Deep learning, large datasets, genome-wide analysis to determine novel pathways in the progression of age-related macular degeneration

Lead Research Organisation: University College London
Department Name: Institute of Health Informatics


Background- We have no effective treatment for most forms of age-related macular degeneration, the commonest cause of legal blindness in the western world, with numerous failed clinical trials. Age related macular degeneration, in its earliest form, is characterised by subtle build-up of deposits under the retina (the light sensitive structure coating the inside of the eyeball). The deposits under the retina can be of different size or location. Some patterns of deposits have been identified as important risk factors for progression to, late-stage AMD. These deposits can be visualised on optical coherence tomographic imaging (OCT) and to lesser extent on colour digital photography of the retina. Eyes with deposits under the retina are associated with faster progression rate to vision loss. It is well established that advanced AMD has a genetic component. AMD genome wide association studies (GWAS) have been highly successful in identifying strong and highly replicated associations of a number of genetic markers (single nucleotide polymorphisms). Large genetic studies have to date not graded for the subtle subretinal deposits form of AMD, a form strongly associated with devastating visual loss. Therefore, neither the genetic associations nor risk factors for development are understood. Being able to actively identify and quantify these deposits in an automated way will aid both the busy clinician managing such patients as well as being able to identify the phenotype for genetic and risk factor studies. Understanding these risk factors represent a tractable problem with the potential for specific interventions to limit their devastating consequences.

Objectives-1- Is it possible to a develop deep learning algorithm to identify the different patterns of deposits under the retina, on OCT imaging, that will quantify them and provide a readout to the treating clinician? 2- Can we identify different patterns of AMD deposits to already collected large (n>50 000 patients) that have genotypic data 3 - Can we Identify novel genetic risk factors that guide treatment approaches to dry AMD with a deep learning framework to test the predictive capacity of statistically significant single nucleotide polymorphism (SNPs)?


10 25 50

Studentship Projects

Project Reference Relationship Related To Start End Student Name
EP/S021612/1 01/04/2019 30/09/2027
2265357 Studentship EP/S021612/1 01/10/2019 30/09/2023 Roy Schwartz