By improving access to services and reducing demand for face-to-face appointments, Skin Analytics is making a big difference to dermatology departments across the UK.

Since its founding in 2012, research-led company Skin Analytics has been committed to helping more people beat skin cancer. They use AI to catch skin cancer early:

‘The machine learning algorithm built by Skin Analytics, Deep Ensemble for the Recognition of Malignancy (DERM), recognises the most common malignant, pre-malignant, and benign skin lesions.’

This enables dermatology departments to rapidly triage patients so that they can receive care as soon as possible. This makes the identification of skin cancer quicker and easier, improving patient outcomes.

Rapid action

Skin Analytics had noticed that stage one skin cancer survival rates were as high as 97%, so they made it their goal to ensure that every single patient who goes to see their doctor about their melanoma does so in stage one.

The challenge lay in minimising the amount of time between starting to worry about an unusual mole to actually getting in front of a specialist.

DERM saves ‘unnecessary hospital appointments for patients and dermatologists. The Skin Analytics system also enables local dermatologists to review referred cases remotely, allowing them to direct patients to the best assessment or treatment option.’

DERM takes seconds to assess patients, allowing them to immediately seek further care or have their worries put to rest.

Tried and tested

Skin Analytics client Simon Miller volunteered to have one of his moles scanned as part of a demonstration and ended up proving how vital the app truly is: it flagged one of his moles as cause for concern, and it transpired to be a malignant melanoma. The early detection meant that a few months later Simon was cancer-free.

A study found that DERM could be as accurate in diagnosing suspected skin cancer as clinical specialists; it identified 100% of melanomas, with over half in the earliest stage of malignancy.

Assisting clinicians, not replacing them

Automation of their jobs isn’t something doctors would have anticipated. This development won’t displace dermatologists from their roles though, it will make them easier.

Skin Analytics CEO Neil Daly stated that we only have the equivalent of 650 full-time dermatologists across the UK and they are currently dealing with over 1million dermatology appointments each year. So this AI won’t be risking any jobs, it will be helping with workloads by allowing clinicians to spend more time on treatment than diagnosis.

NHS GP and advisor to global health and technology companies, Dr Harpreet Sood, commented that:

“This technology provides a potential tool for clinicians to better manage more dermatology cases in the community, enable appropriate referrals and reduce variation in diagnosis and management of skin-related cases.”

This reduction of variation in diagnosis is key to Skin Analytics’ goal of providing equal access to healthcare.

Transforming patient pathways

The UK is facing a real capacity constraint in the health service, especially during the COVID-19 pandemic. This has had knock-on effects for cancer care with many screenings being postponed. MacMillan analysis has found that ‘there are 50,000 people now missing a cancer diagnosis who are at risk of becoming the forgotten ‘C’ of the coronavirus crisis.’

With the NHS facing daunting wait times amidst the COVID-19 pandemic, health apps are speeding up access to care and helping clear the patient backlog.

The rapid turnaround of the Skin Analytics app has been an absolute saviour amidst the pandemic; DERM has the potential to reduce the number of patients needing face-to-face appointments by 40-60%.

DERM’s challenges

This is not an arena where an algorithm can afford to make mistakes, so Skin Analytics opted for a step up from retraining an existing neural network:

Skin Analytics have designed ‘all aspects of our machine learning architecture from the ground up for the specific problem we are trying to solve. This includes specifically tailored machine learning architectures, training methodology, and data augmentation for detection of skin cancer.’

Training machine learning algorithms is an incredibly complex process, and part of training DERM involved showing it as many moles and lesions as possible. I actually came across the development of this app when at the Dermatology Department myself, patients were asked if they’d like to participate by allowing the team to take pictures of their moles so the AI could learn to identify them.

It’s key that AIs such as DERM are trained using images of as many different skin colours as possible so they can accurately detect skin cancers across the population. Skin diseases are often pictured on white skin, which makes finding skin disease information online difficult for people of colour – this kind of health inequality is among those Skin Analytics aims to combat.

Accessing Skin Analytics

Skin Analytics is currently running an NHS pilot study with University Hospitals Birmingham. Patients are referred to Skin Analytics for triage instead of going to the hospitals themselves – the aim is to flatten the demand curve and reach that goal of treating as many patients at stage one as possible.

They are also working with Bupa to provide a new remote skin assessment service:

“Available the next working day, customers will receive a special kit through the post which includes a smartphone and a dermoscopic (magnifying) lens to take high-resolution photos of moles or lesions.”

If you are concerned about a mole, the NHS ABCDE checklist can help differentiate between a normal mole and a melanoma. A key thing to look out for is change – whether that’s in size, shape, or colour – as melanome is a mutation by definition. Skin Analytics can help monitor these changes and provide tips on when to worry about a mole regardless of if you access their services.

What next?

Skin Analytics is growing rapidly: they recently raised £4million in Series A funding, which CEO Neil Daly stated will be used to expand NHS partnerships and branch out into the US. They recently received the FDA’s breakthrough device designation, which will fast track their entrance onto the US market.

Conclusion

Skin Analytics overall aim is to improve access to skin cancer assessments and reduce costs to the health system – and they are certainly succeeding.

Their support has been invaluable during the COVID-19 pandemic and the speed with which DERM can offer specialist level assessment will cement its position as a key part of the future of dermatology.

About the Author: Leo Hynett

Leo Hynett is a contributing Features Writer, with a particular interest in Culture, the Arts and LGBTQ+ Politics.