Advice
The technology
The technology
The technologies included in this briefing are standalone software platforms that use varying degrees of system processing and adaptive artificial intelligence (AI) to review and inform the detection of possible melanoma. Version updates and periodic maintenance activities are needed for these technologies and can be done remotely. All technologies reported on are for people aged 18 or older. Other similar technologies may be available but are not included in this briefing, for example because of the lack of a current publicly available evidence base.
nomela (Moletest Scotland)
nomela is an advanced image analysis technology for melanoma screening. It is designed for use by healthcare professionals when reviewing suspicious pigmented skin lesions as a 'rule out' test. Hardware needed to use this technology includes an iOS tablet with the nomela software loaded (with all other functions disabled except image capture and analysis) and it can be used offline. However, an internet connection via Wi-Fi or 4G technology is needed to intermittently upload results and reports. The tablet scans a skin lesion using the camera, and the nomela application then identifies the perimeter of the suspicious lesion before analysis. This analysis is made using 5 signal processing algorithms to identify texture, asymmetry, colour, average gradient and edge irregularity. The values of these measurements are compared with reference ranges and if 1 measurement is found to fall outside, then the test provides a binary result within 1 minute, which is shown on screen as either 'no evidence of melanoma' or 'melanoma not excluded' (McKenna et al. 2020).
DERM (Skin Analytics)
DERM (Deep Ensemble for the Recognition of Malignancy) is an AI-based skin lesion analysis device intended for use in the screening, triage and assessment of skin lesions suspected as being skin cancer, including melanoma, in adults in any body location except where specific exclusions apply. DERM can be used within a pre-primary care model in a community diagnostic hub in addition to a primary care setting to assist GPs in referring people with suspected melanoma. It can also be used in secondary care to triage cases on the 2-week wait referral pathway. DERM needs a dermatoscope, a smartphone and connection kit, internet connection to submit images and receive results, and a computer screen to view the results. If a lesion exhibits features of more than 1 lesion type, DERM uses a risk hierarchy in its final analysis. The algorithm was trained on a proportion of historical and prospectively collected images from a UK population. DERM does not update itself automatically but the algorithm needs retraining with additional images. DERM has been granted a phase 4 AI in Health and Care Award by the NHS AI Lab (formerly known as NHSX).
Moleanalyzer pro (FotoFinder Systems)
Moleanalyzer pro is intended to be used by a medical professional for non‑invasive visual documentation of primary skin lesions and aims to help the recognition of melanoma lesions. Moleanalyzer pro is predominantly intended for use by dermatologists in a secondary care setting. Lesions can be between 2 mm to 20 mm and should be intact without additional psoriasis, eczema, acute sunburn or on hair-covered parts of the body (see FotoFinder's information of performance and safety). Moleanalyzer pro needs a video dermatoscope (such as medicam 1000) with FotoFinder universe software. It uses a deep-learning algorithm and complex machine learning to continuously update the algorithm with the goal of increasing sensitivity and specificity.
Lesions are assigned a score from 0 to 1, according to structural characteristics (with 0 indicating no suspicion of melanoma and 1 indicating a high suspicion of melanoma), but a threshold value is not provided to guide excision or biopsy.
SkinVision
SkinVision is a smartphone app intended to provide immediate risk indication for the most common types of skin cancer of a specific lesion on the skin. SkinVision is a patient-facing technology and is designed for people to self‑examine suspicious skin lesions; its place in the pathway is before formal interaction with a healthcare professional in a primary care setting. SkinVision needs a smartphone, internet connection and a subscription, as well as the mHealth app (which SkinVision's algorithm is integrated into). SkinVision's AI algorithm assesses an image of the suspicious skin lesion to provide a recommendation on whether there is a need to visit a healthcare professional for further review. There are 3 risk indications for skin lesions: low risk, low risk (plus symptoms) and high risk. SkinVision's algorithm rates images as high risk when there is some similarity between the image and the skin cancer images in its database. The algorithm is trained and validated on SkinVision's large clinical database of skin lesion images. It quantifies variations in texture, the lesion's borders, the ratio in intensity between the skin and skin lesion and the number of distinct regions with different intensity values (Maier et al. 2015). If the skin lesion is high risk or the person has concerns, they share the high‑quality image and the risk assessment directly with their healthcare professional through the app. The healthcare professional decides after viewing the image and risk assessment whether to book the person for follow‑up treatment or not. SkinVision reports to allow people to self-check and monitor their condition using an interactive body map and reminders in their smartphone app to monitor their skin health over time.
Innovations
Digital technologies, with or without AI capabilities, that analyse skin images for melanoma may help increase diagnostic accuracy and reduce waiting times for onward referral by improving triaging for specialist care from both primary and secondary care.
Current care pathway
Most skin lesions will first be examined in a primary care setting. Experts report variation in practice across the country and significant variability in staff and set up in primary care. NICE's guideline on the recognition and referral of suspected cancer includes a weighted 7-point checklist, where suspicious pigmented skin lesions with a weighted score of 3 or more should be urgently referred under the 2-week rule. GPs may or may not use a dermatoscope (a handheld, specialised magnifying device) to support their assessment of the lesion. Any lesions that cannot be considered definitively or unequivocally non-cancerous should be referred to a skin specialist.
As outlined in the British Association of Dermatologists' UK guidelines (Marsden et al. 2010), when a suspicious skin lesion presents and there is a need to exclude melanoma, the lesion will be examined using a dermatoscope (carried out by healthcare professionals trained in this technique) and, when indicated, a biopsy carried out for histopathological review.
The following publications have been identified as being relevant to this care pathway:
Population, setting and intended user
There are reportedly around 16,700 new melanoma skin cancer cases in the UK every year (see Cancer Research UK's melanoma skin cancer statistics). These technologies are intended to be used in addition to current methods of assessing, detecting and characterising adults with suspected melanoma.
