Artificial Intelligence (AI) and Smartphone apps to aid melanoma diagnosis 

Melanoma Focus are asked from time to time if there are any smartphone apps we would recommend to aid in the diagnosis of melanoma. This is a very complex and fast-moving area and uses AI. 

Artificial intelligence (AI) is software that can use large amounts of data to assess and make predictions. It is intelligentbecause it works out patterns in the data and tests them, rather than just identifying what it is instructed to find. With AI development happening so rapidly, and healthcare providers using AI more and more, it is vital that more people know the important questions to ask about how reliable these applications are – the quality of the data they are based on, and whether we can depend on them to be safe. 

Example of AI in use

There are many smartphone apps currently available which use AI and are intended to help people to identify suspicious moles and help them to decide when to self-refer to a GP. But which app is best?

We have approached the British Association of Dermatologists for advice as they have a working group currently looking into this area. In summary, the evidence supporting these apps is weak and they do not recommend any specific app for use. Their position is that while AI is a rapidly advancing area which holds great promise for the future, the current evidence for effectiveness and safety of apps for diagnosing melanoma is limited. The studies that have been carried out to validate these apps have often been done in artificial conditions (e.g. using images only without any clinical information from a patient) or with limited numbers of people and therefore may not reflect real-life clinical situations.  This means there is a clinical and safety concern that these apps may not pick up some melanomas or will incorrectly raise suspicions about completely harmless moles. There is no app available at the moment that is equally as good at detecting a melanoma as it is for ruling out melanoma in all individuals.

If you are interested in further information, here is the link to the statement from BAD.

One of the issues to consider when choosing an app is what the evidence is that underpins the app – and this can sometimes be difficult to find. At the moment the current regulatory systems are being tightened up as they are not as robust as they could be – this means that sometimes apps will claim to have regulatory approvals for a purpose e.g. have a CE mark but this does not necessarily guarantee that there is sufficient evidence to prove that it is a reliable tool to make a melanoma diagnosis.

Things for you to particularly consider:

  • What does the app claim to do and is it realistic?
  • Is it going to help solve your problem? Many apps will say in the small print that they are not claiming to make a diagnosis but if you want an app that will make a diagnosis, it is unlikely that this will help you.  
  • How the data which was used to train the AI app was collected?
  • Does the data represent the patients for whom the AI is being used?
  • Has the app been properly tested in the real world?
  • Does using the app involve you giving your images or personal data and if so does the app clearly explain how your images and/or personal data will be used?

Apps that are used to monitor your moles and/or skin lesions at intervals over time, by simply storing your photographs for you to view any changes over time do not necessarily use AI and are not part of this review, but can be a very useful resource.

It is important to point out that smartphone apps are different to teledermatology.  Teledermatology is where a person or a GP may take a picture of a concerned mole or lesion to send to the NHS for a swift consultant dermatologist review to decide which people they should see face-to-face for further examination. This area of practice is rapidly increasing and is very much supported by the NHS to speed up potential diagnoses of melanomas. 

If you have any concerns or would like to discuss anything, please contact our:

Melanoma helpline