scanner. For iohexol administration, spin-lock signals were acquired at different times before and after injection of 2 mL 1.3 M iohexol solution. R_1p was calculated for each voxel, and $$$\triangle R_{1p}$$$ was calculated by comparing values at power of 100 and 3162 Hz. S_p was obtained by fitting the R_1p dispersion to Chopra’s equation. To increase SNR in fitting S_p, we applied a Gaussian filter to the spin-lock images, averaged all the $$$\triangle R_{1p}$$$ images acquired after injection, and omitted voxels with no significant concentration of agent.

RESULTS

Fig. 1 shows the R_1p dispersion, $$$\triangle R_{1p}$$$, and S_p for the four iohexol samples for different pH but constant concentration (a, c, and e), and the three iohexol samples for different concentrations but constant pH (b, d, and f). Note that $$$\triangle R_{1p}$$$ is sensitive to concentration, but not pH. In contrast, S_p is sensitive to pH, but not concentration. Fig. 2a shows the time course of the differences of $$$\triangle R_{1p}$$$ values in rat tissues before and after injection. It was found that $$$\triangle R_{1p}$$$ signals in tumors increase quickly after injection and decrease slowly up to 1 h, whereas intact brain showed little change. Fig. 2b, 2c, and 2d shows maps of water longitudinal relaxation rate (R_1w) and $$$\triangle R_{1p}$$$ before and at 10 mins after injection. Note that the $$$\triangle R_{1p}$$$ contrast in tumor is enhanced after injection. S_p in tumors was fitted to be 4282 ± 786, 3384 ± 941, 4328 ± 1524, and 4614 ± 356 s^-1 for the four rats. By comparing them with S_p in Fig. 1e, this corresponds to an extracellular pH ranging from 6.6 to 7.0. Fig. 3 shows the S_p maps from the four rats.

DISCUSSION

It is challenging to separate the effects of agent concentration and exchange rate in CEST. Here, we show that spin lock can be used to derive quantities such as $$$\triangle R_{1p}$$$ and S_p, which depend separately on these two parameters. In the fitting of S_p, although we used a single-pool fit to a two-pool system, the S_p of iohexol should depend on the $$$\triangle\omega_s$$$ and k_sw of the two exchanging groups, so the derived value represents a form of average. Fig. 1d and 1f indicate that S_p can provide concentration-independent pH-weighted imaging.

CONCLUSION

We show that spin lock can be applied to study the effects of exogenous agents with exchanging groups, and exchange dependent contrast may have some advantages over conventional relaxation agents and CEST.

Acknowledgements

No acknowledgement found.

References

1. Cobb JG, Xie JP, Li K, et al. Exchange-mediated contrast agents for spin-lock imaging. Magn Reson Med. 2012;67(5): 1427-1433. 2. Zaiss M, Bachert P. Exchange-dependent relaxation in the rotating frame for slow and intermediate exchange - modeling off-resonant spin-lock and chemical exchange saturation transfer. NMR Biomed. 2013;26(5):507-518. 3. Chopra S, McClung RED, Jordan RB. Rotating-frame relaxation rates of solvent molecules in solutions of paramagnetic ions undergoing solvent exchange. J Magn Reson. 1984; 59:361-372.