UK tech company Skin Analytics has received regulatory approval for DERM - an autonomous, AI-powered skin cancer detection system using a smartphone. The technology allows technicians to photograph a patient's lesion and receive a near-instant diagnosis.
When the sun shines, people generally take off their clothes and go to the beach.
The sensation of warm rays on the skin is a pleasant thing for sun lovers. But these rays can also trigger skin cancer, one of the most treatable types of cancer if detected early.
In the UK, according to the National Health Service (NHS) England urgent referrals for skin cancer have skyrocketed by 170 per cent in 10 years and 11 per cent of urgent cases are waiting more than a month just to be seen.
The reason is that the number of dermatologists is very limited in Europe. On average, there are only 30 dermatologists per million population.
In London, Skin Analytics demonstrated its new DERM AI technology. This simple app runs on most smartphones. With an additional lens called a dermascope , the app allows technicians to photograph a patient’s lesion and receive a near-instant diagnosis.
Product manager for DERM, Kirsten Vehlow, showed how fast the process is. She demonstrated it to a colleague and said:
"It's pretty simple and easy. So, as you can see here, I open up the app which has a camera here. And then this is a dermatoscope that we can basically plug into our phone, turn on the light, and then we can see if I take a picture of Audrey's lesion, we can see the lesion very clearly. If I decide to use this picture, the app will check the quality of the picture to make sure that it's good enough for the AI to examine. And then I send it to examine it here. And within a few seconds, we get the results right on our phone. We can see the results in more detail about what they mean. And the picture will be stored there with the results and the location of the lesion," Vehkow said.
This AI technology compares a single photo of a lesion against a library of images before providing a diagnosis.
Skin Analytics says DERM achieves a 99.8 percent accuracy rate in determining whether or not a disease is cancerous.
Neil Daly, founder and CEO of Skin Analytics explains that the technology to support DERM is already available – usually in people’s pockets.
He said: "This technology is so accessible because you can use a variety of smartphones to take the images. We use it with a dermatoscope that is mounted on a smartphone to make sure we get very high quality images which allows us to see a lot of patients very, very quickly because of this hardware."
DERM has been awarded the European Union CE Class III medical device designation under the European Medical Device Regulation, making it the world’s first autonomous AI authorised to detect cancer. DERM has been used at 21 NHS sites in the UK and has screened 135,000 patients. The tool has also helped detect more than 13,000 cancers, says Skin Analytics.
"DERM is a tool that is another asset that we can leverage to try and reduce the number of people who die from cancer. When we talk about health care, it's a very human discipline. And at the end of the day, diagnostics is just one small part of that process. So by using technology like DERM to be able to see more patients, to find the right patients to see a dermatologist, we can really increase the efficiency of the health care system and be able to see more patients more quickly and give them better outcomes," Daly added.
DERM has been in development for the past 12 years and the AI relies on a large dataset of photos including skin cancers, pre-cancerous skin conditions, and even harmless skin cancer mimics.
Daly added, "So the way DERM works is we train it on a large set of images. Once it has that collection, it comes back with pathology results for biopsied lesions. And the AI is able to track features that can identify cancerous and precancerous lesions. And then the AI is able to learn from those features so that when we feed it a new image, it can identify, is this something we need to be concerned about, do we need to take action, or is this something that is not concerning and we can safely send the patient home?"
One NHS patient who used DERM was Cherry Thompson. She found a lesion and was concerned it might be melanoma. She had never heard of AI diagnosis for skin cancer but found the process efficient.
"I think this is probably a step forward and it might be very useful. I know a few dermatologists and one of them is probably the best in the country, and I have a lot of confidence in them. That said, maybe computer diagnosis could be more efficient, I don't know," Cherry said.
Thompson no longer has to worry because the lesion on his skin is not melanoma. This reassuring news came much sooner thanks to DERM.
The sun exposure in the UK is not as much as Neil Daly is used to in his native Australia. This, Daly says, makes people in the UK not feel the need to protect their skin from sun exposure. As a result, the risk of skin cancer is greater.
"I'm really interested in skin cancer because I grew up in Australia, where we're two to three times more likely to get cancer in our lifetime. So, almost every Australian will get skin cancer in their lifetime. But skin cancer is also one of the few cancers that is visible on the surface of the skin, so we can still do something about it. So, in the early stages of skin cancer, the cure rate is almost 99%, about 97%. So if we can get it treated, the mortality rate is almost zero. That's in contrast to a lot of other cancers. So, I'm really, really interested in making sure that we're applying this technology to really improve the cure rate for cancer," Daly said.
In NHS settings where DERM is used, the technology has reduced the number of dermatologist appointments by between 60 and 95 per cent, freeing up time for dermatologists to deal with more serious cases. DERM can automatically reduce urgent skin cancer referrals by up to 40 per cent.
We may be in the grip of winter right now, but summer is just around the corner. With summer comes the sun.
DERM could be an important tool for early cancer diagnosis and potentially save many lives.
Post a Comment for "AI App Claims 99.8% Accuracy in Detecting Skin Cancer"