Compensation and evaluation of metal-implant-induced artifacts in MR-based attenuation correction

Ziel/Aim:

MR-based attenuation correction (MRAC) using tissue type segmentation suffers from metal-implant-induced artifacts (MAs) in the underlying MR scan. We propose an improved MRAC segmentation algorithm compensating MAs in MR-based attenuation maps (MRMaps) and evaluate the quantitative influence of these artifacts on the reconstructed PET images.

Methodik/Methods:

MA-based cavities in MRMaps are filled by the developed algorithm using a delineation of the patient's body contour that is derived from the PET emission image . PET emission data of 11 patients with MAs (endoprotheses, sternal cerlages) examined in a Philips Ingenuity PET/MR were reconstructed with the vendor-provided method for attenuation correction (MRMap1,PET1) and additionally with our MRAC algorithm (MRMap2,PET2). Both types of MRMaps were visually inspected for segmentation errors. The segmentation errors in MRMap1 were classified into four classes (L1 and L2 artifacts inside the lung, and B1 and B2 artifacts inside the remaining body depending on the assigned attenuation coefficients). The average relative SUV differences (eps_relav) between PET1 and PET2 in all regions showing erroneous attenuation coefficients in MRMap1 were calculated.

Ergebnisse/Results:

MRMap1 showed erroneous attenuation coefficients in regions near metal implants and inside the patient's lung in all 11 patients. In MRMap2, all MA regions were filled with the soft tissue attenuation coefficient and the lung was correctly segmented in all cases. MRMap2 showed small residual segmentation errors in 8 patients. eps_relav was (mean+-sd): (-57+-1)% in B1, (-43+-4)% in B2, (19+-19)% in L1 and (128+-50)% in L2 regions.

Schlussfolgerungen/Conclusions:

MAs severely disturb MR-based attenuation correction and SUV quantification in PET/MR. The developed algorithm is able to compensate for these artifacts, improves SUV quantification accuracy distinctly and is suitable for clinical application.