A compressed sensing accelerated radial MS-CAIPIRINHA technique for extended anatomical coverage in myocardial perfusion studies on PET/MR systems

2019 
Abstract Purpose Simultaneous acquisition of myocardial first-pass perfusion MRI and 18F-FDG PET viability imaging on integrated whole-body PET/MR hybrid systems synergistically delivers both functional and metabolic information on the tissue state. While PET viability scans are inherently three-dimensional, conventional MR myocardial perfusion imaging is typically performed using only three short-axis slices with a temporal resolution of one RR-interval. To improve the integrated diagnostics, an acquisition and image reconstruction method based on “Multi-Slice Controlled Aliasing In Parallel Imaging Results IN Higher Acceleration (MS-CAIPIRINHA)” was developed extending anatomical coverage for MR perfusion imaging to six short-axis slices per RR-interval. Methods An ECG-gated radial TurboFLASH MR pulse sequence with dual band excitation was implemented on an integrated whole-body PET/MR system and a model-based reconstruction technique was developed to fully reconstruct the undersampled CAIPIRINHA acquisitions. An 18F-FDG viability PET scan was performed simultaneously to the MR protocol, additionally complemented by a late enhancement MRI acquisition (LGE). Results and conclusion The developed imaging technique was tested in five patients with known collateralized coronary total occlusions, resulting in improved characterization of perfusion across areas of decreased tissue viability as indicated by the simultaneously determined 18F-FDG uptake. While conventional MR perfusion with only three slice positions was occasionally missing substantial parts of the viable area, the new approach achieved LV coverage only slightly inferior to LGE imaging and therefore better comparable to PET results. The quality of first-pass enhancement curves was comparable between conventional and radial MS-CAIPIRINHA-based acquisitions.
    • Correction
    • Source
    • Cite
    • Save
    • Machine Reading By IdeaReader
    34
    References
    2
    Citations
    NaN
    KQI
    []