Clinical Trials

AI tool detects early signs of eye disease

AI tool detects early signs of eye disease 1024 683 Toku Eyes

The trial of an Artificial Intelligence retinal screening tool, developed in New Zealand, showed it did not miss any referable disease for diabetic patients at risk of eye disorders.

THEIA has been created by Toku Eyes and was tested for diabetic retinopathy. Co-founder Ehsan Vaghefi says the tool is being further developed to vastly improve early detection of eye disease such as macular degeneration and glaucoma.

Its use was trialed at Counties Manukau DHB in its central and satellite units, as well as a private optometrist in Palmerston North.

More than 900 patients, who are part of the diabetic screening programme, had their de-identified images screened as usual by retinal specialists, as well as by THEIA.

The study, conducted between March and May of this year, found the AI did not miss any patients with more than mild or vision threatening disease and the level of agreement with the clinicians was between 91-98 percent.

Vaghefi says people wait around six weeks to get their results from the current screening process, but with THEIA they would get them instantly. He says the tool could expand existing screening capacity by 50 percent and reduce labour costs by nearly a third, allowing more patients to access these services, including those in under-served areas.

Diabetic retinography is a leading cause of blindness in working age people, but there are screening backlogs nationwide.

Jacob Benefield, optometrist at Naylor Palmer Optometry in Palmerston North, says:

“Toku Eyes shows how AI can really help increase proactive retinal screening, maximising the time optometrists and ophthalmologists can spend with patients whose needs are more complex. Putting this tool into the hands of lab technicians and medical imagers will expand screening services, ensuring more people are seen sooner and improving their chances of early detection of diseases like diabetes,” 

Toku Eyes recently raised $3.6 million from investors to advance the commercialisation of THEIA and is applying for FDA approval in the US. Vaghefi says the AI platform is designed to handle lower quality images and anyone can be trained to operate the necessary machine, freeing up the time of specialists to focus on the 15-20 percent of people who need a consultation.

The use of AI also makes screening more accessible to people who struggle to attend hospital appointments. He would like to see a national diabetic screening programme developed with satellite units nationwide, supervised from the centre.

If you would like to provide feedback on this news story, please contact the editor Rebecca McBeth.

NEWS – editor Rebecca McBeth
Article originally on

Top scores for Kiwi AI screening tool

Top scores for Kiwi AI screening tool 1024 683 Toku Eyes

Toku Eyes has completed its first prospective clinical trial of its artificial intelligence (AI) screening platform Theia, designed to support more cost-effective diagnosis and prognosis of eye disorders.

The trial, which focused mainly on diabetic retinopathy (DR), included 1,000 patients from Middlemore Eye Hospital and Naylor Palmer Optometry. Revealing topline results in June, Theia achieved a sensitivity of 98.9%, a specificity of 92.6%, a negative predictive value of 99.5% and an imageability rate of 99.9%, which is better than any other FDA-approved AI imaging tool in the world, said Associate Professor Ehsan Vaghefi, Toku Eyes co-founder and CEO.

“Theia is the most accurate algorithm of its kind available today, while providing the highest level of diagnostic granularity. This diagnostic tool has the potential to vastly improve early detection of eye disease.”

In Toku Eyes’ economic modelling, in New Zealand alone, Theia could expand existing screening capacity by an estimated 50% and reduce labour costs by 30%, said Dr Vaghefi. Camera agnostic, Theia is designed to handle lower quality images, works with any number of images per eye and automatically identifies images belonging to the same eye and the same patient, he explained. “Using a practice’s existing camera, operated by a technician after basic training, Theia frees up specialists’ time currently spent scanning and writing reports for every patient, when just an estimated 5-10% actually need a referral.”

As well as reliably detecting DR and several other diseases, Theia provides a patient-specific disease progression rate, which can be used in educational conversations with the patient about preventative measures, said Dr Vaghefi.

Naylor Palmer optometrist Jacob Benefield said the platform is user-friendly and doesn’t appear to miss anything. “Occasionally it might flag an epiretinal membrane as retinopathy but you’d rather it flag too many things than miss something.”

Benefield said using Theia means he saves enough time to see one, maybe two more potential DR patients a week. “Where I normally spend time grading the images, Theia now grades them for me, using a traffic light system, where green is clear, amber indicates that I need to have a look at the photos and red requires a referral.”

The real potential time saviour, however, is if a non-optometrist staff member could do the scans for him, he said. “I would then only step in as required for those amber and red alerts. The photos are saved in the system, so those could be looked at later in the day.”

Naylor Palmer is keen to continue using Theia beyond the trial, said Benefield, but ultimately it depends on what Mid-Central DHB (or Health NZ) want to do.

Toku Eyes also announced it raised $3.6 million from investors, including Australasian ophthalmologists, to progress Theia’s capabilities. “This funding round will advance Theia’s commercialisation and complete validation of its HbA1c (blood sugar), cataract developments, macular degeneration and hypertension diagnosis programmes, through partnership with US-based clinics,” said Dr Vaghefi.

