Abstract
1398
Objectives To evaluate the utility of a prolonged single bed liver acquisition incorporated into a whole body PET/MRI protocol.
Methods 8 patients, 4 imaged with 18F-FDG and 4 imaged with 68Ga-DOTA-TOC were included. Gadoxetate was used for all studies. Imaging was performed on a 3.0T time-of-flight PET/MRI (GE Healthcare, Waukesha, WI). A single 15-minute bed position was acquired in additional to a multi-bed position whole body PET/MRI using a 135 second acquisition at each bed position. PET data was reconstructed using respiratory binning (6 bins total) and respiratory compensation (“Q-static”, one image encompassing 40% of the RR at end expiration). MR imaging included breath-hold post-contrast sequences as well as respiratory navigated hepatobiliary phase (HBP) images.
Results 17 lesions were visualized in the 8 patients that were avid on either DOTA-TOC or FDG PET imaging. SUVmax values from whole body PET/CT and PET/MRI imaging resulted in similar results (p=0.28). Qstatic reconstruction resulted in SUVmax values 35% higher than the single bed 15 minute acquisition (average SUVmax difference of 5.2, p < 0.001). The 15-minute bed position acquisition had a qualitatively better appearance than the respiratory compensated technique. All breath-hold techniques misregistered with acquired PET data, while navigated HBP was robust and accurately registered to PET imaging data.
Conclusions Leveraging a single bed liver acquisition is beneficial for PET/MRI, allowing for respiratory compensation techniques. Respiratory compensation resulted in an increased SUVmax either due to increased noise or improved respiratory gating. Respiratory compensation is particularly relevant in the setting of PET/MRI multiple breath-holds are performed. Additionally, HBP imaging is useful as it can be acquired using navigated sequences that accurately fuse.
Research Support Grant support from GE Healthcare.