Characterization of the FDOPA Uptake of the Brainstem and Diencephalon in Parkinsonian Syndromes by PET/MRI

Complete text of publication follows: Aim/Introduction: FDOPA PET/CT is commonly used for the investigation of dopaminergic denervation, the diagnosis being based on striatal uptake. But other brain structures, including monoaminergic neurons of the brainstem and diencephalon, metabolize FDOPA and are likely to be affected. Using the values automatically generated by a homemade PET / MRI atlas, we looked for alterations in the FDOPA uptake in parkinsonian syndromes. We retrospectively selected 120 patients who underwent FDOPA PET/MRI examination, 49 with Parkinson's syndrome according to clinical and scintigraphic data, 71 others considered very unlikely to have this condition (the control group). Materials and Methods: the T1 images of each patient were registered on the T1 images of a model where 28 anatomical regions had been previously identified. The adjustment parameters were then applied to FDOPA images acquired 90 min after injection. In each region VOIs of different sizes were then generated: 1 voxel (1mm³), 27 voxels or a volume adapted to the size of the region to be studied. The mean SUV, alone or compared to that of the other regions, was used to distinguish the 2 groups, by the way of a Student's test. Results: the results were similar, using different sizes of VOIS. In the parkinsonian group, a significant reduction in SUV was observed in the substantia nigra (p = 0.002), amygdala (p = 0.01) and pulvinar (p = 0.03). We did not observe any significant difference in the A8 zone, the reticular or the raphe nuclei. In the supratentorial area, the occipital region was unaffected too, in contrast to the striatum. The best discriminations between parkinsonian and control groups, obtained by combining the activities of the whole regions, were provided by the difference between the SUV of, on the one hand, the putamen and, on the other hand, the A8 area, the hypothalamus, the pulvinar, the substantia nigra and the caudate nucleus, all with p less than 10-20. Conclusion: for the first time in nuclear medicine, we think we have been able to automatically demonstrate the involvement of nuclei of the brainstem and diencephalon in Parkinson's syndrome. This opens perspectives in terms of early detection, staging and differential diagnosis of this disease