Implications of pulse sequence in structural imaging of trabecular bone

Area of Science:

• Biomedical Imaging
• Radiology
• Bone Imaging

Background:

• Trabecular bone (TB) imaging is crucial for assessing skeletal health.
• Accurate measurement of TB structure is essential for diagnosing and monitoring bone diseases.
• Various MRI sequences exist, each with unique properties for soft tissue and bone imaging.

Purpose of the Study:

• To compare the signal-to-noise ratio (SNR) and image characteristics of four 3D MRI sequences for trabecular bone imaging.
• To evaluate fast large-angle spin echo (FLASE), steady-state free precession (SSFP), gradient-recalled acquisition in steady state (GRASS), and spoiled GRASS (SPGR) for their utility in TB structural analysis.

Main Methods:

• Theoretical and experimental SNR analysis was performed on phantoms, bone specimens, and in vivo.
• Trabecular thickness and bone volume fraction were compared across sequences, with attention to echo time (TE) dependence.
• Simulations modeled trabeculae as cylinders to investigate intra-voxel spin-dephasing in SSFP and GRASS.

Main Results:

• FLASE demonstrated the highest SNR, followed by SSFP, GRASS, and SPGR, within hardware constraints.
• Gradient-echo sequences (GRASS, SPGR) resulted in 10-45% greater trabecular thickness compared to FLASE.
• SSFP images provided better delineation of partial-volume trabeculae, but simulations revealed artifactual thickening due to off-resonance effects in SSFP and GRASS.

Conclusions:

• FLASE offers the highest SNR and is robust against susceptibility dephasing for TB imaging.
• While SSFP provides better SNR than GRASS, its practical use may be limited by off-resonance effects and duty cycle constraints.
• Gradient-echo sequences may overestimate trabecular thickness due to intra-voxel dephasing, necessitating careful interpretation.