Towards implementation of genomic prediction models for glaucoma screening and personalised care

People fear losing their vision as much as developing dementia or cancer. Glaucoma is the leading cause of incurable blindness globally, affecting over 80 million people. In the early stages of glaucoma, there are usually no symptoms and people are unaware they have the disease.

By the time vision loss is apparent, there has already been considerable irreversible damage. Therefore, early detection of glaucoma is critical for preventing blindness. In the UK, glaucoma is detected opportunistically at community optometrists when people attend for eye checks.

There is no formal screening programme as current screening tests are inadequate. Once diagnosed, treatment is usually for life and glaucoma is a major burden on healthcare services with over 1.4 million National Health Service (NHS) outpatient visits per year in the UK. We are currently unable to predict which glaucoma patients need most intensive therapy, meaning some patients still develop blindness despite treatment and other patients are likely overtreated.

We are also unable to predict which of the many treatment options will be optimal for individual patients and the current standard of care is trial and error. All these challenges, in combination with the rapidly increasing number of glaucoma sufferers due to the ageing population, makes glaucoma a major public health challenge. We urgently need to innovate how we detect and treat glaucoma to prevent blindness and make best use of our limited resources.

Glaucoma is a hereditary condition but is "complex", meaning that it is caused by large number of genetic factors acting together. Mr research programme builds on my continued discovery of hundreds of these genetic factors for glaucoma using a combination of methods including artificial intelligence and leading a collaboration of all studies globally.

Put together in a prediction model, these factors helped predict who in a population will develop glaucoma. Many more factors need to be discovered, but this opens up the possibility of targeted population screening to enable earlier diagnosis and prevention of irreversible sight loss. In my renewal period, I aim to:

1. Increase the discovery of the genetic causes of glaucoma in non-European populations, to ensure our prediction models can be deployed equitably.

2. To test the glaucoma-detection accuracy of our genetic prediction models and artificial intelligence algorithms applied to retinal photographs.

3. Lead the cross-disciplinary design of a prospective randomised trial for genomics and artificial intelligence enhanced targeted glaucoma screening.

4. Develop a readily interpretable visual tool to present our prediction models that is accessible via the commonly used ophthalmic electronic medical records.

If successful, my proposed research will have considerable potential impact. Being able to identify people at high risk of glaucoma on the basis of a genetic test alone will allow targeted screening of these individuals, enabling earlier diagnosis and prevention of irreparable vision loss. If genetic factors predict the course of the disease, it will be possible to help decide which patients need most intensive treatment to prevent blindness and which patients are at low risk to avoid overtreating them.

Similarly, it may become possible to select the most effective treatment for each individual patient using genetic information. Put together, genetic testing may allow us to personalise the care of glaucoma patients and direct our limited resources to the patients that need it most. Innovative approaches such as this are essential to deal with the rapidly increasing numbers of glaucoma patients.

Additionally, the work identifying the key biological processes that underlie glaucoma will help the development of new drug treatments.