High-tech solutions top the list in the fight against eye disease

From Engadget - July 12, 2017

Cataracts are the single leading cause of blindness worldwide, afflicting roughly 42 percent of the global population, including more than 22 million Americans. The disease, which causes cloudy patches to form on the eye's normally clear lens, can require surgery if left untreated. That's why Google's DeepMind AI division has teamed with the UK's National Health Service (NHS) and Moorfields Eye Hospital to train a neural network that will help doctors diagnose early stage cataracts.

The neural network is being trained on a million anonymized optical coherence tomography (OCT) scans (think of a sonogram, but using light instead of sound waves) in the hopes it will eventually be able to supplement human doctors' analyses, increasing both the efficiency and accuracy of individual diagnoses.

"OCT has totally revolutionized the field of ophthalmology. It's an imaging system for translucent structures that utilizes coherent light," Dr. Julie Schallhorn, an assistant professor of ophthalmology at UC San Francisco, said. "It was first described in 1998 and it gives near-cell resolution of the cornea, retina and optic nerve.

"The optic nerve is only about 200 microns thick, but you can see every cell in it. It's given us a much-improved understanding of the pathogenesis of diseases and also their response to treatments." The new iteration of OCT also measures the phase-shift of refracted light, allowing doctors to resolve images down to the capillary level and observe the internal structures in unprecedented detail.

"We are great at correcting refractive errors in the eyes so we can give you good vision far away pretty reliably, or up close pretty reliably," Schallhorn continued. "But the act of shifting focus from distance to near requires different optical powers inside the eye. The way the eye handles this when you are young is through a process called 'accommodation.'" There's a muscle that contracts and changes the shape of the lens to help you focus on close objects. When you get older, even before you typically develop cataracts, the lens will stiffen and reduce the eye's ability to change its shape.

"The lenses that we have been putting in during cataract surgery are not able to mimic that [shapeshifting] ability, so people have to wind up wearing reading glasses," Schallhorn said. There's a lot of work in the field to find solutions for this issue and help restore the eye's accommodation.

There are two front-runners for that: Accommodating lenses, which use the same ciliary muscle to shift focus, and multifocal lenses, which work just like your parents' multifocal reading glasses except that they sit directly on the eye itself. The multifocals have been on the market for about a decade, though their design and construction has been refined over that time.

To ensure the lenses that doctors are implanting are just as accurate as the diseased ones they are removing, surgeons are beginning to use optiwave refractive analysis. Traditionally, doctors relied on measurements taken before the surgery to know how to shape the replacement lenses and combined those with nomograms to estimate how powerful the new lens should be.

The key word there is "estimate." "They especially have problems in patients who have already had refractive surgery like LASIK," Schallhorn explained. The ORA system, however, performs a wavefront measurement of the cornea after the cataract has been removed to help surgeons more accurately pick the right replacement lens for the job.

Corneal inlays are also being used. These devices resemble miniature contact lenses but sit in a pocket on the cornea that's been etched out with a LASIK laser to mimic the process of accommodation and provide a greater depth of focus. They essentially serve the same function as camera apertures. The Kamra lens from AcuFocus and the Raindrop Near Vision Inlay from Revision Optics are the only inlays approved by the FDA for use in the US.

Glaucoma afflicts more than 70 million people annually. This disease causes fluid pressure within the eye to gradually increase, eventually damaging the optic nerve that carries electrical signals from the eye to the brain. Normally, detecting the early stages of glaucoma requires a comprehensive eye exam by a trained medical professional -- folks who are often in short supply in rural and underserved communities. However, the Cambridge Consultants' Viewi headset allows anyone to diagnose the disease -- so long as they have a smartphone and 10 minutes to spare.

The Viewi works much like the Daydream View, wherein the phone provides the processing power for a VR headset shell -- except, of course, that instead of watching 360 degree YouTube videos, the screen displays the flashing light patterns used to test for glaucoma. The results are reportedly good enough to share with you eye doctor and take only about five minutes per eye. Best of all, the procedure costs only about $25, which makes it ideal for use in developing nations.

And while there is no known cure for glaucoma, a team of researchers from Stanford University may soon have one. Last July, the team managed to partially restore the vision of mice suffering from a glaucoma-like condition.

Normally, when light hits your eye, specialized cells in the retina convert that light into electrical signals. These signals are then transmitted via retinal ganglion cells, whose long appendages run along the optic nerve and spread out to various parts of the brain's visual-processing bits. But if the optic nerve or the ganglion cells have been damaged through injury or illness, they stay damaged. They wo not just grow back like your olfactory sensory nerve.


Continue reading at Engadget »