Two Distinct Trajectories of Clinical and Neurodegeneration Events in Parkinson’s Disease
Cheng Zhou,Linbo Wang,Wei Cheng,JinChao Lv,Xiaojun Guan,Tao Guo,Jingjing Wu,Wei Zhang,Ting Gao,Xiaocao Liu,Xueqin Bai,Haoting Wu,Zhengye Cao,Luyan Gu,Jingwen Chen,Jiaqi Wen,Peiyu Huang,Xiaojun Xu,Baorong Zhang,Jianfeng Feng,Minming Zhang
DOI: https://doi.org/10.1038/s41531-024-00635-z
2023-01-01
npj Parkinson s Disease
Abstract:Increasing evidence suggests that Parkinson’s disease (PD) exhibits disparate spatial and temporal patterns of progression. Here we used a machine-learning technique—Subtype and Stage Inference (SuStaIn) — to uncover PD subtypes with distinct trajectories of clinical and neurodegeneration events. We enrolled 228 PD patients and 119 healthy controls with comprehensive assessments of olfactory, autonomic, cognitive, sleep, and emotional function. The integrity of substantia nigra (SN), locus coeruleus (LC), amygdala, hippocampus, entorhinal cortex, and basal forebrain were assessed using diffusion and neuromelanin-sensitive MRI. SuStaIn model with above clinical and neuroimaging variables as input was conducted to identify PD subtypes. An independent dataset consisting of 153 PD patients and 67 healthy controls was utilized to validate our findings. We identified two distinct PD subtypes: subtype 1 with rapid eye movement sleep behavior disorder (RBD), autonomic dysfunction, and degeneration of the SN and LC as early manifestations, and cognitive impairment and limbic degeneration as advanced manifestations, while subtype 2 with hyposmia, cognitive impairment, and limbic degeneration as early manifestations, followed later by RBD and degeneration of the LC in advanced disease. Similar subtypes were shown in the validation dataset. Moreover, we found that subtype 1 had weaker levodopa response, more GBA mutations, and poorer prognosis than subtype 2. These findings provide new insights into the underlying disease biology and might be useful for personalized treatment for patients based on their subtype.