Starting this autumn, roughly nine million more people in England will have their suspected skin cancer referrals triaged first by an algorithm rather than a dermatologist. NHS England South East announced on October 7, 2026 that it is deploying DERM, an AI tool built by UK company Skin Analytics, across all 18 NHS trusts in the region — a move the NHS describes as its largest deployment of autonomous medical AI to date.
How DERM works
DERM analyzes images of suspicious skin lesions taken during a patient’s referral and decides whether the case needs specialist dermatology evaluation. If it judges a lesion benign, the NHS can discharge the patient at that triage stage without ever booking an outpatient appointment; anyone the system flags as higher risk is fast-tracked to a specialist. The tool holds Class III CE certification for autonomous use — the highest risk classification under European medical device rules — and operates inside NHS clinical safety, security and data governance requirements. According to Skin Analytics, clinical studies and post-market surveillance put its detection rate at roughly 98% of skin cancers, with a 99.5% rate of correctly ruling cancer out.
How we got here
DERM’s path to this scale took years. In May 2025, England’s National Institute for Health and Care Excellence (NICE) gave the tool a conditional recommendation for NHS use, to be rolled out over three years while more real-world evidence is gathered — NICE noted at the time that early data suggested DERM could roughly halve urgent dermatology referrals for suspected skin cancer. Northern Ireland went further, approving DERM for autonomous skin cancer assessment across its entire population of 1.9 million. By the time of the South East announcement, DERM was already running in 25 NHS trusts, had assessed more than 250,000 patients, and had identified over 25,000 cancers subsequently confirmed through histopathology — the tissue-sample gold standard that gives the company’s figures real clinical weight rather than resting on the algorithm’s own say-so.
The people behind the numbers
“DERM has helped more than 250,000 patients find more than 25,000 cancers over six years,” said Neil Daly, founder and chief executive of Skin Analytics. Daghni Rajasingam, regional chief medical officer for NHS England South East, framed the expansion as a way to combine AI with clinical expertise — speeding reassurance for lower-risk patients while freeing specialists to focus on complex cases. The rollout is explicitly tied to the government’s 10-Year Health Plan, which has pushed the NHS toward more diagnostic automation as a way to manage rising demand against a flat dermatology workforce.
The case for caution
Not everyone treats the rollout as an unambiguous win. The headline accuracy figures — 98% detection, 99.5% rule-out — are company-reported, drawn from Skin Analytics’ own clinical studies and post-market surveillance rather than a fully independent trial. That is part of why NICE structured its approval as conditional rather than unconditional: the three-year evidence-gathering period reflects regulators’ own acknowledgment that autonomous discharge decisions in cancer care need continued real-world scrutiny, not a one-time sign-off. A published commentary examining post-deployment performance of an earlier DERM algorithm version found zero skin cancers among patients the system judged eligible for discharge in a review of more than 1,700 cases — a reassuring result, but one tied to a different software version than the one now scaling across the South East, underscoring that performance can shift as models are updated and deployed more broadly.
What the expansion is meant to solve
The pressure driving this is concrete: suspected skin cancer referrals in England have climbed for years, straining a dermatology workforce that hasn’t grown at the same pace. NHS modelling for the three-year South East deployment projects that up to a third of urgent suspected skin cancer referrals could be discharged before a patient’s first outpatient consultation, freeing an estimated 13,500 consultant hours a year — the equivalent of roughly 27,000 additional specialist appointments that could instead go to patients who need them.
What’s next
If the South East expansion performs as modelled, it is likely to become the template other NHS regions use to justify similar rollouts, further extending autonomous discharge decisions into frontline cancer care. The open question is whether the conditional evidence-gathering NICE required will surface problems serious enough to slow that momentum, or whether DERM’s six-year track record of catching more than 25,000 confirmed cancers will be enough to settle the remaining skepticism before the three-year review period even concludes.
Photo: Leolo212 / PIXABAY via Pixabay
