Explainable multimodal imaging–based machine learning for cardiovascular risk stratification in early Parkinson’s disease

Year 2026, Volume: 7 Issue: 2, 313 - 322, 27.03.2026

Esra Demir Ünal , Ahmet Kadir Arslan

https://izlik.org/JA93BG98HZ

Abstract

Aims: This study aimed to characterize convergent biochemical and cardiovascular imaging endophenotypes of atherosclerotic risk in early PD and to implement an interpretable machine-learning architecture integrating multimodal clinical and imaging parameters for precision cardiovascular risk profiling and cardiometabolic classification.
Methods: A total of 125 early-stage idiopathic PD patients and age- and sex-matched controls were analyzed. Feature space was reduced using cross-validated Least Absolute Shrinkage and Selection Operator (LASSO) with penalty tuning and stability selection to mitigate multicollinearity. XGBoost, Random Forest, RBF-kernel Support Vector Machine, and Stochastic Gradient Boosting models were trained under a train–hold-out framework with Bayesian hyperparameter optimization. Model performance was assessed via bootstrapping, and interpretability was provided using SHapley Additive exPlanations (SHAP).
Results: PD patients showed increased hypertension (54.4%; OR 3.2, p=0.007) and hypercholesterolemia (OR 2.8, p=0.01), with excess left ventricular systolic dysfunction (28.8% vs 0%, p<0.001), aortic insufficiency (20.0% vs 0%, p=0.001), and high-risk carotid plaques (unstable 40.8% vs 2.0%; calcified 49.6% vs 18.0%; both p<0.001). Inflammation was elevated (CRP 7.70±7.22 vs 3.34±1.31 mg/L, p<0.001) with reduced lymphocyte-to-monocyte ratio (LMR) (3.99±1.94 vs 5.13±1.45, p<0.001). XGBoost achieved superior discrimination (accuracy 0.941, 95% CI 0.911–0.971; sensitivity 0.923; specificity 0.889; PPV 0.960, 95% CI 0.930–0.990; AUC 0.95, 95% CI 0.91–0.983; Brier 0.07), with explainability dominated by LMR (SHAP 100%), followed by CRP (92.6%) and glucose-metabolic markers (65.6%).
Conclusion: PD demonstrates a convergent inflammatory–metabolic endophenotype with elevated carotid and cardiac risk burden. An explainable XGBoost-based multimodal framework implicated LMR and C-RP as discriminative biomarkers for targeted surveillance in early PD.

Keywords

Accelerated atherosclerosis , carotid plaque vulnerability , lymphocyte-to-monocyte ratio , machine learning risk stratification , Parkinson's disease , XGBoost algorithm

Ethical Statement

The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Ethics Committee of Ankara Etlik City Hospital (AEŞH-BADEK2024-331). Informed Consent Statement: Written informed consent was obtained from all subjects involved in the study.

