Primary care clinicians have a vital role to play in detecting and managing patients with skin lesions suspected to be skin cancer, as timely diagnosis and treatment can improve patient outcomes, particularly for melanoma. However, detecting skin cancer can be challenging, as common non-malignant skin lesions share features with less common skin cancers. In recent years there has been huge interest in the application of artificial intelligence (AI) to medical diagnosis in general. One key area is cancer detection and diagnosis, and emerging AI technologies and diagnostic aids may soon become available for use in primary care and community settings to aid the triage of suspicious skin lesions (Esteva et al, 2016; Marchetti et al, 2018).

We aim to address the following questions:
1. Which AI technologies for detecting possible skin cancer may be ready for evaluation prior to implementation in the primary care setting?
2. What will be the costs and benefits of such implementation to patients and to the NHS?The project will have three main parts:
Study 1: a scoping review of the current evidence about AI technologies for detecting possible skin cancer.
Study 2: a discrete choice experiment (DCE) with primary care clinicians, aiming to understand the attributes they consider important in AI technologies for detecting possible skin cancer, and the trade-offs between attributes that they are willing to accept.
Study 3: a health economic analysis, using routinely available data and relevant research findings (Wilson et al, 2013), aiming to establish the cost and accuracy attributes that are necessary for AI technology to be incorporated safely and cost-effectively into primary care to aid the triage of possible skin cancer.