Introduction

Hyperpolarized ¹³C MRI often suffers from an inhomogeneous B₁⁺ transmit field, as it requires the use of a secondary transmit coil system within the proton-only body coil. These field inhomogeneities lead to varying flip angles, making the image signals difficult to interpret in a large field-of-view. The varying spatial distribution of the B₁⁺ field for a given transmit coil can be observed using a variety of methods;^1,2 however, applying these methods during in vivo ¹³C scans is infeasible due to timing constraints. Moreover, referencing a pre-measured B₁⁺ map to minimize variation in an in vivo scan is prone to error, as the field distribution varies with coil loading and positioning changes.

To account for changes in the B₁⁺ distribution that result from coil loading, some form of B₁⁺ calibration is often performed using a ¹³C reference phantom placed near the region of interest. However, this calibration becomes inconsistent in large field-of-view applications as there is a distribution of B₁⁺ fields within the target region.

In this work, a method that rapidly obtains multiple B₁⁺ measurements during in vivo scans from multiple ¹³C reference phantoms is proposed. Placing these reference phantoms strategically within the coil allows for simultaneous calibration of areas that differ in B₁⁺. This provides an opportunity for larger field-of-view imaging than was previously attainable with ¹³C MRI.

Methods

A transmit profile of a flexible ¹³C transmit-receive quadrature coil (Clinical MR Solutions, Brookfield, WI) (Figure 1) was measured using the double angle method. This spatial map of the transverse component of the B₁ field, B₁⁺, is used to determine where to place ¹³C phantoms for calibration. Phantoms are placed near anatomical regions of interest that show different B₁⁺ in the transmit profile.

A short TR pulse-acquire sequence with linearly increasing flip angles was designed to rapidly measure the B₁⁺ at the reference phantoms. With TR = 310 ms and 30 iterations, the total scan time for the measurement is 9 s. The signal evolution from the train of excitations for each phantom creates a characteristic signal response for the particular B₁⁺ at the corresponding location. These values are discrete samples of the B₁⁺ field that is unique to each subject's positioning in the scanner.

The characteristic response is fit to a non-steady state model which determines the measured B₁⁺ of the phantom and outputs the power adjustment required to achieve the desired B₁⁺. These adjustment values will be input as control variables for the sequence, such that the power adjustment occur automatically. This will minimize the variability in the B₁⁺ field that occurs towards the edges of the coil.

Results & Discussion

Two axial slices of the pre-computed B₁⁺ map used to determine the location of the reference phantoms is shown in Figure 2. The relative B₁⁺ distribution of these slices, one central to the coil and one near the edge, show the variability that occurs towards the coil perimeter. Such variability supports the need for a method that can correct for this in large field-of-view imaging.

The rapid (9 s) B₁⁺ measurement sequence was tested during human scans (n=14) using a ¹³C-urea phantom, showing that this method is capable of computing the B₁⁺ field accurately within ±1μT. A comparison of B₁⁺ measurements found using this method to manual measurements for each subject is shown in Figure 3. A sample calibration plot showing the non-steady state model fit to the signal responses of two reference phantoms obtained simultaneously is shown in Figure 4. This validates the ability of this method to acquire multiple point B₁⁺ measurements simultaneously.

In the future, this method will be included in the scan protocol for hyperpolarized ¹³C MRI of the heart and kidneys to enable rapid calibration for slice-specific B₁⁺ scaling.

Conclusion

A method was developed and validated for acquisition of rapid (9 s) B₁⁺ field measurements from multiple ¹³C reference phantoms used during human ¹³C metabolic imaging. The method shown here can be used to minimize the variability in flip angle due to inhomogeneous B₁⁺ in large field-of-view images, such as those containing the thoracic and upper abdominal cavities.