Integrating retinal and brain imaging biomarkers for diagnosis of parkinson’s disease: findings from a time-lagged cross-sectional persian cohort study
Parkinson’s disease (PD) is a multisystem neurodegenerative disorder with both central and peripheral manifestations. This study aimed to comprehensively characterize structural and functional alterations in the retina and brain, as well as systemic blood biomarkers, in patients with clinically diagnosed PD compared with healthy controls, using a multimodal imaging approach. Blood-derived biomarkers and OCT data were collected from 29 PD patients and 25 healthy participants over an average interval of 2.7 years. MRI scans were performed using a Philips 1.5 Tesla scanner. Retinal and brain structural and perfusion metrics were analyzed using linear regression models adjusted for age, with statistical significance determined through false discovery rate (FDR) correction. The diagnostic performance was assessed using logistic regression, LASSO-penalized logistic regression model and ROC curve analysis. The study indicated that reduced ALT serum is the only significant hematological marker (p = 0.0253). Structural MRI revealed significant volume reductions in the left amygdala and caudate (p = 0.0241 and 0.0035 respectively); However, these findings necessitate careful interpretation due to limitations in spatial resolution. Perfusion MRI showed lower cerebral blood flow in gray matter and the whole brain (p = 0.02) in PD patients. The best imaging biomarker (PCASL total CBF) achieved an AUC of 0.79. A LASSO-penalized logistic regression model integrating left amygdala volume, left caudate volume, and SGPT achieved a significantly elevated AUC of 0.954, with 100% sensitivity and 86.7% specificity. The integration of central and peripheral biomarkers encompasses complementary aspects of Parkinson’s disease pathology, providing a more comprehensive diagnostic framework than any singular modality.