George and Matilda Eyecare and IBM Research Australia will work together to determine how retinal fundus photos, optical coherence tomography (OCT) images, and AI algorithms can be used to inform and potentially help guide practitioners in detecting glaucoma.
According to the World Health Organisation, glaucoma is the leading cause of irreversible blindness worldwide. The profound visual loss and blindness caused by glaucoma is usually preventable with early detection, diagnosis and management. Despite available health care, 50 per cent of all patients with glaucoma are still undiagnosed, and many progress to profound blindness.1
Retinal imaging, such as retinal fundus photography and OCT can help diagnose early signs of many sight threatening eye diseases and as such, are fast becoming essential tools to help clinicians diagnose and manage eye disease in eye care practices.
Using retinal photos and OCT images, IBM Research Australia scientists and George and Matilda Eyecare optometrists will train and validate AI algorithms – such as deep learning neural networks – and use the resulting image analytics models to look at the relationship between several characteristics of the optic nerve in determining instances of glaucoma. Teams will aim to determine the correlation between glaucoma severity and higher eye pressure readings (intraocular pressure), higher cup to disc ratios, and the thinning and thickness of certain retinal layers (the retinal nerve fibre layer and ganglion cell layer).
Chris Beer, CEO for George and Matilda Eyecare, said, “Our vision at George and Matilda Eyecare is to help the world see better, one person at a time, by supporting and uniting local independents to build the best optometry community. This project is very exciting for us in looking at new ways of embracing technology to improve what we can offer our patients. We think this research will allow us the ability to improve best practice for the patients that entrust their eyesight to our care. We are also excited about what this collaboration with IBM Research Australia will bring in new learnings for the patients that see us in the future, and how it will help all Australians, and even those across the world, to reduce preventable blindness in the future.”
In previously published work, scientists from IBM Research Australia were able to teach AI algorithms to detect the level of diabetic retinopathy2 to a very high degree of accuracy from looking at retinal photos alone.3
References
- www.iapb.org/knowledge/what-is-avoidable-blindness/glaucoma/
- www.cera.org.au/wp-content/uploads/2015/11/OutOfSightReport.pdf
- www.ibm.com/blogs/research/2017/04/spotting-diabetic-retinopathy/
