Machine versus man in skin cancer diagnosis

In certain high-need settings, such as developing countries or healthcare systems with limited human resources, AI may bring additional value, particularly as a screening tool. Furthermore, AI could also be useful for developing clinical support tools for not-experienced physicians and general practitioners, in view of an increasing demand for points of care for dermatology screening (9). However, in a not-distant future, automated classifiers and AI algorithms, not independently but in a context of “human-machine collaboration”, will be integrated into clinical dermatology practice for a more accurate and effective triage of lesions. While there is little doubt that AI and ML algorithms for skin cancer diagnosis are already gaining a central role in dermatology research, it is possible to anticipate that their application to clinical practice will require severe and robust validation in large prospective studies.