AI, Sustainability & Site Training Challenges 

Was there an overall message or theme that stood out from this year’s European Congress of Radiology (ECR) 2025? 

ECR 2025 highlighted the integration of AI technologies into medical imaging processes. The goal is to enhance diagnostic accuracy and improve patient outcomes. There was also an emphasis on sustainability in healthcare, highlighting the importance of responsible resource management in radiology. 

Which presentations were you particularly excited to see at this year’s event?  

The advancements in AI applications in radiology, especially with the CHAIMELEON and EuCanImage projects, were exciting because they aimed to create structured repositories for health imaging data, supporting cancer management and precision oncology.^i,ii  

A presentation on the PUMA system showed the use of multiplexing to combine multiple tracer images into one colored image.ⁱⁱⁱ This method shows the presence of disease and provides surrounding physiological context, helping clinicians better understand tumors and their environments. 

GE HealthCare’s Freelium magnetic resonance imaging (MRI) system, a helium-free sealed magnet platform, aims to reduce the need for liquid helium (a limited, non-renewable resource required to run an MRI scan) without sacrificing power efficiency, operational security, or clinical performance.^iv  

What were some of the key trends from the event? 

While much of the focus was on advances in imaging processes and the broader use of AI in clinical settings, presentations on optimizing imaging techniques and integrating advanced technologies in pediatric imaging stressed the importance of minimizing radiation exposure without sacrificing diagnostic quality. One notable presentation on iodine-125 seeds for use during pregnancy in breast cancer patients indicated that fetal radiation doses are comparable to conventional X-rays, alleviating concerns about their use.^v  

The potential for AI applications to transform imaging optimization and enhancement will allow for automatic segmentation in  oncologic imaging. Although there are challenges that lie ahead, such as data harmonization, regulatory navigation, and ethical use, discussions around human oversight and potential solutions are crucial for the successful implementation of AI in clinical settings. 

How do you see these trends impacting the clinical research landscape? 

The renewed focus on AI and data integration has already shown promising advancements in personalized cancer management, creating more accurate and faster diagnoses, better patient outcomes, and a streamlined research process in cancer management.  

Many of the ongoing trials, such as the MASAI trial for interval cancer detection and the DENSE-2 trial for breast cancer screening, reflect this shift towards more tailored and personalized treatment strategies, that continue to provide increasingly effective treatment programs.^vi,vii

What did you take away from the conference to help you in your role supporting PSI’s sponsors? 

Engaging with the innovative attendees offered a glimpse into cutting-edge technologies that allow us to demonstrate to prospective clients that PSI is well-versed in the applications of AI in radiology, including its advantages and limitations. By doing so, we can help mitigate risks and prevent failures in clinical trials. 

Through an in-depth understanding of the critical role of AI in diagnostics and treatment planning, we can better identify potential partnerships and collaborations that align with our strategic goals. Furthermore, the discussions surrounding regulatory challenges and the ethical use of AI will empower us to assist our sponsors in integrating these technologies responsibly.