Objective: To provide a semi-automated segmentation and quantification method of the changes in the substantia nigra (SN) in Parkinson’s disease (PD) using neuromelanin (NM)-sensitive MRI in a large multi-centre dataset (Parkinson’s MR imaging repository: PaMIR).

Background: Previous studies investigating small groups of patients (<30) consistently showed reduced NM-related signal in the SN [1-2], which may have the potential to serve as a diagnostic imaging marker of PD. Hence, evidence of this via large-scale multi-centre studies is needed. In addition, manual segmentation of SN is time-consuming and therefore a faster and reliable segmentation method is desirable.

Methods: Participants from five sites (n= 192 PD [66.0±8.9 years; 130 males] and 92 healthy controls (HC) [64.2± 9.9 years; 34 males] underwent NM-MRI at 3T and MDS-UPDRS as part of the multimodal PaMIR study. A user-friendly interface was designed to automatically divide the SN into anterior “aNM-SN” and posterior “pNM-SN”, once a region containing the whole SN (wNM-SN) is determined manually. It also automatically computes 1) the suprathreshold hyperintense volume of the NM-rich region following a previously published approach to correct for sequence- and platform-dependent differences (Figure 1) [1] and 2) the contrast to noise ratios (CNR) of aNM-SN, pNM-SN and wNM-SN vs. background signal. Non-parametric tests were used for between-group comparisons controlled for the midbrain volume, sex, and age.

Results: We found significant loss of NM-rich volume and contrast in PD vs. HC as markers of depigmentation (Figure 2). Additionally, both metrics in pNM-SN were lower than aNM-SN in line with an anterior-posterior pigmentation gradient in PD. [figure1] [figure2]

Conclusions: Our results provide further evidence of NM volume and contrast abnormalities in PD, reflecting pigmented cell loss in the SN and indicate the validity of the current semi-automated SN segmentation and quantification approach. This paves the way for the assessment of using volumetric and contrast measures of the SN as a potential multi-site diagnostic and severity marker in our large-scale NM study. Our ongoing work focuses on quality assurance, interrater reliability and using the follow-up data to clarify the role of NM-sensitive MRI in tracking the progression of disease.

References: 1. Schwarz ST., Xing Y. Tomar P, Bajaj N, Auer, DP. In vivo assessment of brainstem depigmentation in Parkinson disease: Potential as a severity marker for multicenter studies Radiology, 2017. 2. Schwarz ST., et al. Mov. Disord., 2011. 26: p.1633–8.

« Back to 2018 International Congress

MDS Abstracts - https://www.mdsabstracts.org/abstract/semi-automated-segmentation-and-quantification-of-the-substantia-nigra-depigmentation-in-parkinsons-disease-a-multicentre-case-control-mri-study-in-284-participants/