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're 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're 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're 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 won't just grow back like your olfactory sensory nerve.

However, the Stanford team found that subjecting mice to a few weeks of high-contrast visual stimulation after giving them drugs to reactivate the mTOR pathway, which has been shown to instigate new growth in ganglion cells, resulted in "substantial numbers" of new axons. The results are promising, though the team will need to further boost the rate and scope of axon growth before the technique can be applied to humans.

Researchers from Japan have recently taken this idea of cajoling the retina into healing itself and applied it to age-related macular degeneration cases. AMD primarily affects people aged 60 and over (hence the name). It slowly kills cells in the macula, the part of the eye that processes sharp detail, and causes the central focal point of their field of vision to deteriorate, leaving only the peripheral.

The research team from Kyoto University and the RIKEN Center for Developmental Biology first took a skin sample from a human donor, then converted it into induced pluripotent stem (IPS) cells. These IPS cells are effectively blank slates and can be coerced into redeveloping into any kind of cell you need. By injecting these cells into the back of the patient's eye, they should regrow into retinal cells.

In March of this year, the team implanted a batch of these cells into a Japanese sexagenarian who suffers from AMD in the hope that the stem cells would take hold and halt, if not begin to reverse, the damage to his macula. The team has not yet been able to measure the efficacy of this treatment but, should it work out, the researchers will look into creating a stem-cell bank where patients could immediately obtain IPS cells for their treatment rather than wait months for donor samples to be converted.

And while there isn't a reliable treatment for dry-AMD, wherein fatty protein deposits damage the Bruchs membrane, a potent solution for wet-AMD, which involves blood leaking into the eyeball, has been discovered in a most unlikely place: cancer medication. "Genentech started developing a new drug when an ophthalmologist in Florida just decided to inject the commercially available drug into patients eyes," Schallhorn explained.

"Generally this is not a great idea because sometimes things will go terribly wrong," she continued, "but this worked super-well. It basically stops and reverses the growth of these blood vessels." The only problem is that the drugs don't last, requiring patients to receive injections into their eyeballs every four to eight weeks. Genentech and other pharma companies are working to reformulate the drug -- or at least develop a mechanical "reservoir" -- so it has to be injected only once or twice a year.

Stem-cell treatments like those used in the Kyoto University trial have already proved potentially effective against a wide range of genomic diseases, so why shouldn't it work on the rare genetic condition known as choroideremia? This disease is caused by a single faulty gene and primarily affects young men. Similar to AMD, choroideremia causes light-sensitive cells at the back of the eye to slowly wither and die, resulting in partial to complete blindness.

In April of 2016, a team of researchers from Oxford University performed an experimental surgery on a 24-year-old man suffering from the disease. They first injected a small amount of liquid into the back of the eye to lift a section of the retina away from the interior cellular wall. The team then injected functional copies of the gene into that same cavity, replacing the faulty copies and not only halting the process of cellular death but actually restoring a bit of the patient's vision.

Gene therapy may be "surely the most efficient way of treating a disease," lead author of the study, Oxford professor Robert MacLaren, told BBC News, but its widespread use is still a number of years away. Until then, good old-fashioned gadgetry will have to suffice. Take the Argus II, for example.

The Argus II bionic eye from Second Sight has been in circulation since 2013, when the FDA approved its use in treating retinitis pigmentosa. It has since gotten the go-ahead for use with AMD in 2015. The system leverages a wireless implant which sits on the retina and receives image data from an external camera that's mounted on a pair of glasses. The implant converts that data into an electrical signal which stimulates the remaining retinal cells to generate a visual image.

The Argus isn't the only implantable eyepiece. French startup Pixium Vision developed a similar system, the IRIS II, back in 2015 and implanted it in a person last November after receiving clearance from the European Union. The company is already in talks with the FDA to bring its IRIS II successor, a miniaturized wireless subretinal photovoltaic implant called PRIMA, to US clinical trials by the end of this year.

Ultimately, the goal is to be able to replace a damaged or diseased eye entirely, if necessary, using a robotic prosthetic. However, there are still a number of technological hurdles that must be overcome before that happens, as Schallhorn explained.

"The big thing that's holding us back from a fully functional artificial eye is that we need to find a way to interface with the optic nerve and the brain in a way that we transmit signals," she said. "That's the same problem we're facing with prosthetic limbs right now. But there are a lot of smart people in the field working on that, and I'm sure they'll come up with something soon."

Read more:
High-tech solutions top the list in the fight against eye disease - Engadget

Comments are closed.