Introduction

Chemical exchange saturation transfer (CEST) MRI is sensitive to detecting tissue changes following acute stroke, tumors, and epilepsy^1-3. CEST MRI is versatile, but the CEST measurement depends not only on labile proton concentration and exchange rate but also on experimental parameters, such as the duration of radiofrequency (RF) saturation (T_s) and the relaxation recovery delay (T_d). Although the use of long RF irradiation enhances the CEST effect, it unavoidably prolongs the total scan time, making CEST MRI challenging for routine clinical use. For typical human CEST scans, the RF saturation time can range from 0.2 to 3.5 s at 3 T^3-5. Therefore, CEST protocols vary substantially among centers, and it is urgent to unify results obtained under different T_s and T_d. A quasi-steady-state (QUASS) CEST algorithm has been recently proposed to reduce impacts of experimental T_s and T_d parameters on the CEST measurement⁶. We hypothesized that the QUASS algorithm can be generalized for the multi-slice acquisition in addition to T_s and T_d correction and tested it for fast multi-slice imaging in tumor patients.

Theory

A schematic diagram of the fast multi-slice spin echo (SE) echo planar imaging (EPI) CEST MRI sequence is shown in Figure 1. The apparent CEST Z-spectrum is calculated by normalizing the saturated scan signal ($$$I'_{sat}$$$) with the unsaturated control scan signal ($$$I'_{0}$$$)⁷:
$$\frac{I'_{sat}\left(\Delta\omega\right)}{I'_{0}}=\frac{\left(1-e^{-R_{1w}T_{d}}\right)e^{-R_{1\rho}\cdot T_{s}}+\frac{R_{1w}}{R_{1\rho}}\cos^{2}\theta\left(1-e^{-R_{1\rho}\cdot T_{s}}\right)}{1-e^{-R_{1w}\cdot \left(T_{s}+T_{d}+pld\right)}}e^{-R_{1w}\cdot pld}+\frac{1-e^{-R_{1w}\cdot pld}}{1-e^{-R_{1w}\cdot \left(T_{s}+T_{d}+pld\right)}}\qquad\qquad\left[1\right]$$
where $$$R_{1w}$$$ is the bulk water longitudinal relaxation rate, $$$I_{0}$$$ is the equilibrium magnetization, $$$R_{1\rho}$$$ is the spin-lock relaxation rate and $$$\theta=\arctan\left(\frac{\gamma B_{1}}{\Delta \omega}\right)$$$, in which $$$\gamma$$$ is the gyromagnetic ratio and B₁ and $$$\Delta \omega$$$ are the amplitude and offset frequency of the RF saturation, respectively. Under the assumption that T_s and T_d are much longer than the post-label delay (PLD) time, the PLD-corrected apparent CEST Z-spectrum can be derived as,
$$\frac{I'^{pldcor}_{sat}\left(\Delta \omega\right)}{I'^{pldcor}_{0}}\cdot \left\{\frac{1-e^{-R_{1w}\left(T_{s}+T_{d}\right)}}{1-e^{-R_{1w}\cdot T_{d}}}\right\}=e^{-R_{1\rho}\cdot T_{s}}+\frac{R_{1w}\cdot \cos^2 \theta}{R_{1\rho}\cdot \left(1-e^{-R_{1w}\cdot T_{d}}\right)}\cdot \left(1-e^{-R_{1\rho}\cdot T_{s}}\right)\qquad\qquad\left[2\right]$$
in which the superscript pldcor denotes the PLD-corrected apparent signals. $$$R_{1\rho}$$$ can be numerically solved, and the QUASS-corrected CEST Z-spectrum can be calculated as
$$\left(\frac{I_{sat}\left(\Delta \omega\right)}{I_{0}}\right)^{QUASS}=\frac{R_{1w}}{R_{1\rho}}\cos^{2}\theta\qquad\qquad\left[3\right]$$

