Identification of Parkinson PACE subtypes and repurposing treatments through integrative analyses of multimodal clinical progression, neuroimaging, genetic, and transcriptomic data

Abstract

The Parkinson's disease (PD) is a heterogeneous neurodegenerative disease, of which the etiological and pathological mechanisms remain unclear to date. PD has been associated with diverse movement dysfunctions and non-motor symptoms (i.e., symptom heterogeneity) and progression patterns of these symptoms differ from patient to patient (i.e., progression heterogeneity). To address these, the present investigation aims at comprehensively considering full progression course of early PDs to identify subtypes, each of which can reflect unique PD progression pattern. We retrospectively analyzed the Parkinson's Progression Markers Initiative (PPMI) and the Parkinson Disease Biomarkers Program (PDBP) as the development and validation cohorts, respectively. An unsupervised deep learning model was built to model progression trajectories in diverse clinical manifestations and cerebrospinal fluid (CSF) biomarkers to produce a representation vector for each patient, encoding his/her symptom progression profile. Then by performing clustering analysis on the patients' representation vectors, we identified three subtypes with distinct PD progression patterns in the PPMI cohort: Subtype I, mild baseline severity and mild symptom progression; mild baseline severity and moderate progression; and Subtype III, rapid symptom progression. Replication in the PDBP validation cohort demonstrated reproducibility of the subtypes. After that, we explored demographic factors, CSF biomarkers, neuroimaging biomarkers in brain regional atrophy, and genetic factors of the subtypes. Last, to enhance usability of the subtypes, predictive model of subtypes that relies on data at baseline and 1-year follow-up was trained. In conclusion, the identified subtypes revealed significant symptom progression patterns of PDs. Patients with similar baseline severities can even suffer from different progression pattern, leading to distinct prognosis. Demographic factors, biomarkers, and genetic components of the subtypes suggested distinct biological mechanisms and pathways potentially leading to those progression patterns. Our findings may benefit pathophysiological study, clinical practice, and clinical trials to advance PD.

Transcriptomics profiling of Parkinson’s disease progression subtypes reveals distinctive patterns of gene expression

“…In the replication study of the PD progression subtype identification, it has been found that the most severe subtype (S3) had distinctive clinical features when compared to the two less severe subtypes (S1 and S2). Moreover, it was observed that there was limited signal in baseline variables to discriminate between the less severe subtypes [22]. These observations are in line with our results, as the performance of our classifier is poor in discriminating between S1 and S2 (0.576 AUROC).…”

“…The identification of progression subtypes is of extreme importance in order to attempt settling the heterogeneity of PD. Recent research has shown that people with PD can exhibit a variety of progression patterns from diagnosis onwards [5, 8, 19, 20, 22, 24]. The identification of diseasemodifying treatments can be fostered by finer-grained diagnoses and biomarkers identification, pursuing a precision medicine approach.…”

“…The analysis of comprehensive clinical and biological data resulted in the identification of three distinct subtypes: in brief, subtype I (S1) starts with mild motor and non-motor symptoms, and motor impairment increases with a moderate rate over time; subtype II (S2) has moderate motor and non-motor symptoms at baseline, with a slow progression rate; subtype III (S3) has significant motor and non-motor symptoms at baseline, and its impairment progresses rapidly over time, thus accounting for a worse prognosis [20]. An improved iteration of this approach, using an LSTM coupled with a Deep Progression Embedding (DPE) model, was shared as a preprint in 2021, aligning with earlier findings but awaiting peer-review [22]. Other authors developed their own algorithm for the identification of progression subtypes [23], but the heterogeneity in the results dependent on the features selected for analysis, and unavailability of clustered subject IDs, made us prone to focusing on PD progression subtypes identified in [20].…”

“…Other authors developed their own algorithm for the identification of progression subtypes [23], but the heterogeneity in the results dependent on the features selected for analysis, and unavailability of clustered subject IDs, made us prone to focusing on PD progression subtypes identified in [20]. Not only the latter is ongoing research, still needing peer-revision for its latest update [22], but has open access to the full analysis code and tables through GitHub.…”

Phenomics and Robust Multiomics Data for Cardiovascular Disease Subtyping

“…As in clinical settings, the explicability of model prediction is as important as its accuracy; it will be critical to push for the development of interpretable machine learning models and diagnostic criteria for characterizing their output. 139 In this vein, popular measures such as the Shapley Additive Explanations have been adapted to the subtyping context to characterize better the generated patient clusters 140 or even to discover new clusters. 141 Alongside technical advancements, implementing these models into clinical practice poses several challenges.…”