The National Institute for Health and Care Excellence (NICE) recently published its HTE24 guidance under the Health Technology Evaluation (HTE) programme, marking a milestone for artificial intelligence in dermatology. At the centre of this evaluation is DERM, an AI-enabled diagnostic tool developed by Skin Analytics. While DERM received a conditional green light for NHS use, Moleanalyzer pro, a rival technology, was excluded. Why?
To understand this, we need to unpack how NICE evaluates technologies, particularly under the Late Stage Assessment (LSA) model, and what “recommended with evidence generation” really means in scientific and commissioning terms.
HTE is NICE’s structured route for evaluating medical technologies, with a growing focus on digital health, AI and real-world applications.
The programme does not provide immediate reimbursement decisions, but it does enable adoption within the NHS, especially where a technology addresses a clinical or operational gap. It also provides robust scientific scrutiny even when a technology lacks the randomised controlled trial (RCT) evidence traditionally required for full guidance. In addition, the programme also helps the NHS and commissioners make procurement and implementation decisions during a controlled evidence generation period.
Late Stage Assessment (LSA)
Introduced in 2023, the LSA approach focuses on technologies already in widespread or established use in the NHS, especially those evolving rapidly (like AI). The LSA aims to:
- Level the playing field by comparing incremental technological innovations (e.g., newer AI algorithms vs. older decision tools).
- Ensure NHS procurement decisions are based on evidence of clinical and economic value, not just vendor promises.
- Address the historic issue where technological upgrades are adopted unevenly across the NHS without critical assessment of their marginal benefit.
Why Was DERM Recommended (with Conditions)?
DERM uses a deep ensemble AI model, a technique where multiple neural networks classify the same lesion and vote on the result, to improve accuracy and reduce error. It supports clinicians within tele-dermatology services and is trained on tens of thousands of dermoscopic and clinical images. NICE found several reasons to support its cautious use:
- Fit-for-purpose use case: DERM is intended to triage patients referred under the urgent suspected skin cancer (2-week wait) pathway -a high-priority bottleneck in NHS dermatology.
- Existing real-world use: Already deployed in NHS trusts, DERM has demonstrated practical workflow integration and clinician acceptance.
- Evidence Generation Commitment: Skin Analytics proposed and committed to a post-implementation evidence plan, including concordance with dermatologists, lesion classification accuracy and impact on referrals.
Importantly, this is not reimbursement approval. Rather, it is a “recommendation for use with evidence generation”, meaning:
“DERM can be used within NHS care, but data on safety, accuracy and impact must continue to be collected and analysed before full commissioning decisions are made.” (NICE HTE24, 2025)
This mirrors the conditional coverage with evidence development frameworks seen internationally.
NICE’s recommendation for evidence generation is not vague. It expects structured, real-world evaluations, often designed in line with international frameworks such as:
- DECIDE-AI: Reporting standards for early-stage clinical evaluation of AI-based decision support.
- STARD-AI: Standards for reporting diagnostic accuracy studies of AI-based systems.
- Prospective observational or concordance studies comparing AI outputs with dermatologists and biopsy-confirmed diagnoses.
These studies should also assess downstream effects on referral patterns, workload redistribution and service accessibility.
Technologies recommended with evidence generation typically undergo 12–24 months of structured deployment, with data captured in collaboration with NHS Trusts, Academic Health Science Networks (AHSNs) or evaluation bodies. If the collected data shows value, both clinical and economic, then NICE may issue a fully positive recommendation and NHS England may adopt the technology via the MedTech Funding Mandate or Innovation Technology Payment (ITP) mechanisms. Alternatively, local commissioners may negotiate broader roll-out independently.
Why Was Moleanalyzer pro Excluded?
Moleanalyzer pro (by FotoFinder Systems) is also an AI-powered dermatology tool. But NICE ruled it out of scope for HTE24 due to:
- Misalignment with the clinical pathway: Moleanalyzer pro is designed to support biopsy decisions for pigmented lesions, not to triage referrals under the 2WW skin cancer pathway.
- NICE’s evaluation frameworks are strictly scoped: A technology must match the clinical objective and NHS workflow under review.
- This underscores a critical lesson for digital health innovators: Reimbursement and adoption are tied to clinical alignment, not just scientific accuracy.
