Synchronisation strategies in T2-weighted MR imaging for detection of liver lesions: Application on a nude mouse model

Abstract

Aim: The objective of this work was to propose original synchronisation strategies based on T2-weighted sequence performed on a small animal MRI spectrometer in order to improve the image contrast and detect mouse liver lesions at high magnetic field.

Materials and methods: The experiments were performed in vivo at 7T using a 32 mm inner diameter cylindrical volumetric coil for both RF emission and reception. A sensitive pressure sensor was used to detect external movements due to both respiration and heart beats. The pressure sensor was interfaced with a commercial ECG Trigger Unit to use dedicated functionalities (trigger levels, delays and window). To enable T2-weighted imaging with minimised T1 effects, an acquisition strategy with controlled TR spanning over several respiratory cycles was developed. With this strategy, the slices were acquired over several respiratory periods.

Results: The acquisition, performed over several respiratory periods, enables a longer TR than the typical mouse respiratory period. The image contrast is controllable and independent of the respiratory period. The heavily T2-weighted images obtained with the developed strategy allow better visualisation of lesions at high magnetic field. Moreover, double respiratory and cardiac synchronisation, based on a unique sensitive pressure sensor, improves image quality with less motion artifacts, especially in the ventral liver region. The total slice number is independent of respiratory period and thin slices can be acquired to cover the whole liver.

Conclusion: The developed strategy enables high quality pure T2-weighted imaging with minimal motion artifacts. This strategy improves T2-weighted image contrast and quality, especially at high magnetic field, on animals with short respiratory periods. The strategy was demonstrated using a mouse model of liver lesions at 7T. This protocol could be used to carry out a longitudinal follow-up.

Keywords: T2-weighted contrast image; high-field MRI; mice liver; motion artifacts; synchronisation.

Figure 1

Schematic of connection for sensor and devices used for triggering. An air pillow was connected to a pressure sensor for both respiration and heart motion detection. The pressure sensor was interfaced with the ECG Trigger Unit HSB-T. This device ensures the amplification and filtering of input signal as well as the adjustment of the trigger level, trigger delay and trigger window. The trigger pulse generated by the ECG unit is used by the waveform generator to send a burst of trigger pulses to the MRI console.

Figure 2

Diagrams of different synchronisation and slice acquisition strategies used to obtain T2-weighted contrast images. The trigger pulses sent by the trigger unit to the MRI console as well as the excitation timing of each slice is indicated and illustrated for 9 slices: (a) Conventional triggering strategy for T2-weighted contrast images. The repetition time is equal to the respiratory period with TR = T_resp. All slices are acquired within one respiratory cycle (non interleaved case); (b) Heavier T2-weighted image contrast is obtained with minimum controlled TR imposed between two slice packages; (c) New acquisition strategy with balanced slice acquisitions. A trigger pulse is generated for every single slice and acquisition of slices is spanned over several respiratory cycles; (d) Dual cardiac and respiratory triggering with balanced acquisition over several respiratory cycles. The time between two consecutive slices is called inter-slice time (T_is).

Figure 3

Example of the setup for dual cardiac-respiratory triggering: (a) Signal from the pressure sensor with acquisition window enabled between two respiratory cycles; (b) Trigger pulses sent to MRI console based on cardiac cycle.

Figure 4

Images of mice liver obtained at D21: (a) without any synchronisation with TR ≈ 1.5 s; (b) with respiratory triggering and slice excitation spanned over three respiratory cycles (TR ≈ 3 × T_resp ≈ 4.5 s).

Figure 5

Images of mice liver (T_resp = 1.5 s; 12 slices; slice thickness 1.5 mm) obtained on normal specimen: (a) without synchronisation (TR = 1.5 s; TA = 5 min 20 s); (b) with conventional respiratory strategy and a long controlled TR (TR ≈ 3 s; T_is = 90 ms; TA = 9 min); (c) with respiratory triggering strategy with balanced acquisitions over several respiratory periods (TR ≈ 3 × T_resp = 4.5 s; T_is = 105 ms; TA = 9 min); (d) with dual cardiac and respiratory triggering with balanced slice acquisitions (TR = 4.5 s; T_is ≈ 200 ms; TA = 17 min).

Figure 6

Multi-echo images (T_resp ≈ 2 s; 36 slices; 0.5 mm slice thickness) and slice excitation spanned over several respiratory cycles obtained at D21: (a, c, e) with respiratory triggering (TR ≈ 3 × T_resp; T_is = 105 ms; TA = 21 min) for TE = 20, 40, 60 ms; (b, d, f) with cardiac and respiratory triggering (TR ≈ 6 × T_resp; T_is = 155 ms; TA = 35 min) for TE = 20, 40, 60 ms.