Abstract
1830
Objectives MR-based attenuation correction (AC) for cortical bone in neurological studies has recently been improved by employing ultrashort-echo-time (UTE) sequences in combination with various segmentation methods. In this prospective study, we aim to establish a clinically applicable and technically feasible imaging protocol for simultaneous 18F-DOPA brain PET/MR evaluation of Parkinson’s disease (PD) in comparison to PET/CT in terms of quantification accuracy and diagnostic efficacy.
Methods From April to June 2015, 20 scans (10 PET/MR, 10 PET/CT) were performed in 10 patients: 5 PD (M: 1, F: 4; age range: 43~71 y, median: 54 y) and 5 sex & age (range: 41~75 y, median: 53 y) matched normal patients. PET/MR (Biograph mMR) and PET/CT (Biograph mCT) scans were sequentially performed after a single dosage of 18F-DOPA injection (~10mCi). Pre-medication with Carbidopa (200 mg) was given orally 60 min before 18F-DOPA injection. Brain PET/MR began at 30 min post injection with a 30-min LIST mode acquisition. Simultaneous 3-Tesla MR sequences included UTE, 3D SAGITTAL T1, AXIAL (AX) T2 TSE, CORONAL T2 FLAIR, AX DWI (B0, 1000), AX susceptibility weighted imaging (SWI). 18F-DOPA PET/CT for PD was performed at 90 min post injection with a 10-min acquisition. Regions of putamen, cerebellum (CE) and occipital lobe (OL) were drawn automatically on PET/MR and PET/CT using MIMNeuro. Quantification accuracy of PET/MR was evaluated by comparing SUV in putamen with PET/CT in 5 normal patients (18F-DOPA uptake in putamen reaches plateau in less than 30 min). 18F-DOPA brain PET/MR and PET/CT were visually assessed by 2 nuclear medicine physicians, each independently then in consensus. MR images were interpreted by radiologists. The diagnostic parameter for PD on LIST mode PET/MR was defined as the influx index (K) of putamen with OL as the reference region by Patlak analysis. For PET/CT, SUV ratio of putamen-to-CE (SUVr) was used for the differentiation of PD.
Results Quantitation results of 18F-DOPA brain PET/MR using UTE for AC were acceptable when compared with PET/CT, with the SUV difference in putamen between them ranging from -1.8 to 4.7% and a median of 1.3%. All observers agreed that both hybrid fusion and visual delineation of the putamen from the caudate and surrounding brain tissue is better on the PET/MR than on the PET/CT images. All PD patients showed decreased 18F-DOPA uptake in putamen on both PET/MR and PET/CT, indicating presence of neuronal dopaminergic dysfunction. No evidence of micro-hemorrhage was seen on SWI in all PD patients to suggest secondary PD. By Patlak analysis on PET/MR dynamic data, K value was significantly lower in PD patients than controls (0.0030±0.0011 vs 0.0069±0.0006, p<<0.05). By PET/CT, SUVr was also significantly lower in PD than normal patients (1.85±0.22 vs 2.91±0.21, p<<0.05). Strong statistical correlation was found between the K value of PET/MR and SUVr of PET/CT by 2-tailed Pearson correlation analysis (coefficient=0.967, p<<0.05).
Conclusions UTE-based AC was able to achieve comparable quantification accuracy for 18F-DOPA brain PET/MR when compared with PET/CT. PET/MR could provide better image quality on brain tissue differentiation with additional advantage for exclusion of micro-hemorrhage. LIST mode 18F-DOPA brain PET/MR at 30 min post injection might provide sufficient data for evaluation of Parkinson’s disease with reasonably high diagnostic accuracy and shortened scanning time.