At the end of the course: - The student will learn about how AI is embedded in different types of medical devices, spanning from genomics applications to diagnostic imaging. - The student will understand about the process of developing, deploying, and maintaining (MLops) Artificial intelligence-based medical devices in different areas, such as bioinformatics or biomedical images and signal analysis. - The student will understand the regulatory requirements needed to put these systems on the market as medical devices. - The student will be able to create a prototype of a medical system embedding AI, she/he will be able to analyze the steps needed to obtain the certification of their software as a medical device, and define the possible steps for the AI-based software maintenance.
Course Prerequisites
Knowledge about the Artificial Intelligence and Machine Learning taxonomy (such as the definition of supervised, unsupervised and reinforcement learning)
Teaching Methods
The course combines both lectures and laboratory activities. Lectures will be supported by PowerPoint presentations, which will be made available to students on the KIRO platform. These materials will cover topics such as the principles of MLOps, the regulation of medical devices, and several case studies of devices already on the market or under research. During the laboratory sessions, students – possibly organized into groups – will work together to develop a prototype of an AI-based medical device. Each group will be free to choose the application domain of their simulated device (e.g., bioinformatics, medical signal analysis, or imaging) and will define the steps required to obtain device certification. Students will then implement the AI model underlying the device, possibly using public datasets and following the main phases of AI model development (in particular, training and testing). They may also design a graphical interface for interacting with the model.
Assessment Methods
The exam consists of an oral presentation in which students will present their project based on the implementation of an AI-based medical device. The evaluation will take into account adherence to MLOps principles as well as regulatory aspects. Students may work in groups and develop the project collaboratively.
Texts
Powerpoint presentations
Contents
It is shown which are the medical devices that include AI components currently on the market, and what are the lines of research in this area. Subsequently, the regulation necessary to place a medical device on the market is illustrated, with particular attention to the requirements that AI must meet in this area (MDR/IVDR and AI act). Finally, the principles of AI systems development, deployment and maintenance (MLOps) in production in an effective and reliable way are illustrated.
