Wagner, Siegfried Karl; (2023) Retinal bioimaging for neurodegenerative and cardiovascular diseases. Doctoral thesis (Ph.D), UCL (University College London).

Preview

Text SKWagner_Thesis_v1.1.pdf - Accepted Version Download (8MB) | Preview

Abstract

Imaging of the retina using new high-resolution techniques enables in-vivo and minimally invasive assessment of the body’s cardiovascular, neurological, and metabolic systems. The ability to interrogate these imaging-derived data using advanced statistical and computational approaches has significantly improved in the past 15 years. While retinal signatures of systemic disease (‘oculomics’) from in-vivo imaging are increasingly postulated, our understanding of the cardiometabolic and neurodegenerative associations with ocular phenotype has been hampered by a lack of large labeled retinal imaging datasets. In this thesis, I report methods and findings from my analysis of clinical and retinal imaging data from two health datasets – UK Biobank, a prospective cohort study of >500,000 volunteers residing in the UK, and AlzEye, where ophthalmic data from 353,157 patients attending Moorfields Eye Hospital has been deterministically linked with hospital admissions data across England. I report that acute conditions affecting the retinal vessels (retinal artery occlusion), optic nerve (non-arteritic ischaemic optic neuropathy) and cranial nerves (third, fourth and sixth palsies) represent sentinel events for cardiovascular dysfunction and all-cause mortality bearing implications for the interdisciplinary management of these conditions. I show that differences in retinal morphology, resembling cardiometabolic, neurodegenerative and inflammatory dysfunction are seen in individuals with schizophrenia and periodontitis. The observation that individuals with Parkinson’s disease exhibit thinner inner nuclear layers, the retinal niche of dopaminergic activity, and that these differences are detectable, on average, seven years prior to clinical diagnosis warrants further attention. Retinal imaging can also reveal insights into the mechanisms of systemic disease - bilateral retinal structural differences in individuals with unilateral amblyopia in childhood has informed a novel finding of heightened cardiometabolic dysfunction in these individuals. Finally, I demonstrate that while group-level differences exist, individual-level prediction of all-cause dementia using retinal imaging remains challenging and does not confer improved performance beyond sociodemographic and clinical risk factors. Future work should consider high-dimensional modelling approaches, alternative imaging modalities, and more granular case definitions of dementia.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONCSVXML

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item