Patient Perspectives On The Use Of Artificial Intelligence In Malignant Melanoma Diagnosis

Introduction: Early diagnosis of malignant melanoma (MM) is challenged by an increased incidence and a shortage of dermatologists. This has raised expectations on the use of artificial intelligence (AI) for diagnosing MM. Implementation of AI-based diagnostic tools rely on diagnostic performance as well as on patient trust and acceptance. Therefore, understanding patients' perspectives is essential when developing trustworthy AI-based diagnostic tools. Objectives: The aim of this study was to investigate perspectives and expectations of AI-based computer-aided diagnosis (AI-CAD) for MM among individuals at high risk of developing MM. Methods: An inductive qualitative study using an interpretive description methodology was carried out. Ten semi-structured interviews incorporating vignettes to investigate expectations regarding trust in AI, physician-patient interaction, diagnostic efficiency, and data security were conducted. Results: Participants preferred AI being used as decision support for physicians. The interaction with a physician (human-in-the-loop) was deemed key to the experience of a proper diagnostic process, and participants recognized AI's potential to streamline the process. Conclusions: Participants trust AI and recognize the potential of AI-CAD in the detection of MM. However, they emphasized the importance of human interaction as essential to a positive diagnostic experience. As trust is fundamental to safe implementation of new technologies in clinical practice, we recommend conducting post-market prospective studies employing a mixed methods approach. This will help ensure both patients’ trust in the technology and physicians’ confidence in its clinical utility.