Andrew Leynes1, Jaewon Yang1, Dattesh Shanbhag2, Sandeep Kaushik2, Florian Wiesinger3, Youngho Seo1, Thomas Hope1, and Peder Larson1
1University of California San Francisco, San Francisco, CA, United States, 2GE Global Research, Bangalore, India, 3GE Global Research, Munich, Germany
The current clinical standard for extracranial MR-based
attenuation correction (MRAC) on hybrid PET/MRI systems is the use of a
Dixon-type sequence to generate a continuous-value fat-water map. The exclusion
of bone in Dixon MRAC contributes a clinically significant amount of
underestimation in bone lesion uptake. Bone information from a zero echo-time
(ZTE) MRI pulse sequence is combined with the Dixon MRAC to produce a hybrid
ZTE-Dixon MRAC. The work demonstrates, using PET/MR patient data, that the
Dixon MRAC (neglecting bone) is underestimating bone lesion uptake by a
clinically significant amount (>10%) when compared to the hybrid MRAC
(including bone).