Article originally on

Toku Eyes raises millions for AI Tech that helps stop diabetes, preventable blindness

Toku Eyes raises millions for AI Tech that helps stop diabetes, preventable blindness 1024 683 Toku Eyes

Auckland startup Toku Eyes has raised $3.6 million to develop its AI software – which founder Associate Professor Ehsan Vaghefi says can analyse eye images faster than a human clinician and pick up things they might miss, helping to spot conditions such as type 2 diabetes – the most common cause of blindness or vision impairment in New Zealanders aged 20 to 60.

The medtech company doubles as something of a personal crusade for Vaghefi.

“When I was going to school when I was little, other children would get help crossing the road and I would be helping my father across,” he tells the Herald.

His dad was just 4 years old when he went blind as a result of congenital glaucoma – a preventable condition, even then.

“He’s probably the most talented person I know,” Vaghefi says. “He grew up to become a lawyer then a law professor. But he could have achieved even more with his sight.”

That was decades ago, in Iran, but Vaghefi says that today, in 21st-century New Zealand and other developed countries, many cases of preventable blindness or vision impairment are still missed. Around 270,000 Kiwis have been diagnosed with type II diabetes, but experts estimate the true rate could be 30 to 40 per cent higher.

Vaghefi says his company’s software will help to make analysis of fundus camera scans cheaper and easier to analyse, making ocular care more accessible to those on lower incomes.

A fundus camera is a combination of a digital camera and a microscope that takes an image of the back of your eye. Because your eye is transparent, its the easiest way to get a picture of your veins, which in turn can reveal early signs of cardiovascular problems or conditions like diabetes – which has implications for your the health of your whole body, and which can also cause macular degeneration that impacts on your vision.

Toku Eyes’ AI software, called THEIA, can also catch the signs of hypertension (high blood pressure), cataracts or early signs of kidney disease, among other conditions.

“The AI sits in the cloud. If a fundus camera has an internet connection, it can upload an image that THEIA analyses in 10 seconds, down to the pixel level.”

Challenge to the status quo

It’s a challenge to the traditional business model of private practices in which highly trained optometrists manually view the scans of every patient. AI means these specialists can focus on the more complex cases and staff with less training can operate the scanner (a similar concept to other local medtech startups employing artificial intelligence – the Peter Beck-backed HeartLab in the cardiac field and Formus Labs in orthopedic implant).

Vaghefi hopes the tool will expand existing screening capacity by up to 50 per cent and reduce labour costs by a third, allowing more patients to access ocular, including Maori and Pasifika people in under-served areas, who disproportionately suffer type 2 diabetes.

“A lot of people still go blind for needless reasons, even in this country,” he says.

“THEIA can clinicians more quickly work out who needs supervision or specialist care. At the moment, identifying the 10 per cent who might need help takes a lot of time.”

Toku Eyes’ THEIA software uses artificial intelligence to analyse a retinal image then return a result in around 10 seconds. Image / Supplied

Vaghefi first conceived of what became THEIA while working on imaging technology during his PhD in biomedical engineering at Auckland University (where he remains a senior research fellow and senior lecturer on top of his startup duties).

After graduating, he formed Toku Eyes (or “My Eyes”) and set about developing his AI product with help from local ophthalmologist and options and capital from Uniservices (Auckland University’s commercialisation arm), Icehouse Ventures and the rich-list Mafsen family.

On top of the $3.6m just raised, the company has also received around $4m in grants from organisations including the Health Research Council, MBIE (through its National Science Challenge) and Callaghan Innovation.

Today, he has 10 staff and THEIA is going through the US Federal Drug and Administration, which should serve as a springboard for approval in other territories. Meanwhile, THEIA successfully completed its latest clinical trial in May 2021 at Middlemore Eye Hospital, Vaghefi and is being used in private practices across the country.

China crisis

Right now, it’s being used to review analysis that’s already been carried on image scans. Once it gains regulator approval, Vaghefi sees the AI being used for real-time anal􀁜sis.

He hopes FDA approval for Toku Eyes first wave of software will arrive by the end of this year. If it does, he sees his company going to market in early 2022.

He sees the US and New Zealand as Toku Eyes primary markets initially, but will also use around a third of the $3.6m just raised on entering China – where that county’s burst of wealth in recent decades has led to an explosion of bad eating and, according to some estimates, somewhere between 100m to 150m people suffering type 2 diabetes.

Development continues, and the founder is aiming to soon have an algorithm that can spot glaucoma.

Backing from US non-profit

The oversubscribed funding round just closed – Toku Eyes first external raise – was led by Icehouse Ventures and supported by a second local venture capital outfit, Artemis Capital and US non-profit Kera Link International, a social venture that invests in technology to address blindness in developing countries.

The funds were raised at a pre-money valuation of $8.5m, and at the end of the day, Vaghefi wants to make profits – the better to fund his socially conscious startups, and to expand to new initiatives such as cheaper, more accessible fundus cameras.


Article originally published on

By: Chris Keall
Business Writer, NZ Herald

You're viewing our
UK site.

For other regions, please click below: