Introduction

Transmit gain (TG) calibration is performed prior to MR image acquisition to determine the radiofrequency (RF) output necessary to produce desired excitation angles. Accurate TG calibration is crucial for reproducible imaging of nonproton hyperpolarized (HP) agents because each HP signal excitation attenuates the total available HP signal pool, which directly effects future measurements.

We have previously implemented an automated script to perform saturation-recovery ¹³C TG calibration on a preclinical 7T MRI for metabolic imaging of HP [1-¹³C]-pyruvate in small animals. This approach leverages strong signal from a heavily doped 8M ¹³C-urea phantom. While a thermal ¹²⁹Xe gas phantom was recently introduced for HP gas MRI¹, this approach may not easily translate into a small calibration phantom that can fit alongside small animals in a 7T MRI.

We developed a TG calibration sequence for nonproton HP MRI using the Bloch-Siegert shift^2,3 in preparation for upcoming HP ¹²⁹Xe gas MRI small animal studies on a 7T MRI. Our method allows for fast, accurate, and robust TG calibration of nonproton HP MRI studies with minimal error (<5%) for ¹H and ¹³C.

Methods

Bloch-Siegert Effect: The Bloch-Siegert shift describes the effect by which an off-resonant RF pulse induces a phase shift without altering spin nutation of the target nuclei:

$$$\phi_{BS}=\int_{0}^{T}\frac{(\gamma B_{1}(t_{BS}))^{2}}{2(2\pi f_{BS})(t_{BS})}dt_{BS}\tag1$$$

Briefly, an off-resonant RF pulse of duration t_BS at offset frequency f_BS is applied following a single excitation pulse (Figure 1). Two repeated measurements with ±f_BS allow for the quantification of the Bloch-Siegert phase difference, φ_BS, which is proportional to the square of the amplitude of the RF field.

Imaging: All imaging and spectroscopy were carried out using a 7T Biospec USR7030 small animal MR scanner (Bruker Preclinical Imaging, Billerica, MA) and a dual-tuned ¹³C/¹H volume coil with a 35 mm inner diameter. An optimized Fermi shaped RF pulse⁴ was used for off-resonant Bloch-Siegert pulse generation. An analysis script was implemented as a standalone MATLAB (The MathWorks, Natick, MA) executable function that can be called automatically following data acquisition.

Phase Error Correction: During sequence development we discovered a persistent phase error (φ_Switch) as a result of switching the frequency of the synthesizer to f_BS, for the duration of the Bloch-Siegert pulse (t_BS), between signal excitation and readout. In order to isolate φ_Switch, the Bloch-Siegert pulse was driven with 0 W (φ_BS = 0°) and tested at several f_BS values ranging 0 to 2000 Hz at ¹H (Figure 2). Other important pulse parameters included t_BS = 3.5 ms, TR = 10,000 ms, two repetitions, and two dummy scans.

Phase induced by switching the synthesizer frequency can be written as a function of the offset frequency, pulse duration, and a minor delay associated with the frequency switch:

$$$\phi_{Switch}=2\pi f_{BS}(t_{BS}+t_{Switch})\tag2$$$

The total phase (Δφ) observed following the Bloch-Siegert pulse reflects a combination of these two effects:

$$$\triangle\phi=\phi_{BS}+\phi_{Switch}\tag3$$$

Bloch-Siegert Validation: Following characterization of φ_Switch, the Bloch-Siegert pulse was driven with power values ranging 0 to 20 W at ¹H (Figure 3) and ¹³C. Measured Δφ values were corrected for phase induced by the frequency switch. Next, φ_BS was used to calculate TG calibration factors at each Bloch-Siegert power level. Finally, percent difference between our measured Bloch-Siegert TG calibration factors and the validation TG calibration factors for ¹H and ¹³C, measured automatically by the scanner and using our existing saturation-recovery ¹³C TG calibration method, were calculated.

Results

φ_Switch was measured to be ~7° when t_BS = 3.5 ms, f_BS = 2000 Hz and t_Switch was calculated to be ~10 µS.

Figure 4 provides a summary of percent difference between our measured Bloch-Siegert TG calibration factors and the validation TG calibration factors for ¹H and ¹³C at power levels ranging 0 to 20 W. The reference power required for the scanner to generate a 90°, 1 ms hard excitation pulse was measured to be 3.05 W and 11.69 W for ¹H and ¹³C, respectively.

Discussion and Conclusion

We successfully utilized the Bloch-Siegert effect to develop a robust pulse sequence that can be used for fast and automated TG calibration of nonproton nuclei on a 7T small animal MRI with minimal error (<5%). Our method performs well for ¹H and ¹³C. We are currently in the process of testing our method using HP ¹²⁹Xe gas and a dual-tuned ¹²⁹Xe/¹H volume coil. Our Bloch-Siegert pulse and supporting developmental tools can be found at: https://github.com/mda-mrsl/PV-BSXmit

Acknowledgements

This work was supported by funding from the National Cancer Institute of the National Institutes of Health (R21CA280799, T32CA196561). The content is solely the responsibility of the authors and does not necessarily represent the official views of the sponsors.