Lin, Wei

The accurate prediction of motor progression in Parkinson’s disease (PD) remains a major clinical challenge that limits personalized treatment planning and efficient clinical trial design. In this study, we developed and validated a machine-learning framework integrating a targeted panel of plasma proteins measured by Olink proximity extension assays with clinical variables to stratify patients according to their progression risk. We analyzed baseline plasma samples from 211 early-stage PD patients enrolled in the Parkinson’s Progression Markers Initiative (PPMI) cohort using four targeted Olink panels, from which 28 circulating proteins were retained after quality-control filtering. Patients were classified as rapid or slow progressors based on their annualized change in MDS-UPDRS Part III scores. Among the algorithms tested, Random Forest achieved the highest discriminative performance with an area under the receiver operating characteristic curve (AUC) of 0.751 (95% CI: 0.684–0.811), which exceeded that of clinical predictors alone (AUC 0.666). The integration of targeted proteomic and clinical features further improved model performance (AUC 0.773; p = 0.009). Nested cross-validation confirmed minimal optimistic bias (AUC 0.743). To enhance clinical interpretability, we applied SHapley Additive exPlanations (SHAP) analysis, which identified interleukin-6 (IL-6), brain-derived neurotrophic factor (BDNF), and vascular endothelial growth factor A (VEGF-A) as the most influential predictors. SHAP feature rankings were highly stable across cross-validation folds (mean Spearman ρ = 0.91). The robustness of these findings was confirmed through sensitivity analyses using extreme quartile comparisons (AUC 0.823), treatment-naïve subgroup analysis (AUC 0.738), and a clinically anchored outcome definition based on the minimal clinically important difference (AUC 0.739). A decision curve analysis demonstrated a net clinical benefit across threshold probabilities of 0.25–0.70. Our results establish targeted plasma protein profiling combined with interpretable machine learning as a promising tool for PD motor progression risk stratification, with potential applications in individualized patient counseling regarding motor prognosis and the selection of candidates for disease-modifying trials.

Abstract Background Parkinson's disease (PD) represents a growing global health challenge, with prevalence doubling from 2.5 million cases in 1990 to 6.1 million in 2016, and projections suggesting it will exceed 12 million by 2040. This exponential growth has been termed the "Parkinson's pandemic." China and the United States, which together account for nearly half of the global PD population, have developed markedly different approaches to end-of-life (EOL) care, shaped by their distinct healthcare systems and cultural contexts. Understanding these differences is crucial for developing effective global strategies. This systematic review and meta-analysis comprehensively compares advanced-stage PD care between these two nations to quantify disparities and inform evidence-based policy interventions. Methods We conducted a rigorous systematic review following PRISMA 2020 guidelines, searching six databases (PubMed, Embase, Scopus, Web of Science, CNKI, Wanfang) from January 2015 to October 2025. The timeframe was selected to capture contemporary care patterns following major healthcare reforms in both countries (Affordable Care Act implementation in the US; Healthy China 2030 initiative). Inclusion criteria encompassed peer-reviewed studies reporting on advanced PD (Hoehn-Yahr stage ≥ 3) with outcome measures including survival, quality of life (QoL), palliative care access, caregiver burden, rehabilitation utilization, or healthcare costs. The sample size threshold (n ≥ 100) balanced statistical power with inclusivity. Two independent reviewers screened 3,847 records with excellent agreement (Cohen's κ = 0.82–0.89). Quality assessment employed the Newcastle-Ottawa Scale (NOS ≥ 8 indicating high quality) and ROBINS-I for observational studies, with GRADE criteria determining evidence certainty. Meta-analyses utilized random-effects models with comprehensive heterogeneity assessment (I² statistics, meta-regression) and publication bias evaluation (Egger's test, trim-and-fill analysis). Sensitivity analyses included high-quality studies only and examined the influence of large database studies. Results Our analysis of 42 high-quality studies encompassing over 125,000 patients revealed profound and systematic cross-national disparities across multiple domains. In terms of care access, US patients demonstrated significantly higher hospice enrollment rates (28–35% vs. <1% in China, p < 0.001; HIGH certainty evidence) and greater access to specialized neuropalliative programs (15% of US centers vs. <1% in China). These differences translated into measurable quality of life disparities, with Chinese patients scoring 17.6 points worse on the PDQ-39 scale (95% CI 14.2–21.0; MODERATE certainty, I²=68%), a difference exceeding the minimal clinically important difference nearly four-fold. The burden on families showed equally striking patterns: Chinese families provided 94% of EOL care compared to 61% in the US, resulting in significantly higher caregiver depression rates (47% vs. 32%, OR = 1.89, 95% CI 1.45–2.47) and burden scores (Zarit difference: 10.2 points, 95% CI 7.8–12.6). Access to interventions revealed additional gaps, with only 9.2% of Chinese patients receiving physical therapy versus 14.4% in the US (OR = 0.61, 95% CI 0.49–0.76), and a 5.25-fold disparity in deep brain stimulation access (4.2% vs. 0.8%). Economic analysis demonstrated substantial out-of-pocket burden differences: Chinese families paid 58% of total costs ($4,200 annually) versus 12% for US families ($2,800). Meta-regression identified healthcare system characteristics as the primary driver of these disparities (R²=0.42, p < 0.001). Conclusions This comprehensive analysis reveals that systematic differences in healthcare infrastructure, insurance coverage, and cultural care models produce markedly divergent EOL experiences for PD patients in China versus the United States. These findings have immediate policy implications. For China, we recommend: (1) establishing pilot neuropalliative programs in tertiary centers with phased national expansion; (2) developing caregiver support infrastructure including respite care and training programs; (3) leveraging telemedicine platforms to address urban-rural disparities. For the United States, priorities include: (1) addressing racial disparities in rehabilitation access through culturally competent outreach; (2) expanding Medicare coverage for maintenance therapy; (3) earlier integration of neuropalliative care. Implementation science research is urgently needed to test culturally adapted interventions. As global PD prevalence accelerates, lessons from this bilateral comparison provide a framework for other nations confronting similar challenges in neurological care delivery.