A Framework for Predicting X-Nuclei Transmitter Gain Using 1H Signal

Abstract

Commercial human MR scanners are optimised for proton imaging, containing sophisticated prescan algorithms with setting parameters such as RF transmit gain and power. These are not optimal for X-nuclear application and are challenging to apply to hyperpolarised experiments, where the non-renewable magnetisation signal changes during the experiment. We hypothesised that, despite the complex and inherently nonlinear electrodynamic physics underlying coil loading and spatial variation, simple linear regression would be sufficient to accurately predict X-nuclear transmit gain based on concomitantly acquired data from the proton body coil. We collected data across 156 scan visits at two sites as part of ongoing studies investigating sodium, hyperpolarised carbon, and hyperpolarised xenon. We demonstrate that simple linear regression is able to accurately predict sodium, carbon, or xenon transmit gain as a function of position and proton gain, with variation that is less than the intrasubject variability. In conclusion, sites running multinuclear studies may be able to remove the time-consuming need to separately acquire X-nuclear reference power calibration, inferring it from the proton instead.

Trial registration: ClinicalTrials.gov NCT05215938.

Keywords: X-nuclei imaging; carbon; magnetic resonance imaging; radio frequency setting; sodium; xenon.

Figure A1

Phantom and coil placements in carbon transmit gain examinations. Bicarbonate point phantom with no proton loading placed in the centre between the two Helmholtz loop coils (left). Loading setup using head (middle) and body (right) load rings with the head dimethyl silicone ball inside. Distances between the Helmholtz loop coil elements are noted in cm. Three distances were evaluated in the bicarbonate setup (20, 25, and 30 cm).

Figure A2

Tabulated $\left|{ϵ}_r\right|$ and $\left|\sigma \right|$, in SI units, for a variety of human tissues.

Figure 1

Proton and sodium transmit gain with a linear regression model fit.

Figure 2

(A) Multi-organ measured and predicted transmit gain of proton and sodium with indication of anatomies of interest. (B) Circadian measured and predicted transmit gain of proton and sodium in kidney imaging.

Figure 3

(A) Circadian measured and predicted proton and sodium transmit gain on a group level. (B) Actual flip angle spread of proton and sodium circadian measurements.

Figure 4

(A) Proton and xenon transmit gain with a linear regression model fit. (B) Proton and xenon transmit gain variation.

Figure 5

Proton and carbon transmit gain changes according to loading and distance. (A) Transmit gain changes according to changes in distance between the Helmholtz loop coil elements with a bicarbonate point phantom. (B) The transmit gain of proton and carbon in the head and body load with dimethyl silicone phantom. Carbon is shown with and without bicarbonate point phantom adjustment for distance between the Helmholtz loop coil elements in (B). (C) Proton and carbon transmit gain with a linear regression model fit.