Project overview
Psoriasis is a common skin disease and affects over 1 million people in the UK. Light therapy is an effective treatment for psoriasis. In this project I want to train artificial intelligence (AI) to improve light therapy for psoriasis. At the start of light therapy, patients are tested to a range of light doses (this is called light testing) so that they won't sunburn during treatment. The results of light testing are assessed by simply looking at the skin but this way of assessing light test results is not accurate. Patients are also asked questions about their skin colour (also called skin tone) and how easily they sunburn and suntan so that the best doses for light testing can be used. However, using these questions to choose the doses for light testing is not reliable. Treatment with light therapy exposes most of the skin to light, which means that unaffected skin may sunburn during treatment. This sunburn is painful and can increase the risk of skin cancer. To avoid this, we use lower doses of light but this means the psoriasis takes many weeks to clear. AI is a computer system that performs tasks usually done by humans and could be used to improve light therapy for psoriasis. I want to train AI using a colour measuring machine (called a spectrophotometer) that measures skin tone and skin redness. This is because the machine is more accurate at reading the light test results, and is better than using questions to choose the best doses for light testing. I also want to train AI to recognise psoriasis as it improves during treatment so that a device that shines light on the psoriasis only can be developed in the future. This will allow us to use higher doses of light to treat psoriasis while avoiding normal skin. This would mean that the psoriasis would improve more quickly. I also want to find out if people with psoriasis would be happy with AI helping to treat their psoriasis. Aims: To develop an AI approach that improves light therapy for people with psoriasis by: 1. Training AI to measure a person's skin tone. 2. Training AI to measure skin redness. 3. Training AI to recognise psoriasis as it improves during light therapy. 4. Checking that people with psoriasis agree with the use of AI in light therapy. Methods: 1. After taking informed consent, I will train AI to measure skin tone using photographs of peoples' skin where I have measured their skin tone with the colour measuring machine. 2. I will also train AI to measure skin redness when patients are being tested for light therapy. This will be done by taking photographs of the skin test sites and measuring the redness with the colour measuring machine. 3. I will train AI to recognise psoriasis as it improves during treatment using photographs of psoriasis taken during treatment. 4. I will also interview people who have psoriasis to understand their views on the idea of using AI for light therapy in the future. Patient and public involvement: Patients have been involved in helping me plan this research. My patient group will be involved throughout this fellowship. They will review the research documents, help develop questions for interviews with psoriasis patients, and share the results of the project with the public and patient charities. Sharing the results: The results will be shared with study participants, patients, members of the public, and the research community through: - Talking to national groups such as the Psoriasis Association. - Presentations at conferences. - Articles in medical journals.