Synopsis

Keywords: Data Acquisition, Pulse Sequence Design

MRI serves as a non-invasive way of assessing the degree of iron overload. This is done via T2*-mapping commonly performed with muti-echo gradient echo sequences. This approach fails in cases with extreme iron overload where the relaxation times drop down to submillisecond range. A method allowing ultra-short T2* mapping is simulated, employing a temporal acquisition offset in a spin-echo pulse sequence. Two acquisitions are suggested: one – before the spin-echo centre, and another – after, resulting in two different T2*-weighted images, acquired with half-line radial encoding. This method allows obtaining T2*-weighted images with weightings starting effectively at zero echo time.

Introduction

Iron overload is a frequent consequence of a number of systemic disorders such as anemia or thalassemia. Iron overload occurs in different body parts, including spleen, heart, and liver. MRI often serves as a non-invasive way of assessing the degree of iron overload in the selected organ. This is done via T₂*-mapping commonly performed with muti-echo gradient echo sequences¹. This approach fails in the cases with extreme iron overload, where the T₂* relaxation times drop down to single- or sub-millisecond range². In those cases common combinations of radiofrequency (RF) and gradient pulses can take up to a few milliseconds to produce the first detectable signal. Therefore alternative T₂*-mapping methods are needed for extra-short T₂* times mapping.

Theory

A method allowing ultra-short T₂* times mapping can be constructed by employing an acquisition offset in a spin-echo pulse sequence³. While the spin-echo amplitude is T₂-weighted, the echo mirrors the free induction decay and thus follows the T₂* decay, the signal is thus proportional to
$$e^{-\frac{TE}{T_2}}e^{-\frac{|TE_{off}|}{T_2^*}}$$
where TE_off is the time offset from the centre of the spin-echo.
The signal maximum occurs in the centre of the spin-echo far from any gradient switching and RF-pulses contrary to the classic gradient echo sequence, where the signal maximum occurs during or right after the RF-pulse. Two acquisitions can be fit under the envelope of the spin-echo (one – before, and another – after the echo maximum), resulting in two differently T2*-weighted images. A half-line radial encoding (Fig. 1, B) is suggested as an encoding scheme. This method allows acquiring T₂*-weighted images with weightings starting effectively at TE_off=0.

Simulation

The sequence suggested in Fig. 1,B was tested here via Bloch simulation. The source T₂* distribution for the simulations was derived from an image out of the Combined Healthy Abdominal Organ Segmentation (CHAOS) dataset⁴. The image intensity values were used after scaling as the source T₂* values. The scaling was performed to provide mean T₂* = 3 ms. over the image. Additional scaling was applied to the liver and spleen, so that the mean liver T₂* became 1 ms and the mean spleen T₂* bacame 0.5 ms (Fig. 2). The proton density was assumed to be uniformly distributed, as well as the T₁ and the T₂ relaxation time, the latter were both considered infinite for the simulation purposes.
Excitation and refocusing in the SE-UTE sequence was performed with the 3-lobe sinc-shaped pulses, with flip angles of 90 and 180. The duration of the excitation pulse was chosen to be 0.8767 ms (the value found to provide shortest excitation and slice-selection gradient refocusing in the gradient echo sequence), the duration of the refocusing pulse was 0.25 ms. Gradient system parameters were selected as maximum gradient value of 80 mT m^-1, gradient slew rate of 200 T m^-1 s. The encoding parameters were selected to correspond to the CHAOS dataset imaging parameters (i.e., 435×435 mm² imaging FoV, 8 mm thick slice, bandwidth of 523 Hz per pixel). The 256×256 points Cartesian encoding from the CHAOS dataset was substituted with half-line radial encoding with 128 points per spoke and 256*π ≈ 804 spokes uniformly distributed from 0 to 2π. The simulation was performed with TE_{spin echo} = 10 ms, a list of offset TE₁ from 0 to 4 ms with 0.2 ms step, and a list of offset TE₂ from 0.1 to 4.1 ms with 0.2 ms step. The minimum time required for the dephasing and refocusing gradient blip between TE₁ and TE₂ for the desired imaging parameters was found to be 72.3 µs, therefore a 0.1 ms difference between the two acquisitions was feasible. The T₂* relaxation was simulated by splitting each voxel into 301 Lorentzially distributed frequency isochromats with the full with at half maximum corresponding the desired voxel T₂*.
The simulated data from all the isochromats and voxels was summed to provide the complex left- and right-sided half-echo signals, corresponding to offset echo times TE₁ and TE₂ respectively. These were gridded to the 256×256 Cartesian k-space using convolution with a Kaiser-Bessel function kernel⁵. The resulting k-spaces were Fourier-transformed into images and each pixel was fit with a single-exponential decay function to obtain three T₂*-maps. The three maps were ones calculated from only the images acquired at TE₁, only images acquired at TE₂, and maps acquired from the combined TE₁ and TE₂ images. The liver and the spleen relaxation times were extracted through the existing CHAOS segmentation and their distributions compared to the ground truth map.

Results

The T₂*-weighted images were successfully reconstructed from the simulated radial encoding data without significant artifacts: only small intensity drop on the image edges was present, related to the radial encoding and gridding procedures (Fig. 3). The T₂* decays in all three fits (TE₁, TE₂ and combined) were single-exponential and were able to provide millisecond-order and submillisecond order T₂*-mapping (Fig. 2 and Fig. 4).

Conclusion

A spin-echo pulse sequence with two off-center semi-echo acquisitions was proposed here for ultra-short T₂* measurements. The abdominal image acquisition was simulated with the proposed sequence and data for T₂* relaxation measurements was acquired. T₂* relaxation maps were calculated and were found to largely correspond to the simulation ground truth.

Acknowledgements

The work was supported by the Ministry of Science and Higher Education of the Russian Federation (075-15-2021-592).