Skin Analytics

Home Innovations Skin Analytics

Next Previous

Skin Analytics

5th March 2019

Skin Analytics has developed AI algorithms that can assess dermoscopic images of skin lesions to help to identify skin cancer.

“Two-week rule referrals place an enormous burden on Dermatology services across the UK. However, typically fewer than 10% of such referrals turn out to have a significant skin cancer. Teledermatology offers the potential to greatly improve the current referral pathway and ensure that patients with skin cancer receive care, in both a quick and efficient manner.”
Niall Wilson, Skin Analytics Medical Director and consultant Dermatologist, Broadgreen Hospital

Summary:

By enabling high quality screening in Primary Care within current appointment times and without the need for expensive equipment, Skin Analytics’ solution supports:

More accurate identification of melanoma, leading to potentially better health outcomes and reduced treatment cost
Less onward referrals to secondary care, reducing strain on specialist clinics and lowering the cost of finding melanoma

GP practices are provided with a dermascope and image capture device. This can be used during a consultation to capture an image of any pigmented lesions a GP would select for referral. An artificial intelligence (AI) algorithm identifies: suspected melanoma, common types of nonmelanoma skin cancer, benign (noncancerous) lesions.

Challenge/problem identified:

Some areas of the UK have 0.64 dermatologists per 100,000 population, compared to the Royal College of Physicians recommendation of 1.6.

Skin Analytics’ solution can therefore relieve pressure on a specialism which in 2014 required more than 13 million GP appointments*.

Anecdotal evidence provided by Skin Analytics’ dermatologist partners suggests that over 90% of two-week referrals they receive are in the wrong patient pathway. By providing an accurate way to stratify referrals by suspect melanoma and suspect nonmelanoma skin cancers, GPs are empowered to refer through the appropriate channel,
thereby reducing the demand on two-week wait referrals.

*http://www.bad.org.uk/shared/get-file.ashx?id=2348&itemtype=document

Impact:

Based on the 2015 NICE Melanoma Evidence review, GPs operate with a sensitivity and specificity of 60% and 72% respectively when assessing melanoma. Using dermoscopy, where trained, this increases to 75% and 78%. Skin Analytics’ solution operates at >95% sensitivity and >70% specificity (comparing well to dermatologists at 88% and 90%)
Reduced onward referrals by around 50% where trialled. An NHS study conducted in Bristol in 2011, found that reducing the onward referral around this level can save £43,000 per 100,000 population

Deploying Safe and Effective AI in Skin Cancer Pathways

Download a list of supporting evidence here.

Skin Analytics is supported by the NHS Innovation Accelerator in its capacity as a support tool within the cancer pathway, not as a replacement for diagnosis by a healthcare professional. For the most current information on Skin Analytics’ regulatory status, please contact Neil Daly at neil@skinanalytics.co.uk.

Fellow: Neil Daly