MRI-Based Attenuation Correction for Emission Tomography using Ultrashort Echo Time Sequences

Abstract

One of the difficulties in PET-MRI and SPECT-MRI is performing attenuation correction, which is necessary for obtaining quantitative PET or SPECT images. In PET-MRI and SPECT-MRI, the attenuation map should ideally be derived from the MRI image. This is difficult as there is no direct correlation between MRI image intensity and the attenuation coefficient of tissue. In conventional MRI images, there is also no contrast between air, lung, and cortical bone. This is caused by the low proton density and very short T2 relaxation time of these tissues. Conventional MRI images therefore do not allow discrimination of air, lung, and cortical bone, although these tissues have strongly different attenuation coefficients. Ultrashort echo time (UTE) sequences acquire the MRI signal very quickly after RF excitation, before the signal in lung and cortical bone has disappeared. The discrimination of air, lung, and cortical bone is therefore possible based on UTE-MRI images. Different methods for performing UTE-MRI-based attenuation correction have been proposed. Most methods have only been tested on human brain PET data. Preliminary investigations into whole-body UTE-MRI-based attenuation correction have also been done. The feasibility of using a UTE-MRI-based attenuation correction for micro-SPECT of rats has also been investigated. Further investigations are still needed to improve these methods and expand their applicability. Keywords: UTE; PET-MRI; SPECT-MRI; attenuation correction; quantification