There are different options for using these technologies in the NHS care pathway for melanoma:
-
Pre-primary care. Patient-facing technologies such as SkinVision are positioned to allow people to proactively monitor their skin over time. DERM is currently being piloted in a community diagnostic hub. Both of these technologies provide an immediate risk assessment and indication on whether to visit their GP and with what urgency.
-
Primary care. These technologies can be used as a tool to help GPs make a more informed diagnosis of skin malignancies such as melanomas to improve the appropriateness of onward referrals.
-
Secondary care. These technologies can be used to support remote triage of cases that have already been referred on the 2-week wait referral pathway, thus streamlining workflow and reducing backlog. If the technologies can correctly identify people without malignancies, then theoretically this will reduce the number of unnecessary biopsies. These technologies, particularly DERM, can be used for triaging people in both primary and secondary care settings. They have the ability to reduce inter‑clinician and department variability and improve patient outcomes (sensitivity or negative predictive value) compared with standard care.
Using these technologies in any of these settings has the potential to reduce the high number of people in specialist clinics by reducing the number of inappropriate referrals:
-
nomela's place in the care pathway may be in primary care, a diagnostic centre or a specialist GP community service setting.
-
DERM can be effectively used in the pre-primary (pre-GP) setting on a self‑referred population, in a primary care setting as a decision support for GPs or in a secondary care setting as part of a community hub (when a dermatologist forms part of the care team).
-
Moleanalyzer pro may be used in the secondary care setting.
-
SkinVision is a patient-facing technology, placed as a pre-GP tool. It is a smartphone-only technology, allowing people to proactively monitor their skin for signs of skin cancer and providing indication for referral to the GP when needed.
Costs
Technology costs
nomela
The company has not yet listed any price for nomela, but they plan to supply and maintain the iOS devices for free, together with the use of their infrastructure, which can deliver eConsents and reports securely to NHS systems as needed. They are currently working on 2 provisional pricing models: a per-test fee in the range of £30 to £50 (excluding VAT) or a yearly one-off licensing fee calculated on the estimated annual trust patient volume.
DERM
The price of the technology includes basic hardware, software, service and support. The hardware includes an iPhone, a dermatoscope and a connection kit. This technology does not replace the use of a dermatoscope but is used alongside. DERM can be used either as a standalone web-based service or with a cloud-based platform called Ov2. This cloud-based teledermatology platform allows secure collection of consent, medical and lesion history, and contextual clinical and dermoscopic images of lesions and produces a final report. There are multiple possible options for the image capture hardware kit based on client preference, with prices starting from around £350 (excluding VAT). One kit is shared across several clinical users within the same trust (typically 3 to 5 kits per site). A secondary care model reviewing up to 10,000 people per year would need around 5 to 6 kits. Third-party client software is included within the price of the technology. The company report the current pricing model is either a pay per use or a block contract model and is around £40 per use (excluding VAT). Use is estimated on previous years' activity levels or estimated referral volume.
Moleanalyzer pro
The price of the technology is £53,574 (excluding VAT) for a total body mapping system. The price includes the master tower, medical components, raw imaging kits, dermatoscope components, software and licence. The machine comes with 2 years' free maintenance, which covers both software and hardware. Depending on the system purchased, there is an annual maintenance cost at the end of the 2-year warranty. The lifespan of the technology is around 7 to 8 years. The technology can also be purchased separately from the total body mapping system, in which case a monthly cost of £200 or £1.50 per suspected mole analysed is charged per trust. Any dermatoscope connected to an iOS or android device can be used for this.
Cost of standard care
The current practice is for GPs to refer a suspected melanoma case to a dermatology clinic for further testing. GPs with a particular interest in the disease area may perform dermoscopy. In primary care, GPs do a basic screening, and this may or may not include use of a dermatoscope (ranging in price from £700 to £1,000 excluding VAT, with an expected lifespan of 5 to 10 years).
In secondary care, a specialist uses a dermatoscope for examination and then refers on to biopsy if needed. The cost of screening for melanoma per person in a secondary care setting would be a standard outpatient consultation (dermatology), costing £121 excluding VAT (National Cost Collection 2019/20).
Resource consequences
There is no publicly available information on the resource consequences (economic studies reporting the cost effectiveness) of the technologies. However, Skin Analytics (DERM) states that a study done to assess the resource consequences of deploying the technology (DERM Health Economics Study, 2020/21) was presented at the American Academy of Dermatology conference in 2022 and will be published as a full-text article. If these technologies are adopted in a primary care setting, this could result in a reduction of inappropriate referrals to specialist dermatologists. Resource consequences of adopting the technology will be contingent on use or volume settings, and further data on the effectiveness and cost effectiveness of these technologies across the NHS are needed.
For all technologies, there is a necessary set-up and training period that will require additional resources. For example, users will get training on how to capture good-quality clinical images for optimal performance of the AI algorithm, and there will be IT costs associated with establishing cloud connectivity through the NHS firewall and limited options for connection to external sites. For each local installation, this may take some time to be negotiated because there are currently no national guidelines for using cloud‑based technology solutions in the NHS.
The DERM company website states that the technology is hosted on a cloud‑based service. DERM is intended to be used under the supervision of a healthcare professional as part of a predetermined or agreed clinical pathway only. It is also intended to be used when users responsible for capturing images and interpreting the results have been appropriately trained to perform these tasks. nomela is designed for use solely by trained medical professionals, with a training and certification system that has been used for clinical trials. For Moleanalyzer pro, the installation and training will be done by FotoFinder UK and is included in the price. SkinVision users need a smartphone device with a camera to download and use the SkinVision app.