NICE Approves AI Tool to Triage Skin Cancer

NICE Approves AI Tool to Triage Skin Cancer

On 1 May 2025, the National Institute for Health and Care Excellence (NICE) issued a major endorsement: the artificial intelligence tool DERM (Deep Ensemble for Recognition of Malignancy), developed by Skin Analytics, has been conditionally recommended for NHS use over the next three years. Here’s why this matters — and what it means for patients, GPs, and dermatologists.

1. What is DERM?

DERM is an AI-powered skin cancer detection system designed to assist teledermatology pathways. When a patient is referred for suspected skin cancer, clinicians take a magnified image of the lesion using a smartphone coupled with a dermoscopic lens. The image is uploaded to DERM’s online platform, where an AI algorithm analyses it against a database of known skin conditions. Based on the result, patients are either:

• Triaged as benign — discharged with advice and no specialist referral

• Flagged as suspicious/pre‑cancerous — referred swiftly to a dermatologist for further assessment.

NICE’s analysis suggests DERM could slash urgent referrals by up to half, compared with teledermatology alone — addressing a critical bottleneck. The flowchart looks like this:

1. GP refers to urgent pathway

2. Image taken & analysed by DERM

3. Benign cases — fast discharge

4. Suspicious cases — dermatology follow-up

2. Why NICE’s Conditional Approval Is a Big Deal

A digital leap in diagnostic care

In the context of NHS’s "analogue to digital" strategy, this marks a pivotal step. DERM is the first AI dermatology tool to receive CE Class III marking and NICE Early Value Assessment approval.

Patient safety remains paramount

The system achieves a 97% cancer detection rate, with a 99.8% negative predictive value for melanoma, rivaling dermatologist accuracy. The approval, however, is conditional—primarily to collect more data, particularly on patients with darker skin, where initial evidence was limited.

Strategic trial period

NICE has recommended a three‑year “use and learn” phase, during which real-world performance will be closely monitored. Post‑trial, the guidance will be updated england.nhs.uk.

3. The Case for AI Skin Cancer Detection in the NHS

Battle against long waits

Dermatology services routinely handle over 1 million referrals per year, 60% being urgent — but only 6% are confirmed cancer. Meanwhile, many trusts miss the 18‑week referral‑to‑treatment target, with large backlogs. By discharging benign cases early, DERM alleviates pressure, allowing faster specialist attention for those at genuine risk.

Proven real-world impact

Hospitals like Chelsea & Westminster NHS Trust have already implemented similar AI workflows. Images are taken and analysed in seconds; suspicious cases get fast-tracked, while benign lesions are ruled out in minutes. Early results cite a five‑fold speed advantage over traditional in-person diagnosis.

In just over two years, the technology has been deployed at 20 NHS hospitals, resulting in 13 000 detected cases, with nearly half of patients receiving an immediate all-clear. For patients, this means dramatically reduced anxiety and waiting times; for dermatologists, it means laser‑focus on critical cases.

4. Teledermatology + AI: The Perfect Fusion

Teledermatology has been a game-changer — particularly since the pandemic — allowing dermatologists to triage suspect skin lesions via remote image review. But AI brings true scale, speed, and consistency. DERM delivers instant analysis, supporting virtual workflow with:

• Smart triage — distinguishing cancerous vs benign

• Resource optimisation — routing non-cancer cases away from specialist queues

• Support for remote GPs — enabling triage even in peripheral or underserved sites

• Better equity, with built-in second reads for less‑studied groups like darker skin tones.

5. Implementation Roadmap and Challenges Ahead

Phased roll-out

During the NICE trial, most centres are using ‘second read’ methodology: DERM assesses the image first; a human clinician confirms or counters. This approach builds confidence before autonomous use.

Data-driven refinement

Real-world implementation at NHS trusts must feed data to:

• Monitor diagnostic accuracy

• Compare teledermatology with DERM+human review

• Evaluate cost-efficiency and resource savings.

Triumphs and troubles

So far, trust results are promising, but evidence from darker skin types remains limited — thus the ongoing checks . Additionally, robust guidance on automatic consent, ethics, and safety-net advice is already provided via NHS England materials .

6. The Future of AI in Dermatology

DERM is not just a one-off — it’s the forerunner of a major AI transformation in dermatology, with many promising trails:

• Dermatologist-led AI for rare or hard-to-detect variants

• Metadata-powered risk models (like the “C4C risk score”) that combine lesion descriptors and algorithms.

• Deep learning and explainable AI, advancing from detection to explanation and reporting for clinicians.

Innovations are extending to patient-managed apps, GP-led triage support, and AI-assisted decisions in culturally diverse and underserved communities.

7. Key Benefits for Patients and Clinicians

• Faster reassurance for benign lesion cases

• Timely specialist referrals for high-risk lesions

• Reduced diagnostic delays, critical for survival rates in melanoma

• Better resource use, allowing dermatologists to focus on complex cases

• Scalable pathways, reducing regional disparities and reducing pressure on overwhelmed services

• Built-in equity safeguards, ensuring fair outcomes across all demographic groups.

8. Final Thoughts: The Dawn of a New Era

DERM’s NICE-backed conditional approval marks a crucial inflection point in AI skin cancer detection integration within the NHS. It aligns with national priorities to modernise healthcare, digitise services, and enhance cancer outcomes. Over the next three years, data and experience from real-world use will either strengthen this case — or further refine and redefine how, when, and where AI support is best deployed.

As Dr Anastasia Chalkidou (NICE) states, “DERM is an example of how we can harness artificial intelligence to benefit both patients and healthcare professionals” 

Likewise, Ashley Dalton, Minister for Public Health, emphasises speed and disruption in cancer care – an immediate win for patients.

Conclusion

DERM’s conditional approval by NICE signifies a leap forward in AI-driven skin cancer detection in the NHS. Its deployment promises quicker reassurance for non-cancer cases and timely intervention for high-risk lesions — all while easing pressure on dermatology services. As this “use and learn” phase unfolds, RatedDoctor.com will continue tracking developments, sharing best-practice case studies, and equipping GPs and hospitals to embed AI into their clinical workflow responsibly and efficiently.

Stay connected with RatedDoctor.com for expert insights, practical implementation guides, and updates on how AI innovation is reshaping NHS dermatology and beyond.