Open-Source Magnetic Resonance Imaging with Minimal Echo Time

The minTE repository provides parametrized, radial, short/ultrashort echo time Magnetic Resonance Imaging (MRI) sequence in the Pulseq format [1] along with reconstruction functions. Sequence variations include 2D and 3D partially-rewound RF-spoiled GRE and CODE [2] generated with the PyPulseq [3] library.

Quickstart

Option 1: Local installation

1. Download or clone the repository
2. Set up and activate the virtual environment making sure all packages listed in requirements.txt are installed;
3. Note that if virtual-scanner is not automatically installed by the IDE through requirements.txt, you need to run the following on the command line with the virtual environment activated: pip install git+https://github.com/imr-framework/virtual-scanner.git@2ed303d42038449e28ebe44fa4177541cabe8eb2
4. From the root, run python -m minTE.tests.run_all_unit_tests to ensure all functions work;
5. From the root, run python -m minTE.quickstart.seq_gen_demo to generate and visualize the sequence variations.

Option 2: Google Colab Notebook

Open the demo notebook in Google Colab and follow the instructions to generate and display a sequence file.

Sequences and example parameters

Shared parameters are FOV = 253 mm, 2D slice thickness = 5 mm, flip angle = 10 degrees, and TR = 15 ms.

Sequence type	Dimensions	Half-pulse option	Matrix size	# spokes	TE (ms)	Acq. time (min:sec)
UTE / rGRE	2D	Yes	256	804	0.45 - 0.82	00:12, 00:24
UTE / rGRE	3D	Yes	64	13056	0.12 - 0.66	03:01, 06:32
CODE	3D	No	64	13056	0.32	03:16

Example Usage

# Make a 3D half pulse UTE sequence 
seq, TE, ktraj = write_UTE_3D_rf_spoiled(N=64, FOV=250e-3, slab_thk=253e-3, FA=10, TR=15e-3, ro_asymmetry=0.97,
                                         os_factor=1, rf_type='sinc', rf_dur=0.05e-3, use_half_pulse=True)

# Make a 2D full pulse UTE sequence
seq, TE, ktraj = write_UTE_2D_rf_spoiled(N=256, Nr=804, FOV=253e-3, thk=5e-3, slice_locs=[0],
                                         FA=10, TR=15e-3, ro_asymmetry=0.97, use_half_pulse=False, rf_dur=1e-3,
                                         TE_use=None)
                                         
# Make a CODE sequence
seq, TE, ktraj = make_code_sequence(FOV=253e-3, N=64, TR=100e-3, flip=10, enc_type='3D',
                                    os_factor=1, save_seq=False, rf_type='gauss')

Reconstruction

The functions reconstruct_2d and reconstruct_3d are provided here that use PyNUFFT [4,5] to generate images from non-cartesian k-spaces. The 'cg' option is recommended for image quality.

Images

ACR phantom

Ex vivo procine muscle/bone sample

Feedback

Please communicate all questions and feature suggestions on the Github Issues tab. We appreciate your inputs!

References

1. Layton, Kelvin J., et al. "Pulseq: a rapid and hardware‐independent pulse sequence prototyping framework." Magnetic resonance in medicine 77.4 (2017): 1544-1552.
2. Park, Jang‐Yeon, et al. "Short echo‐time 3D radial gradient‐echo MRI using concurrent dephasing and excitation." Magnetic resonance in medicine 67.2 (2012): 428-436.
3. Ravi, Keerthi, Sairam Geethanath, and John Vaughan. "PyPulseq: A Python Package for MRI Pulse Sequence Design." Journal of Open Source Software 4.42 (2019): 1725.
4. Lin, Jyh-Miin. “Python Non-Uniform Fast Fourier Transform (PyNUFFT): An Accelerated Non-Cartesian MRI Package on a Heterogeneous Platform (CPU/GPU).” Journal of Imaging 4.3 (2018): 51.
5. J.-M. Lin and H.-W. Chung, Pynufft: python non-uniform fast Fourier transform for MRI Building Bridges in Medical Sciences 2017, St John’s College, CB2 1TP Cambridge, UK