TY - JOUR T1 - Respiratory compensated PET with DOTA-TOC PET/MRI using motion registered and gated reconstruction JF - Journal of Nuclear Medicine JO - J Nucl Med SP - 1987 LP - 1987 VL - 57 IS - supplement 2 AU - Dora Tao AU - Miguel Hernandez Pampaloni AU - Vahid Ravanfar AU - Eric Nakakura AU - Emily Bergsland AU - Thomas Hope Y1 - 2016/05/01 UR - http://jnm.snmjournals.org/content/57/supplement_2/1987.abstract N2 - 1987Objectives Previous approaches to respiratory compensation using PET/MRI have gated data acquired during simultaneous acquisition and limited data to that acquired during end expiration. Although this results in removal of respiratory blurring and ghosting, there is an increase in image noise due to the decreased amount of included PET data in the reconstruction. In this work we have evaluated the utility of registering gated PET data to a single phase, allowing for more PET data to be included in the final reconstruction.Methods : 6 patients were imaged after Ga-68 DOTA-TOC injection using a 3.0T time-of-flight PET/MRI (GE Healthcare, Waukesha, WI). Data was acquired during three minute bed positions during the whole body acquisition and during a 15 minute bed position simultaneous with dedicated liver imaging. Both PET data sets were reconstructed in three ways. The first used all PET data without any respiratory compensation (ungated). The second used respiratory compensation whereby only breath holds between 2 and 10 seconds were included and only the last 40% of the respiration was included corresponding to end expiration (gated). The third reconstruction used data from breath holds between 2 and 10 seconds, but was reconstructed into six bins for the duration of the respiratory cycle. Subsequently images from each of the six bins were registered to the end expiratory phase, which was manually selected, using a 3D motion correction algorithm (registered). This results in a single PET data set that includes PET data from throughout the respiratory cycle, but registered to one phase in the respiratory cycle. VOIs were placed over lesions in the liver and SUVmax were measured. Additionally noise was calculated using an ROI placed within the liver (SDliver / SUVavg-liver).Results SUVmax was lowest for the ungated data, and highest for the gated data, with registered data noise levels between gated and ungated (WB-ungated: 46.1, WB-gated: 43.9 WB-registered: 48.0; liver-ungated: 34.7, liver-gated: 49.5, liver-registered: 32.4). Noise levels were lowest for the ungated data, and highest for the gated data, with registered data noise levels between gated and ungated (WB-ungated: 0.19, WB-gated: 0.32, WB-registered: 0.29; liver-ungated: 0.08, liver-gated: 0.13, liver-registered: 0.10).Conclusions By registering respiratory gated PET data it is possible to recover noise typically created with respiratory compensation techniques that discard data at different points in the respiratory cycle. Further complete evaluation of this technique needs to be performed in order to understand its reproducibility Figure 1: Top row reconstructions are using the three minute data sets from the whole body acquisition and the bottom row are from 15 minute dedicated liver acquisition. Gated reconstruction (middle) results in removal of the majority of motion, although noise levels increase. Registered reconstruction have improvement in noise characteristics although have some respiratory motion. $$graphic_01DED769-C733-4736-B7C0-51AE37AB172B$$ ER -