Parkinson’s Disease (PD) is the fastest growing

neurodegenerative disease worldwide1

• Diagnosis relies primarily on clinical evaluation2

• Many diagnostic tools for PD are invasive and expensive

• Retinal imaging reveals structural and vascular changes

mirroring neurodegenerative processes in the brain3,4

• There is still no universally accepted biomarker

• Visible changes in vessel caliber are under-studied

• We employ a novel calculation of the retinal arteriovenous

ratio (AVR) known as retinal vascularity index (RVI)

• We calculated RVI on ultra-widefield (UWF) retinal images

and compared RVI between patients with PD and controls

Methods

Cross-sectional study design (NCT03233646)

• Individuals with PD were diagnosed by neurologists specializing in

movement disorders

• Control individuals had normal cognition and no PD symptoms

• Exclusion criteria: <50 years age, diabetes, glaucoma, vitreoretinal

pathology, any non-PD dementia, corrected Snellen visual acuity

worse than 20/40

• UWF images were acquired using scanning laser ophthalmoscopy

(California, Optos, Marlborough, MA) in undilated eyes

• UWF images were analyzed using a semi-automated software

embedded in the Comprehensive Ocular Imaging Network (COIN,

ocularimaging.net)

• A trained grader manually identified the size and location of the optic

disc and fovea to define the region of interest (ROI) of five disc

diameters (DD)—47° field—centered on the optic disc

• The largest arteriole and venule in each quadrant (superonasal [SN],

superotemporal [ST], inferotemporal [IT], inferonasal [IN]) were

segmented at consistent distances from the optic disc center to

measure vessel diameter (d)

• Cross-sectional area of each vessel was automatically calculated and

averaged (A) for arterioles and venules separately

RVI = the ratio between arteriolar A

and venular A, calculated overall within

our ROI and within each quadrant.

• Statistical analysis: main outcome analysis using generalized

estimating equations with exchangeable correlation structure and

discriminatory analysis using receiver operating curves (ROC)

Results

168 eyes of 108 PD participants and 115 eyes of 71 control

participants were included

• RVI was significantly higher in PD compared to controls

• This difference was significant overall and in each of four

quadrants within our ROI

• ROC analysis revealed moderate discriminative ability

Conclusions

We observed an increased RVI in PD compared to controls

• Increased RVI may reflect changes in retinal vascular

structure related to α-synuclein deposition5, dopamine

dysregulation6, and/or neuroinflammation7

• Discriminative analysis suggests that RVI may not stand

alone as a diagnostic tool

• Further study is warranted to determine if RVI derived from

a 47° ROI on UWF images may serve as a potential retinal

biomarker of PD