Detecting Dementia in the Retina: a Big Data Machine Learning Approach
Detecting Dementia in the Retina: a Big Data Machine Learning Approach
This retrospective case control explores the retinal features of dementia associated with neurodegenerative diseases, particularly Alzheimer's disease. By linking a pseudonymised dataset of three-dimensional retinal scans, called optical coherence tomography, with nationally held data on dementia, corresponding characteristics will be evaluated through descriptive statistics and machine learning techniques.
By 2025, it is estimated that approximately 1 million people in the United Kingdom (UK) will suffer from dementia, a syndrome associated with progressive decline in brain function. While there is currently no cure for most types of dementia, early diagnosis can help patients receive the appropriate treatment and support to help maintain mental function.
The focus of this project is to identify changes in retinal structure associated with dementia. In collaboration with bioinformatics experts at University College London (UCL), the investigators propose to analyse our repository of >1 million retinal scans, termed optical coherence tomography (OCT), performed regularly on patients since 2008. OCT scans will be linked at a patient level to data from the Hospital Episode Statistics (HES) database to identify those who have been diagnosed with dementia or went on the develop dementia. Thus, a pseudonymised classified dataset of retinal scans will be generated for qualitative and quantitative analysis. The primary objective is to characterise changes in the layers of the retina associated with dementia. Machine learning techniques may also be employed to identify novel patterns of retinal change associated with dementia.
Inclusion Criteria:
All patients over the age of 40 who have undergone an optical coherence tomography scan at Moorfields Eye Hospital NHS Foundation Trust
Exclusion Criteria:
Poor quality image