Introduction

Parallel radiofrequency transmission (pTx) system implementations have grown steadily in the last decade^1,2,3. At present, some commercial systems are available for research; however, in-house designs remain a relevant alternative not only from a cost perspective, but also because optimal channel counts and optimal pTx methods for different clinical applications are still under investigation. Often, in-house pTx systems are assembled with commercial and in-house designed components that can require additional measurements to determine certain system characteristics. Although most measurements can be performed “on the bench,” it is advantageous to conduct measurements during imaging for better accuracy and improved insight about true system behaviour, as some MRI applications (e.g. reduced field-of-view imaging) are more sensitive to pTx system uncertainties. In particular, signal-to-noise ratio (SNR) analysis of pTx systems is an important consideration to maximize performance. This work uses a simple multi-purpose method to conduct SNR analysis and noise characterization of a 4-channel pTx system integrated on an existing 3 T MRI system.

Theory

The signal-to-noise ratio is defined as: $$SNR = signal power / noise power$$
The noise factor, F, is defined as: $$F = SNRin / SNRout$$
where SNR_in is the ratio of radiofrequency (RF) signal and noise power at the input of the device under test for a frequency of interest, and SNR_out is the analogous ratio at the output. F is commonly expressed in decibels (dB) as the noise figure (NF) and it is then defined as, $$NF = 10·log(F).$$

Methods

To measure SNR of our 4-channel pTx system³ that is integrated on an existing 3 T MRI system (Magnetom Prisma, Siemens), RF directional couplers (C5497-13 / C7008-102, Werlatone, Patterson, NY) were inserted into the RF transmission pathway as shown in Fig. 1. Normal operation of the MRI system is possible because of the insertion of the dummy load inside the magnet room, enabling RF excitation using the pTx system³. Using a spectrum analyzer (N9010A, Keysight Technologies, Santa Rosa, CA) both signal and noise power were measured for a specified bandwidth of 51 kHz, such that SNR_in and SNR_out were determined separately at different reference points. Corresponding values for F and NF can then be calculated using the above equations. All measurements were performed during gradient-echo imaging (repetition time=2000 ms, echo time=2.46 ms, flip angle=60°, slice thickness=5 mm, field-of-view=220 mm) of a uniform poly-acrylic acid phantom (conductivity=~0.47 S/m; relative permittivity=~80), with 4-channel pTx in quadrature mode (90⁰ phase offset between RF pulses submitted on successive channels).

Results

SNR measurement results are reported in Fig. 2. For additional context, Fig. 3 displays a sample screen capture on the spectrum analyzer at the input of the signal modulator and Fig. 4 displays sample screen captures on the spectrum analyzer at the output of each RF power amplifier (RFPA).

Discussion and Conclusion

Using the above equations, the NF of each pTx channel can be calculated, however on closer inspection, it is believed that the full signal spectrum was not captured in Fig. 3 and requires further analysis. Small-signal measurement can be difficult. Nonetheless, Fig. 4 provided a number of useful observations; overall system noise appears to have increased and channel-to-channel variation in signal and noise power was observed. This channel-to-channel system behaviour is likely to change for different pTx-MRI imaging conditions; a challenge for benchtop procedures to replicate. Therefore, this measurement setup offers many advantages. For example, a real-time calibration procedure can be implemented to adjust output signals (e.g. for undesired channel-to-channel variation) for optimal imaging, similar to pre-scan routines on commercial systems. Other uses include real-time power and specific absorption rate monitoring and noise characterization of specific devices (e.g. signal modulator or RFPA). In conclusion, a useful multi-purpose measurement setup was used to further characterize our 4-channel pTx-MRI system. Additional system measurements are planned prior to the conference proceedings.

Acknowledgements

Special thanks to Fred Tam for the helpful discussion.