OBJECTIVE

Evaluate the correlation between a scalable, automated neuromelanin-sensitive MRI (NM-MRI) approach and I-123 ioflupane single photon emission computed tomography (SPECT) in the nigrostriatal system in Parkinson’s disease (PD).

BACKGROUND

123-I-ioflupane SPECT (DaTScan, GE) is used to assess the nigrostriatal system in the evaluation of PD, however it has limitations due to cost and degraded performance with some medications. NM-MRI detects contrast in SNc that colocalizes with melanized dopamine neurons in radiologic-pathologic correlation studies [1,2], and it robustly detects degeneration of the nigrostriatal dopamine system [3].

Published NM-MRI approaches continue to rely upon manual image processing methods that introduce operator dependent error and that cannot be scaled for applications in large multi-site studies or clinical use [4,5]. Furthermore, the brainstem reference regions in current NM-MRI methods are small, may include outlier voxels, and often have a non-Gaussian signal intensity distribution, which makes interpretation of results difficult [6]. This study evaluates the correlation between DaTScan assessment of presynaptic dopamine transporter (DAT) uptake and an automated, scalable NM-MRI approach for substantia nigra pars compacta (SNc) volume measurement in PD that addresses these issues.

METHODS

PD (n=23) patients meeting MDS diagnostic criteria for PD [8] were recruited in an academic movement disorders clinic and completed a 3 T MRI scan (Siemens Prisma Fit) and an I-123 ioflupane SPECT scan (DaTScan, GE Healthcare). MRI included a magnetization-prepared 2D gradient echo NM-MRI sequence and a T1 structural sequence. SNc volumes were determined using a thresholding method with high scan-rescan reproducibility [7]. A custom algorithm was used for automated image processing with no manual segmentation steps, excluding outlier voxels in the reference region, which allows selection of a larger brainstem reference region as compared to other methods [9] as well as a Gaussian signal intensity distribution in the reference region (Figure, Panel A). Patients were clinically assessed with the Movement Disorders Society Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) and measures of cognitive and sleep phenomenology.

RESULTS

Baseline demographics and clinical characteristics of the PD patient group were determined (Table, Persuasive Data). It was observed that age correlated with DaTScan striatal binding ratios (SBRs) in putamen and caudate, but age did not correlate with SNc volume measured with NM-MRI (Figure, Persuasive Data). Using partial correlation, controlling for age and sex, a statistically significant correlation was observed between SNc volume, measured with NM-MRI, and DaTScan SBRs in the putamen and its anterior subregion. Detailed results are shown in the figure, Panel B.     

CONCLUSIONS

NM-MRI measures of SNc volume obtained with an automated approach that addresses limitations of existing methods are significantly correlated with DaTScan SBRs in the putamen, which undergoes denervation in PD. Establishment of robust and reproducible methods that can be used at scale is necessary for clinical and translational application of imaging tools for Parkinson’s disease and related conditions. These results support the continued development of NM-MRI as a cost-effective alternative to DaTScan for assessment of the nigrostriatal system.   

Image/Figure Caption:

Author