MR-based attenuation correction for torso-PET/MR imaging: pitfalls in mapping MR to CT data.

Area of Science:

• Medical Imaging
• Radiology
• Nuclear Medicine

Background:

• Combined PET/MRI systems require accurate MR-based attenuation correction (AC).
• Validation of MR-AC is essential for clinical PET/MRI data sets.

Purpose of the Study:

• To propose and validate a toolbox for MR-based AC in clinical PET/MRI.
• To assess the accuracy of MR-based AC compared to CT-based AC.

Main Methods:

• Torso scans from ten patients were acquired using PET/CT and 1.5-T MRI.
• MR-based attenuation data were derived from CT post-co-registration and histogram matching.
• PET images were reconstructed using both CT-AC (PET(CT)) and MR-AC (PET(MRI)).

Main Results:

• MR-CT histogram matching resulted in a maximum 12% voxel intensity difference in attenuation images.
• Mean differences between PET(MRI) and PET(CT) reconstructions reached a maximum of 19%.
• Lesion-to-background ratios were similar, except in the lung, where misregistration and intensity transformation biased PET(MRI).

Conclusions:

• The developed toolbox aids in studying MR-AC pitfalls.
• Co-registration accuracy and pixel value transformation are critical determinants of PET(MRI) accuracy.