Jung Hwan Shin, MD, PhD, Jee-Young Lee, MD, PhD, Yu-Kyeong Kim, MD, PhD, Eun Jin Yoon, PhD, Heejung Kim, PhD, Hyunwoo Nam, MD, PhD and Beomseok Jeon, MD, PhD
From the Department of Neurology (J.H.S., J.-Y.L., H.N.), Seoul Metropolitan Government--Seoul National University Boramae Medical Center and Seoul National University College of Medicine; Department of Nuclear Medicine (Y.-K.K., E.J.Y., H.K.), Seoul Metropolitan Government--Seoul National University Boramae Medical Center; Institute of Radiation Medicine (H.K.), Medical Research Center, Seoul National University; and Department of Neurology (B.J.), Seoul National University Hospital and Seoul National University College of Medicine, South Korea.
Correspondence Dr. Lee or Dr. Yu-Kyeong Kim
Objective To elucidate the role of Parkinson disease (PD)-related brain metabolic patterns as a biomarker in isolated REM sleep behavior disorder (iRBD) for future disease conversion.
Methods This is a prospective cohort study consisting of 30 patients with iRBD, 25 patients with de novo PD with a premorbid history of RBD, 21 patients with longstanding PD on stable treatment, and 24 healthy controls. The iRBD group was longitudinally followed up. All participants underwent 18F-fluorodeoxyglucose (FDG) PET and were evaluated with olfaction, cognition, and the Movement Disorders Society–Unified Parkinson’s Disease Rating Scale (MDS-UPDRS) at baseline. From FDG-PET scans, we derived metabolic patterns from the longstanding PD group (PD-RP) and de novo PD group with RBD (dnPDRBD-RP). Subsequently, we calculated the PD-RP and dnPDRBD-RP scores in patients with iRBD. We validated the metabolic patterns in each PD group and separate iRBD cohort (n = 14).
Results The 2 patterns significantly correlated with each other and were spatially overlapping yet distinct. The MDS-UPDRS motor scores significantly correlated with PD-RP (p = 0.013) but not with dnPDRBD-RP (p = 0.076). In contrast, dnPDRBD-RP correlated with olfaction in butanol threshold test (p = 0.018) in patients with iRBD, but PD-RP did not (p = 0.21). High dnPDRBD-RP in patients with iRBD predicted future phenoconversion with all cutoff ranges from 1.5 to 3 SD of the control value, whereas predictability of PD-RP was only significant in a partial range of cutoff.
Conclusion The dnPDRBD-RP is an efficient neuroimaging biomarker that reflects prodromal features of PD and predicts phenoconversion in iRBD that can be applied individually.
Classification of Evidence This study provides Class IV evidence that a de novo PD pattern on FDG-PET predicts future conversion to neurodegenerative disease in patients with iRBD.