MR-based PET attenuation correction using a combined ultrashort echo time/multi-echo Dixon acquisition

Area of Science:

• Medical Imaging
• Biophysics
• Radiological Sciences

Background:

• Accurate attenuation correction is crucial for quantitative positron emission tomography (PET) imaging.
• Current methods often rely on computed tomography (CT) or less precise MR-based techniques.
• Developing MR-only attenuation correction methods for PET/MR scanners is an active area of research.

Purpose of the Study:

• To develop and validate a novel magnetic resonance (MR)-based method for estimating continuous linear attenuation coefficients (LACs) for PET attenuation correction.
• To utilize a physical compartmental model with ultrashort echo time (UTE) and multi-echo Dixon (mUTE) acquisitions for simultaneous water, fat, and short T2 component signal estimation.

Main Methods:

• A 3D mUTE sequence was developed, integrating 3D UTE with multi-echo Dixon imaging using sparse radial trajectories for accelerated acquisition.
• A k-space trajectory mapping sequence was employed to correct for radial trajectory errors during image reconstruction.
• A physical compartmental model fitted multi-echo MR signals to derive component fractions (water, fat, bone) for voxel-wise LAC estimation.

Main Results:

• Phantom and in vivo human studies demonstrated the method's performance against CT-derived LACs.
• The proposed mUTE method showed higher correlation and similarity with reference CT data compared to Dixon- and atlas-based MRAC methods.
• PET images reconstructed using the proposed method exhibited relative absolute errors below 5% across multiple brain regions and white/gray matter in all subjects.

Conclusions:

• The developed mUTE method successfully generates subject-specific, continuous LAC maps.
• This technique offers a viable solution for accurate PET attenuation correction in integrated PET/MR systems.
• The method enhances the quantitative accuracy of PET imaging within a hybrid PET/MR scanner.