Bradford Teaching Hospitals is the first NHS Trust in West Yorkshire to use trailblazing artificial intelligence (AI) which aims to detect skin cancer earlier.
The new approach starts at St Luke’s Hospital, part of the Foundation Trust, this Thursday (April 30) and will see healthcare staff using AI software to rapidly analyse skin lesions and flag small moles that are most likely to be cancerous, with far greater precision than standard techniques.
The DERM (Deep Ensemble for Recognition of Malignancy) technology, developed by Skin Analytics, analyses images to assess and triage skin lesions, redirecting benign cases to non-urgent pathways and significantly reducing waiting times by efficiently triaging patients with suspicious skin lesions.
Consultant Plastic Surgeon and Clinical Lead for skin cancer, Zakir Shariff, hailed the project as
cutting edge and the future of skin cancer diagnosis in this country.
Combining this cutting-edge AI will give us the capacity to pick up potentially serious skin lesions quicker and speedier than current processes, he continued.
We have 5,000 referrals every year for skin cancer at the Trust – all of whom are seen within the two-week cancer referral-to-treatment pathway – yet only 8% or 400 patients are found to have malignant cancer.
DERM technology can instantly pick up potentially serious skin lesions, making diagnosis far speedier than current human processes allow so it will also help our doctors and surgeons concentrate on treating the most urgent cases.
Skin cancer contributes to a whole year of the nine-year life expectancy gap between richer and poorer parts of England, and the National Cancer Plan will prioritise solving inequalities to ensure patients can get fairer and faster access to cancer care.
After a patient has been referred by their GP into the new, thrice-weekly tele-dermatology service, healthcare staff will take detailed photographs of any suspicious skin lesions.
An algorithm will analyse each image for detailed, visual characteristics to provide a suspected diagnosis and direct the most appropriate next steps for patient care. If a lesion is identified as suspicious, the patient will be directed to the consultant dermatologist for further investigation in the ‘one stop clinic’ located next door to the ‘image capturing’ clinic and running alongside it.
The dermatologist will see patients whose mole is diagnosed as cancerous and perform an immediate excision which will be sent to the laboratory for diagnosis. If found to be cancerous, some patients will require further treatment, while others will have their treatment completed there and then, being discharged to outpatient clinics for follow-up care. Those with benign moles and other skin conditions are reassured and offered advice before being discharged.
General Manager for Muscloskeletal and Therapies Clinical Support Unit which covers Plastics and Dermatology, Tom White, explained:
At a time of increasing referrals to dermatology services for suspected skin cancers, early evidence suggests automated use of DERM could be a game changer. It will significantly reduce the number of urgent referrals that require review by dermatologists, reducing waiting times and allowing us to concentrate on the most urgent cases.
We will also have the capacity to see this service go out into the community and GP surgeries which means that, in the future, patients won’t need to come to hospital which we know is more stressful for many.
This new AI highlights higher-risk areas immediately, meaning doctors can take urgent biopsies immediately and send them to our specialist laboratory to be reviewed by our expert cancer team to confirm or rule out cancer.
For many patients, weeks of biopsies and awaiting results can be avoided as the process will reduce prolonged uncertainty and help avoid more invasive surgery.
This AI will help us triage patients into the right place, at the right time, conserving capacity for patients who need us the most as DERM safely triages patients with non-cancerous lesions away from cancer pathways.
Across the UK, Dermatology services receive one million referrals each year from primary care. About 60% are urgent referrals for suspected skin cancer. Of these, typically 6% are confirmed to be skin cancer and the remaining 94% are either non-urgent or non-cancer cases.
DERM represents an important innovation that aligns with the Government’s health priorities to move “from analogue to digital” by rolling out new digital technologies to benefit patients, reduce waiting lists and modernise the NHS.
The first year of the new three-year scheme will see Bradford’s data scrutinised by the Cancer Alliance for evidence and safeguarding. The AI programme is currently running in 25 other NHS trusts around the UK and Bradford is the first hospital in Yorkshire to adopt it.
Mr White continued:
The AI system has numerous safeguards as it uses a risk-based logic, designed to ensure that lesions are routed to the safest, most appropriate management outcome. The new scheme’s first year data will also be evaluated independently for evidence of its success and safety.
According to Skin Analytics Performance Reports, DERM is already known to be 99.7% accurate at ruling out skin cancers meaning diagnosis is quicker and avoids long waits so safer for patients.
This AI technology has been recommended by the National Institute for Clinical Excellence (NICE) for use in the NHS for the next three years, while further evidence is collected. The current evidence suggests that DERM may be able to identify a cancer lesion with similar accuracy to tele dermatology or face-to-face dermatology assessment.