4/6/2023
Digitalisation
News
Artificial intelligence in glaucoma diagnostics
Artificial intelligence can be used well where clear questions meet large amounts of data that contain the necessary information. In glaucoma diagnostics, too, there are already some examples of how AI can be used in research and applied in patient care. Hagen Thieme, Director of the University Eye Hospital Magdeburg, talks about these possibilities.
Glaucoma is a complex eye disease. Gradually, fibres of the optic nerve are lost. The consequences are visual field deficits that progress if the glaucoma is not treated. An important risk factor is the intraocular pressure. If it is too high, the nerve at the optic nerve head, i.e. at the point where it leaves the eye, comes under pressure. The blood supply to the nerve suffers and the nerve fibres die.
AI can already evaluate visual fields today - it is even supposed to recognise defects more reliably than human experts. This means that corresponding applications could support glaucoma diagnosis. It should even be possible to predict how the visual field will develop: A research group had a recurrent neural network evaluate five visual field findings of patients at a time and then predict how a sixth examination would turn out. The result was superior to conventional methods.
Thieme sees another possible application in the assessment of the thickness of the retinal nerve fibre layer based on photographs of the ocular fundus. This application was trained using fundus photographs and RNFL measurements with OCT. Based on the photographs, the programme can distinguish whether a rapid or moderate loss of nerve fibres is to be expected. It thus helps ophthalmologists in the long-term follow-up of glaucoma disease and provides support for therapy decisions even where an OCT examination is not possible.
Possibilities of AI for glaucoma diagnostics
Close-meshed intraocular pressure measurements, visual field analyses and OCT (optical coherence tomography) findings enable detailed diagnostics, with which the course of this chronic disease and the success of treatment must be precisely documented over many years. Enormous amounts of data are generated in the process. Evaluating them is becoming increasingly difficult and time-consuming. Already here, an automated evaluation of the data with AI should be able to provide decisive help.AI can already evaluate visual fields today - it is even supposed to recognise defects more reliably than human experts. This means that corresponding applications could support glaucoma diagnosis. It should even be possible to predict how the visual field will develop: A research group had a recurrent neural network evaluate five visual field findings of patients at a time and then predict how a sixth examination would turn out. The result was superior to conventional methods.
Thieme sees another possible application in the assessment of the thickness of the retinal nerve fibre layer based on photographs of the ocular fundus. This application was trained using fundus photographs and RNFL measurements with OCT. Based on the photographs, the programme can distinguish whether a rapid or moderate loss of nerve fibres is to be expected. It thus helps ophthalmologists in the long-term follow-up of glaucoma disease and provides support for therapy decisions even where an OCT examination is not possible.