NHS England response to early value assessment on using AI for assessing and triaging skin lesions

What does the early value assessment say?

The National Institute for Health and Care Excellence’s (NICE) early value assessment ‘Artificial intelligence (AI) for assessing and triaging skin lesions referred to the urgent suspected skin cancer pathway’ (HTE24) conditionally recommends Deep Ensemble for Recognition of Malignancy (DERM) for use in the NHS over a 3-year period while further evidence is generated to confirm its clinical and cost-effectiveness.

DERM is an AI-based skin lesion analysis device with a Class III CE mark that is manufactured by Skin Analytics. It has the potential to free up dermatology capacity and reduce waiting times by effectively triaging patients with skin lesions where there is a suspicion of cancer.

DERM can be used within tele-(virtual) dermatology services after referral from primary care. It is authorised for use as an automated tool or with a healthcare professional review (known as a ‘second read’) to decide if further assessment by a dermatologist is needed.

A smartphone is used to take images of skin lesions using a dermoscopic lens attachment, and the images are uploaded to the online platform. The DERM platform uses an AI-based fixed algorithm to analyse the dermoscopic images and provide a suspected diagnosis of the lesion. If DERM labels the lesion as benign, the person can be discharged from the urgent suspected skin cancer pathway. They will be sent the results with advice to continue monitoring their lesion and to seek further advice if it continues to change or causes concern.  If DERM labels the lesion as pre-cancer or cancer, an NHS dermatologist will review the case and decide on an appropriate management plan for the person.

In 2023/24 over 650,000 people were referred onto the urgent suspected skin cancer pathway. 120,000 people did not have cancer diagnosed or ruled out within 28 days (the standard). DERM has the potential to free up capacity by discharging around 30% of cases from the pathway. This will allow more people to be seen more quickly and give them faster diagnoses .

In draft guidance, NICE’s independent diagnostics advisory committee had initially recommended more research. After detailed review, a practical ‘use and learn’ approach balancing innovation with patient safety is now recommended. This will allow patients to benefit from DERM within the NHS while further evidence is generated.

NHS England’s response to the early value assessment

NHS England welcomes this early value assessment’s conditional recommendation and the potential for DERM to help increase capacity in dermatology services.

NICE’s assessment indicates that DERM may be able to detect cancer with a similar level of accuracy to a tele-dermatologist or a face-to-face dermatologist review. However, to ensure patients are involved in the decision-making about their care and to support sites to implement DERM safely and effectively, NHS England recommends:

• patients are provided with information and informed consent is obtained prior to DERM being used to analyse images of their lesions. NHS England has developed a template consent form and information leaflets, which can be adjusted for local use. These documents, along with other implementation guidance, will be included in a national implementation toolkit
• when DERM is first introduced at a site, the site should consider using it initially with a second clinical read. Data should be collected locally to monitor the diagnostic accuracy of both DERM and the clinician. Once the site is confident in DERM, they can move to autonomous use
• for patients with black or brown skin, a second read should take place. This should be done throughout the 3-year evidence generation period. Evidence generated by DERM to date shows that performance in these populations is similar to the broader population. However, since the incidence of skin cancer in these groups is lower and they represent fewer of the cases to date, NHS England recommends a second read to ensure accuracy
• data should be collected by trusts in collaboration with Skin Analytics within the framework of NICE’s evidence generation plan (for example, recording dermatology resource use before and after AI implementation). This will help assess the cost-benefit of using DERM plus teledermatology versus teledermatology alone. NHS England will establish an independent data oversight group to offer further guidance on the specific metrics to be collected and real-world study designs.
• sites looking to implement DERM should join the AI in Dermatology Getting It Right First Time (GIRFT) community of practice to learn from other sites’ experiences. To join, contact furtherfaster.girft@nhs.net

If you have any questions about this guidance, email england.digitalinnovation@nhs.net.