Supporting Institution

Not applicable

Thanks

Not applicable

References

• Dong-Chen X, Yong C, Yang X. Signaling pathways in Parkinson’s disease: molecular mechanisms and therapeutic interventions. Signal Transduct Target Ther. 2023;8(1):73. doi:10.1038/s41392-023-01353-3
• Tansey MG, Wallings RL, Houser MC. Inflammation in Parkinson's disease: a parkway to pathogenesis, pathology, and persistence. Neuron. 2022;110:3259-3278. doi:10.1016/j.neuron.2022.07.020
• Becker C, Jick SS, Meier CR. Risk of stroke in patients with idiopathic Parkinson disease. Parkinsonism Relat Disord. 2010;16(1):31-35. doi:10. 1016/j.parkreldis.2009.06.005
• Suri JS, Paul S, Maindarkar MA, et al. Cardiovascular/stroke risk stratification in Parkinson’s disease patients using atherosclerosis pathway and artificial intelligence paradigm: a systematic review. Metabolites. 2022;12(4):312. doi:10.3390/metabo12040312
• Liu Y, Xue L, Zhang Y, Xie A. Association between stroke and Parkinson's disease: a meta-analysis. J Mol Neurosci. 2020;70:1169-1176. doi:10.1007/s12031-020-01524-9
• Li Q, Wang C, Tang H, Chen S, Ma J. Stroke and coronary artery disease are associated with Parkinson’s disease. Can J Neurol Sci. 2018;45(5):559-565. doi:10.1017/cjn.2018.56
• Liu W, Laranjo L, Klimis H, et al. Machine-learning versus traditional approaches for atherosclerotic cardiovascular risk prognostication in primary prevention cohorts: a systematic review and meta-analysis. Eur Heart J Qual Care Clin Outcomes. 2023;9(4):310-322. doi:10.1093/ehjqcco/qcad017
• Stary HC, Chandler AB, Dinsmore RE, et al. A definition of advanced types of atherosclerotic lesions and a histological classification of atherosclerosis. Circulation. 1995;92(5):1355-1374. doi:10.1161/01.cir.92. 5.1355
• Writing Committee Members; ACC/AHA Joint Committee Members. 2022 AHA/ACC/HFSA guideline for the management of heart failure. J Card Fail. 2022;28(5):e1-e167. doi:10.1016/j.cardfail.2022.02.010
• Isselbacher EM, Preventza O, Black JH 3rd. 2022 ACC/AHA guideline for the diagnosis and management of aortic disease: a report of the American Heart Association/American College of Cardiology joint committee on clinical practice guidelines. Circulation. 2022;146(24):e334-e482. doi: 10.1161/CIR.0000000000001106
• Klein AL, Wang TKM, Cremer PC, et al. Pericardial diseases: international position statement on new concepts and advances in multimodality cardiac imaging. JACC Cardiovasc Imaging. 2024;17(8): 937-988. doi:10.1016/j.jcmg.2024.04.010
• Hu Y, Xu S. Association between Parkinson's disease and the risk of adverse cardiovascular events: a systematic review and meta-analysis. Front Cardiovasc Med. 2023;10:1284826. doi:10.3389/fcvm.2023.1284826
• Potashkin J, Huang X, Becker C, et al. Understanding the links between cardiovascular disease and Parkinson's disease. Mov Disord. 2020;35(1): 55-74. doi:10.1002/mds.27836
• Madetko N, Migda B, Alster P, et al. Platelet-to-lymphocyte ratio and neutrophil-to- lymphocyte ratio may reflect differences in PD and MSA-P neuroinflammation patterns. Neurol Neurochir Pol. 2022;56(2): 148-155. doi:10.5603/PJNNS.a2022.0014
• Li F, Weng G, Zhou H, et al. The neutrophil-to-lymphocyte ratio, lymphocyte-to-monocyte ratio, and neutrophil-to-high-density-lipoprotein ratio are correlated with the severity of Parkinson's disease. Front Neurol. 2024;15:1322228. doi:10.3389/fneur.2024.1322228
• Liu Z, Fan Q, Wu S, Wan Y, Lei Y. Compared with the monocyte to high-density lipoprotein ratio (MHR) and the neutrophil to lymphocyte ratio (NLR), the neutrophil to high-density lipoprotein ratio (NHR) is more valuable for assessing the inflammatory process in Parkinson's disease. Lipids Health Dis. 2021;20(1):35. doi:10.1186/s12944-021-01462-4
• Hosseini S, Shafiabadi N, Khanzadeh M, et al. Neutrophil to lymphocyte ratio in Parkinson's disease: a systematic review and meta-analysis. BMC Neurol. 2023;23(1):333. doi:10.1186/s12883-023-03380-7
• Mannarino MR, Bianconi V, Gigante B, et al. Neutrophil to lymphocyte ratio is not related to carotid atherosclerosis progression and cardiovascular events in the primary prevention of cardiovascular disease: results from the IMPROVE study. Biofactors. 2022;48(1):100-110. doi:10.1002/biof.1801
• Yu J, Zhao Z, Li Y, et al. Role of NLRP3 in Parkinson's disease: specific activation especially in dopaminergic neurons. Heliyon. 2024;10(7): e28838. doi:10.1016/j.heliyon.2024.e28838
• Li C, Ke B, Chen J, et al. Systemic inflammation and risk of Parkinson's disease: a prospective cohort study and genetic analysis. Brain Behav Immun. 2024;117:447-455. doi:10.1016/j.bbi.2024.02.013
• Yu H, Sun T, He X, et al. Association between Parkinson's disease and diabetes mellitus: from epidemiology, pathophysiology and prevention to treatment. Aging Dis. 2022;13(6):1591-1605. doi:10.14336/AD.2022.0325
• Sian-Hulsmann J, Riederer P, Michel TM. Metabolic dysfunction in Parkinson’s disease: unraveling the glucose–lipid connection. Biomedicines. 2024;12(12):2841. doi:10.3390/biomedicines12122841
• Grosu L, Grosu AI, Crisan D, Zlibut A, Perju-Dumbrava L. Parkinson's disease and cardiovascular involvement: edifying insights. Biomed Rep. 2023;18(3):25. doi:10.3892/br.2023.1607
• Yan LY, He QF, Lu MY, et al. Association between carotid plaque and Parkinson’s disease. Ann Transl Med. 2019;7(5):94. doi:10.21037/atm. 2019.01.68
• Kasartzian DI, Tsiampalis T. Transforming cardiovascular risk prediction: a review of machine learning and artificial intelligence innovations. Life (Basel). 2025;15(1):94. doi:10.3390/life15010094
• Cao K, Liu C, Yang S. Prediction of cardiovascular disease based on multiple feature selection and improved PSO-XGBoost model. Sci Rep. 2025;15(1):12406. doi:10.1038/s41598-025-96520-7
• Uçar CA, Çokal B, Artık HAÜ, İnanç LE, Yoldaş TK. Comparison of neutrophil-lymphocyte ratio (NLR) in Parkinson's disease subtypes. Neurol Sci. 2017;38(2):287-293. doi:10.1007/s10072-016-2758-8
• Demir Ünal E. The role of metabo-inflammatory syndrome and ferropathological processes in atherosclerotic plaque formation in Parkinson's disease: insights from high-resolution carotid duplex analysis. Turkish J Cardiol. 2025;53(4):228-237. doi:10.5543/tkda.2025. 40390