MRI-based attenuation correction for PET/MRI using ultrashort echo time sequences.

Area of Science:

• Medical Imaging
• Radiology
• Biophysics

Background:

• Accurate attenuation correction is crucial for PET/MRI, but conventional MRI methods struggle with bone detection, necessitating anatomical assumptions.
• These assumptions limit clinical applicability in patients with anatomical abnormalities.

Purpose of the Study:

• To develop and validate a novel MRI-based method for generating accurate attenuation maps for PET/MRI.
• To overcome limitations of existing methods by not relying on anatomical preconditions.

Main Methods:

• Utilized transverse relaxation rate (T2*) from ultrashort echo time (UTE) MRI sequences for tissue classification (bone, soft tissue, air).
• Assigned linear attenuation coefficients based on tissue class without anatomical assumptions.
• Validated accuracy against CT scans and evaluated PET image reconstruction using MRI-derived attenuation maps.

Main Results:

• Achieved 90% voxel-wise accuracy in tissue classification on a piglet phantom.
• Demonstrated clinically acceptable average errors of approximately 5% in human brain PET imaging compared to CT-based methods.
• PET reconstructions using the novel MRI-derived attenuation maps showed comparable results to CT-based corrections.

Conclusions:

• The T2*-derived attenuation map from UTE MRI is feasible for PET/MRI attenuation correction.
• The proposed method provides clinically acceptable accuracy and is robust to anatomical variations.
• This technique holds promise for improved PET/MRI applications, especially in patients with complex anatomy.