Skip to main content
Insight News
July 2019

Scientists from IBM Research and New York University (NYU) have developed a novel, fast and non-invasive method of performing glaucoma visual field tests using artificial intelligence (AI) trained on retinal imaging data.

Eye looking at a hologram

The researchers used a technology developed by IBM to take three-dimensional optical coherence tomography (OCT) images of the retina and train a machine-learning algorithm. The team claims the algorithm they have developed accurately estimates visual field index (VFI) values with an average error rate of only 2% .

The system, according to the researchers, can address the challenges of conventional glaucoma testing. The time consuming test is frustrating for patients, and previous studies have shown that factors such as a patient’s alertness and the time of day can affect results.

“From a biological point of view, we know there are associations between visual function and retinal structure and we can estimate visual function directly from structures in the eye,” Dr Rahil Garnavi, senior research scientist and IBM Research Australia manager, said.

“VFI is a global metric that represents the entire visual field, and accurately capturing that with AI offers to lay the groundwork for future technologies that can potentially use this analysis to quickly estimate a patient’s visual function. Our study suggests the structural information captured by OCT contains information that is highly correlated with functional measurement and could be extremely useful to professionals as they look to make a diagnosis.”

By capturing VFI accurately, AI could help lay the groundwork for technologies that could rapidly estimate and analyse a patient’s visual function. The results of the study have shown it as an improved approach compared to traditional tests.

The tool, when paired with an IBM machine learning system capable of forecasting visual function test results, could provide doctors access to detailed patient information without multiple tests.

“The ability to do this could one day help professionals to better predict the progression and onset of the disease, and adjust treatments accordingly,” Garnavi added.