Photon attenuation correction in whole-body PET/MRI using tissue classification

A challenge for hybrid PET/MRI imaging is the determination of the photon attenuation map to correct the acquired PET data. Segmentation of adapted MRI sequence data in four tissue classes (background, lungs, fat, and soft tissue) is proposed and evaluated using data from 35 PET/CT whole-body oncological examinations as well as two PET/CT and MRI examinations from the same patients. When reconstructing with the segmented attenuation map, small differences in tumor uptake were observed mainly for osseous lesions, but did not change the clinical interpretation in any case. The method appears to be viable for clinical use.