DERM vs. Moleanalyzer pro: A Snapshot Comparison
| Feature | DERM (Skin Analytics) | Moleanalyzer pro (FotoFinder) |
| Intended Use | Triage in urgent suspected skin cancer (2WW pathway) | Decision support for biopsy of pigmented lesions |
| NHS Workflow Alignment | Yes | No |
| NICE Assessment Scope | Included in HTE24 | Excluded |
| Deployment in NHS | Yes, several Trusts | Limited or unspecified |
| Evidence Plan | Yes, aligns with NHS diagnostic needs | No formal evidence generation plan submitted |
Currently, DERM is not reimbursed as a standalone tariffed service. However, with HTE24 NHS Trusts may deploy DERM within existing dermatology budgets, particularly if it’s embedded into tele-dermatology or AI pilots. This HTE recommendation acts as a signal to commissioners and ICSs that funding innovation evaluations in dermatology is endorsed by NICE. Successful evidence generation during this period could lead to:
- National commissioning via NHS England, or
- Inclusion in local digital innovation fund programmes, or
- Procurement under the MedTech Funding Mandate, if cost-effectiveness is proven.
While NICE guidance strongly influences NHS policy, actual adoption depends on:
- Integrated Care Systems (ICSs), who fund and coordinate care pathways locally.
- Trust-level innovation leads and dermatology departments, who pilot and evaluate tools.
- AHSNs and NHSE regional teams, who support data collection and wider roll-out.
Unless a national payment mechanism is introduced, tools like DERM must prove their value within constrained local budgets.
It is critical to emphasise that DERM is not a replacement for clinician expertise. NICE requires that AI-based recommendations be reviewed by trained professionals. This mitigates risks such as:
- False negatives (missed melanomas)
- Over-referral (burdening secondary care)
- Algorithmic bias if trained on unrepresentative data
Maintaining clinical oversight is non-negotiable during the evaluation period.
The evaluation criteria included:
| Domain | NICE Requirements |
| Clinical Effectiveness | Sensitivity, specificity, accuracy vs dermatologists |
| Workflow Integration | NHS digital compatibility, clinician adoption |
| Cost Consequences | Resource use, pathway efficiency, referral impact |
| Patient Safety | Missed melanoma risk, over-referral |
| Evidence Plan | Prospective data collection, adherence to standards (e.g., STARD-AI, DECIDE-AI) |
What Innovators Should Learn from HTE24
- Align with pathway needs: Don’t pitch tools for use cases they are not designed for.
- Propose a robust evidence generation plan: NICE increasingly expects structured data collection aligned with international frameworks (e.g., DECIDE-AI, CONSORT-AI).
- Target LSA if your tool is in NHS use: If already piloted or adopted in pockets of the NHS, the LSA route offers a strategic way to secure national credibility.
- “Recommended” ≠ reimbursed: Financial adoption requires local or ICS funding unless a MedTech Funding Mandate follows.
NICE’s HTE24 allows conditional use of AI in skin cancer triage, not reimbursement. Clinical pathway alignment is essential tools that don’t fit are excluded. Evidence generation is now the gateway to NHS-wide adoption and future funding.
In a post-pandemic NHS increasingly shaped by AI and digital triage, NICE’s shift to Late-Stage Assessment and HTEs is a welcome evolution. It offers structure without stifling innovation, and accountability without demanding perfection.
For Skin Analytics, this means a valuable window of opportunity to gather real-world data and secure broader NHS buy-in.
For FotoFinder, and others, it’s a reminder that clinical relevance and pathway alignment matter just as much as algorithm performance.
#SkinAnalytics – Developer of DERM, the first autonomous AI for skin cancer detection approved by the EU (Class III CE mark) and UK (UKCA Class IIa). DERM is integrated into NHS pathways to triage suspected skin cancers.
#DermaSensor – Offers a handheld, FDA-cleared device that uses AI-powered spectroscopy to evaluate skin lesions for potential cancer, providing results in seconds.
#Proscia – Specialises in AI-driven digital pathology solutions, including tools for melanoma detection, enhancing diagnostic accuracy and efficiency.
#MelaFind – Developed a non-invasive device approved by the FDA and CE for assisting dermatologists in melanoma detection through multi-spectral analysis.
#FotoFinder – Provides AI-based tools like Moleanalyzer Pro for the analysis of skin lesions, supporting dermatologists in early skin cancer detection.
#LifeSemantics – Developed the first AI-powered solution approved in South Korea for skin cancer diagnosis, marking a significant advancement in the region’s digital health sector.
#Helfie – An AI start-up offering diagnostic tools for various diseases, including skin conditions, through photo analysis. However, its claims are under scrutiny by the Therapeutic Goods Administration (TGA) due to regulatory concerns.