Methods

Informed written consent was obtained following IRB-approved protocol. Three healthy volunteers (1 male and 2 females, age 25-39) and a brain tumor patient underwent CEST brain MRI using a 3T MR scanner (Magnetom Prisma, Siemens Medical Solutions, Erlangen, Germany) with a 64-channel head and neck coil. Two sets of multi-slice brain images with different pairs of saturation duration and relaxation delays (T_s/T_d =1 s/1 s and 2 s/2 s) were acquired in the axial orientation. Multiple rectangular RF pulse (duration = 499.8 ms, duty cycle = 99.9 %, B₁ = 0.7 µT) were exploited to emulate continuous wave saturation. The alternating offset frequencies of RF saturation were varied from -5 to 5 ppm, with increments of 0.125 ppm. The imaging parameters were: imaging readout time = 547 ms, TE = 33 ms, FOV = 220 × 220 mm², in-plane matrix = 110 × 110, the number of slices = 8, interleaving slice ordering, slice thickness = 5 mm with 25% slice gap, with fat suppression, 1 average, and readout bandwidth = 1976 Hz/pix. Total imaging time was 3 min 31 s and 6 min 17 s for T_s/T_d of 1 s/1 s and T_s/T_d of 2 s/2 s, respectively. T_1w-weighted images were acquired using eight inversion recovery delay (0.25, 0.5, 0.75, 1.0, 1.25, 1.5, 2.0, and 3.0 ms) with relaxation delay of 5.0 s, and 2 average.

Results/Discussion

Figure 2 shows CESTR′, PLD-corrected CESTR′, and CESTR^QUASS of healthy human brain images at 3.5 ppm with T_s/T_d of 1/1 s and 2/2 s. The apparent CEST effects noticeably differed in the direction of the multi-slice acquisition and under different T_s and T_d (Figs. 2A, B). Although PLD-corrected apparent CEST effects showed little PLD dependency, there was a noticeable CEST contrast discrepancy between T_s/T_d of 1/1 s and 2/2 s (Figs. 2C, D). In contrast, the QUASS CEST effects were more consistent between T_s/T_d of 1/1 s and 2/2 s (Figs. 2E, F). Specifically, for WM, the PLD-corrected CESTR′ was 3.32±1.09% and 3.83±1.11% for T_s/T_d of 1/1 s and 2/2 s, respectively, while CESTR^QUASS being 4.91±1.23% and 4.87±1.20%. In GM, the PLD-corrected CESTR′ was 2.00±1.09% and 2.54±1.09% for T_s/T_d of 1/1 s and 2/2 s, respectively, while their corresponding CESTR^QUASS being 3.36±1.56% and 3.28±1.34%. In Figure 3, the Bland-Altman analysis bias was -0.54% for PLD-corrected CESTR′ and 0.03% for CESTR^QUASS, indicating that the proposed QUASS algorithm minimizes the bias in the CEST measurement, advantageous over simple PLD correction. Compared with the contralateral normal tissue, the tumor region shows greater CESTR in Figure 4. In particular, the apparent CEST effect increased with the increase of T_s and T_d, while the QUASS CEST effects showed little T_s and T_d dependency.

Conclusion

Our study demonstrated fast multi-slice CEST brain MRI with healthy volunteer and brain tumor patient scans. In addition, the QUASS algorithm minimizes the effect of the choice of saturation duration, relaxation delay, and post-label delay time on the CEST MRI quantification. In summary, the QUASS algorithm enables an expedited CEST MRI acquisition strategy, thereby increasing its clinical translatability, particularly at high magnetic fields, while providing accurate steady-state CEST measurement.

Acknowledgements

No acknowledgement found.

References

[1] Sun PZ et al., Detection of the ischemic penumbra using pH-weighted MRI. J Cereb Blood Flow Metab 2007;27(6):1129-1136. [2] Daivis KA et al., Glutamate imaging (GluCEST) lateralizes epileptic foci in nonlesional temporal lobe epilepsy. Sci Transl Med. 2015;7(309):309ra161. [3] Togao O et al., Amide proton transfer imaging of adult diffuse gliomas: correlation with histopathological grades. Neuro Oncol. 2014;16(3):441-448. [4] Harston GW et al., Identifying the ischaemic penumbra using pH-weighted magnetic resonance imaging. Brain 2015;138(1):36-42 [5] Goerke S et al., Clinical routine acquisition protocol for 3D relaxation‐compensated APT and rNOE CEST‐MRI of the human brain at 3T. Magn Reson Med 2021. [6] Sun PZ., Quasi‐steady state chemical exchange saturation transfer (QUASS CEST) analysis—correction of the finite relaxation delay and saturation time for robust CEST measurement. Magn Reson Med 2021;85(6):3281-3289. [7] Jin T, et al., Spin‐locking versus chemical exchange saturation transfer MRI for investigating chemical exchange process between water and labile metabolite protons. Magn Reson Med 2011;65(5):1448-1460.