The translation of in-house imaging AI research into a medical device ensuring ethical and regulatory integrity

Abstract

This manuscript delineates the pathway from in-house research on Artificial Intelligence (AI) to the development of a medical device, addressing critical phases including conceptualization, development, validation, and regulatory compliance. Key stages in the transformation process involve identifying clinical needs, data management, model training, and rigorous validation to ensure AI models are both robust and clinically relevant. Continuous post-deployment surveillance is essential to maintain performance and adapt to changes in clinical practice. The regulatory landscape is complex, encompassing stringent certification processes under the EU Medical Device Regulation (MDR) and the upcoming EU AI Act, which imposes additional compliance requirements aimed at mitigating AI-specific risks. Ethical considerations such as, emphasizing transparency, patient privacy, and equitable access to AI technologies, are paramount. The manuscript underscores the importance of interdisciplinary collaboration, between healthcare institutions and industry partners, and navigation of commercialization and market entry of AI devices. This overview provides a strategic framework for radiologists and healthcare leaders to effectively integrate AI into clinical practice, while adhering to regulatory and ethical standards, ultimately enhancing patient care and operational efficiency.

Keywords: AI Act; Artificial Intelligence; Ethics; Healthcare; Radiology.