Clinical AI – the use of AI solutions in diagnostics or clinical setting

Explore how clinical AI is revolutionizing diagnostics and healthcare. Join us to discover next-gen AI technologies and the human perspectives essential for transforming patient care.

The Clinical AI track will explore AI integration in healthcare, featuring insights for experts, clinicians, and companies.

Sessions:

1. Trustworthy Machine Learning: Philippe Rostalski on AI model reliability in diagnostics.
2. AI in CTG Monitoring: Harshit Mahapatra on AI detecting fetal distress.
3. AI Regulation: Erik Stenberg on the AI Act’s impact.
4. Implemented Case: Human Bytes on real-world AI applications.
5. Change Management: Thomas Karopka on AI in diagnostics.
6. Debate: Moderated by Frederik Mølgaard Thayssen on AI safety, transparency, and barriers.

Program

Welcome and introduction
by/ Carina Østervig Byskov, Innovationsklinikken/ Aalborg University Hospital

Carina Østervig Byskov, Innovationsklinikken/ Aalborg University Hospital, will welcome attendees on behalf of CAIDX. The introduction will outline the goals and highlight of the conference track and its anticipated outcomes, also the launch of a new toolbox, in late 2025, aimed at providing tools and support in the process of developing and integrating AI into clinical workflows.

NEXT AI OPPORTUNITIES IN THE CLINIC

1. Trustworthy Machine Learning for Medical Diagnostics
   by/ Phillipe Rostalski, Fraunhofer IMTE.
   This session will be presented by Philipp Rostalski and Jan Graßhoff, Fraunhofer IMTE. You’ll learn how advanced machine learning models tackle challenges in medical diagnostics. We’ll discuss improving model robustness, reliability and transparency through Bayesian machine learning and addressing ethical as well as regulatory requirements. Real-world examples will be drawn from mechanical ventilation, electrocardiography, and ophthalmology to demonstrate the potential impact of these solutions.
2. Detecting Anomalous Patterns in CTGs using AI
   by/ Harshit Mahapatra, Alexandra Institute.
   Harshit Mahapatra, Alexandra Institute, will discuss using AI to detect anomalous patterns in cardiotocographs (CTGs). This session will focus on AI’s role in identifying irregularities that may indicate fetal distress and present the learnings of Alexandra Institute when applying AI techniques to build a decision support system for CTG monitoring.

DEVELOPMENT AND IMPLEMENTATION OF AI IN THE CLINIC

1. How does the AI Act affect the development of new AI diagnostic solutions in the clinic?
   by/ Erik Stenberg, Innovation Skåne
   Erik Stenberg, Innovation Skåne, will in this session explore the impact of new legal frameworks on hospitals as co-developers of AI solutions, focusing on balancing safety and innovation. Learn how the AI Act and the Medical Device Regulation (MDR) interact, shaping hospitals’ roles in advancing AI technologies within healthcare while ensuring compliance and safety.
2. Implemented Case
   by/ Ulrik Therkildsen, Human Bytes
   Ulrik Juul Rokkedal Therkildsen, Human Bytes is a supplier of AI for healthcare. They cherry-pick some of the best solutions from around the world and bring them to the Nordic countries. In this session Ulrik will show case practical examples of implemented AI solutions within Mammography, Bone fractures, radiation therapy and diabetic eye diseases. We will discuss procurement, implementation and value creation from real clinical settings.
3. There is no AI for this – How to Implement AI in the Clinical Diagnostic Workflow
   by/ Thomas Karopka, Biocon Valley
   Thomas Karopka, Biocon Valley, will lead a session where you’ll explore the practical implementation of AI in clinical diagnostics, focusing on technical, ethical, legal, and regulatory challenges. Learn how AI affects workflows and roles in healthcare. The “Clinical AI-based diagnostics” project CAIDX offers a guide for managing change when integrating AI solutions into clinical settings.

Debate
Moderator: Frederik Mølgaard Thayssen from the Alexandra Institute

Frederik Mølgaard Thayssen from the Alexandra Institute facilitates a debate on integrating AI into clinical practice, focusing on technology, regulation, and human factors.
The discussion covers:

• Balancing innovative AI solutions with safety and compliance requirements from the AI Act and MDR, and exploring successful strategies.
• Identifying barriers to AI integration into clinical workflows and how hospitals can overcome them, with insights from the CAIDX project’s toolbox.
• Ensuring trust and transparency in AI diagnostics, emphasizing the role of Bayesian uncertainty and differential privacy in building trust.