Making dream machines a reality

Heidelberg Engineering head of clinical affairs, Tim Cole, began his presentation at 100% Optical with a brief foray into eye scanning technology – not within hospitals or optometry practices – but the classic science fiction film Blade Runner.

Cole explained that the title of his talk Do androids dream of electric eyes? gives a nod to the Philip K Dick novel that formed the inspiration for the film, Do androids dream of electric sheep?

He observed that the technology used to scan the eyes of potential ‘replicants’ in the 1982 and 2017 films resembles an optical coherence tomography (OCT) device.

However, while the device is a desktop machine in the original Blade Runner, it moves past current widely available OCT technology to a handheld device in the 2017 film.

Cole shared that home diagnostics company Visotec, which is owned by Heidelberg Engineering, is currently developing a prototype handheld OCT device. A patient looks through the smaller, more portable device like a pair of binoculars.

Cole added that there is a lot of demand for a handheld OCT device for domiciliary and paediatric use.

He observed that artificial intelligence (AI) technology is now prevalent within daily life.

“We are really immersed in this world. I think what is interesting is that we are still scared of it. I speak to ophthalmologists who ask, ‘is this going to replace us?’,” Cole shared.

Cole discussed the field of oculomics – where insight on systemic health conditions is gained from the eye.

He noted that gaining understanding about the rest of the body from the eye is not new. For example, Leonardo da Vinci dissected many eyes to learn about human health.

Cole described examples of technology within ophthalmology that use AI. RetinAI has apps for detecting diabetic retinopathy, geographic atrophy, wet age-related macular degeneration (AMD) and dry AMD.

RetInSight uses AI to estimate fluid present in the retina to the nanolitre, while Altris AI enables clinicians to drag and drop OCT scans into the software – which will then suggest a diagnosis and whether to refer.

Cole believes that AI will be particularly important in glaucoma care in the future.

“Glaucoma is a silent thief of sight. There are these subtle changes before sight is lost,” he said.

Cole emphasised the importance of training AI with good data – for example, images that have been mislabelled or with poor image quality will compromise the technology.

“If you feed an AI garbage, it’s going to give you garbage,” he said.

Inconsistencies within referrals from optometry practices to secondary care also pose challenges in establishing the large data sets that are required for AI.

Cole advised optometrists to keep calm and keep learning when it comes to this emerging technology.

“I think we are right at the beginning of this tipping point in human evolution,” he said.