How many times have you noticed a strange new freckle, a slight change in a mole or an oddly coloured patch of skin on your body and chosen to just ignore it? For me, and for many other people, the answer is more often that we’d care to admit.
I’m a natural redhead, so freaking out over every mole and freckle would quickly become a full-time job. But there are a few that, if I had all the time and money in the world, I’d get checked out privately. Or if the NHS had all the time and money in the world, I’d get checked out through my GP.
Unfortunately, neither of us have all the time and money in the world. We’re both pretty skint and knackered, actually. For this reason, many people avoid investigating potential skin cancers, putting them off due to lack of “severity”. But that poses a significant risk. Catching skin cancer early is everything.
See also: What causes cancer?
Take melanoma, for example. It’s the most serious form of skin cancer, it can spread to other areas of the body and is typically caused by sun exposure. A melanoma caught at stage one has a 100 per cent survival rate, whereas the survival rate of a melanoma caught at stage four is 16 per cent.
But what if that reality was changing? What if I could obtain a skin cancer diagnosis within one appointment, and one 30 second scan? What if I didn’t need to involve a doctor at all? What if, eventually, my phone could detect skin cancer through a selfie?
This is the future that Skin Analytics, an award-winning AI-powered technology company based in London, is working towards.
Skin Analytics has developed an AI-assisted medical tool, “DERM”, to assess potential skin cancers from a photograph without the need for a doctor. It’s been in use at the NHS since 2020, but the company was just given the first ever regulatory approval to detect skin cancer with AI alone, no human intervention required.
In what Skin Analytics are calling a “watershed moment” for AI in health, this makes DERM the world’s first AI tool which is legally authorised to independently make clinical decisions on skin cancer without oversight.
And it’s not some distant future. It’s already happening: the process is currently available to BUPA and Vitality customers as well as 21 NHS sites, where it is offered for free. So, me and my moles wanted in on it.
I headed to Skin Analytics’ Clerkenwell offices to test it out and finally put my mole-related misgivings to bed. Practically, it currently works like this: you have one appointment with a medical professional (mine was Kirsten, a former physician associate for the NHS and current Skin Analytics product manager) who photographs any “lesions” you’re concerned about using an iPhone with a lens attachment called a dermatoscope. They’re inputted to the system in the same speed as a Google search and then the AI tool scans them.
My three suspicious lesions took less than 30 seconds to be scanned, with the AI reporting them as having “low suspicion of malignancy”, aka they are deemed cancer free.
.jpeg?trim=334%2C0%2C516%2C0)
Can it be trusted? Indisputably, yes. In fact, it’s even more accurate than a human. A report commissioned by NHS England found that Skin Analytics’ Artificial Intelligence as a Medical Device (AIaMD) achieves a 99.8 pre cent accuracy rate in ruling out cancer, compared to the 98.9 per cent accuracy rate of dermatologists. It’s quicker, too. Last year, DERM saw nearly double the amount of cases that a dermatologist would see over the course of their entire career.
The NHS needs this, badly. The UK has seen a 170 per cent rise in urgent suspected cancer (USC) referrals over the past decade and 11 per cent of these urgent cases wait over a month just to be seen. There is a significant shortage of dermatologists in the UK and abroad, with an average of 30 dermatologists per 1 million population in Europe, even though skin-related issues consistently rank among the top 10 reasons for initial doctor visits.
And to make things even speedier, Skin Analytics is aiming for DERM to eventually function via selfie alone, with the tool ideally available via the NHS app. A person currently needs to attend one in-person appointment for their 30-second DERM scan to be taken by a physician, but there’s a future where they might not have to see anyone at all. They wouldn’t even have to get out of bed to submit it from their phone.
In a nutshell: I left my screening 20 minutes after I arrived, blissfully footloose in the knowledge I am skin cancer-free. Meanwhile, people without access to the NHS sites and insurances offering it will wait up to four weeks following a referral for their urgent suspected cancerous lesions (and that’s not even considering the time it took to get to that point) — whereas mine were just a slight concern. With so many people in the UK scared of overburdening their GPs and, in turn, the NHS, maybe a non-human really is the only way to soothe us British worrywarts.
Symptoms of skin cancer
Cancer Research warns that skin cancers can look very different and the symptoms can vary. Some of the symptoms are also similar to other conditions.
Common symptoms of skin cancer include a sore or area of skin that:
- doesn't heal within 4 weeks
- looks unusual
- hurts, is itchy, bleeds, crusts or scabs for more than 4 weeks
What to look out for
A sore that doesn't heal: The sore can look see through, shiny and pink or pearly white. It can also look red. It may feel sore, rough and have raised edges.
Ulcer: Look out for an area of skin that has broken down (an ulcer) and doesn't heal within 4 weeks, and you can't think of a reason for this change.
A lump: This might be small, slow growing, shiny and pink or red.
Red patches on your skin: These red patches could also be itchy. This could be due to other non cancerous skin conditions. But get it checked to make sure.
Freckles or moles: A change to a mole or freckle can be a sign of another type of skin cancer called melanoma.
Skin Analytics is available to BUPA and Vitality customers and via the NHS in 21 different sites.
