A perfect match

1. Predicting future visual field results

Scientists have used AI to accurately predict future visual field results. The research team, which included Moorfields Eye Hospital ophthalmologist Pearse Keane, used more than 32,000 Humphrey Field Analyzer (HFA) results taken over two decades to train a deep learning algorithm in predicting future changes. Using only a single HFA plot as input, the deep learning model is able to forecast, with reasonable accuracy, visual field results up to five years in the future.

2. Interpreting OCT scans

A pitfall of traditional visual field tests is that they rely on patient feedback and can be influenced by how alert a patient is. But what if a subjective test did not have to be used? IBM Research has collaborated with New York University to develop deep learning techniques capable of estimating the visual field index of a patient from a single optical coherence tomography (OCT) image of the optic nerve. This is significant because it lays the groundwork for technology that could one day estimate a patient’s visual function without the need for multiple time-consuming tests.

3. Personalised monitoring of glaucoma

Dr Siamak Yousefi, from the University of Tennessee Health Science Center, has developed a ‘glaucoma radar’ that could improve the management and tracking of glaucoma disease progression. The tool, which incorporates AI technology, can be used for personalised monitoring of functional visual loss. The radar displays multiple layers of information, including global severity of visual function and patterns of visual functional loss.