Interest in artificial intelligence for the diagnosis of non-melanoma skin cancer: a survey among French general practitioners

Background General practitioners (GPs) are playing a key role in skin cancer screening. Non-melanoma skin cancer is frequent and difficult to diagnose. Objectives We aimed to assess whether GPs are facing difficulties in diagnosing non-pigmented skin tumours (NPSTs) and whether they would be interested in artificial intelligence (AI) software that could help them with this task. Materials & Methods A questionnaire addressing the difficulties in diagnosing NPST and the potential interest in AI as a tool to help diagnose these tumours was emailed to GPs working in two French regions. Results In total, 147 respondents (98%) had faced difficulties diagnosing NPSTs; 86% agreed that an AI diagnostic tool could be useful in a GP’s office and 83% agreed that it could change their practice. Nevertheless, 68% would not be willing to pay for this kind of software. Conclusion GPs are facing difficulties in diagnosing NPSTs and would be interested in an AI tool that could help them achieve this. As referral to dermatology practices can be trying, AI diagnostic aids should be developed to help GPs in their gatekeeping task, however at limited cost to the GP.