We were excited to sponsor one of our partners to attend RSNA 2025 with us, where Dr. Simran Singh Dhesi, Specialty Radiology Registrar at Leeds Teaching Hospitals NHS Trust (LTHT), presented on the Newton’s Tree FAMOS (Federated AI Monitoring Service) project.

Read their full guest blog below to learn more about his presentation on FAMOS, RSNA experience and LTHT’s collaboration with Newton’s Tree:

Attending RSNA 2025 in Chicago offered an invaluable opportunity to join global discussions on the future of radiology and to showcase collaborative work between Leeds Teaching Hospitals NHS Trust (LTHT), Newton’s Tree, and NHS Greater Glasgow and Clyde. As clinical AI continues to develop at pace, this year’s meeting placed particular emphasis on how health systems can deploy these technologies safely, responsibly, and at scale which resonated well with the work we presented.

Our abstract, “Federated AI Monitoring Service (FAMOS) for Safeguarding CXR AI Deployment: A Sociotechnical Approach at Two UK Tertiary Hospitals,” explored the challenge of ensuring that AI tools remain safe and reliable after deployment. While AI systems are extensively evaluated before clinical use, their behaviour can evolve as patient populations, workflows and imaging conditions change. Identifying these changes early is essential for supporting safe and confident use in practice.

The FAMOS project was created to test whether a federated, sociotechnical approach to monitoring could offer a practical way to track real world AI performance across multiple sites. This early work represents an important proof of concept rather than a definitive validation of effectiveness. By combining technical monitoring of input data and model outputs with clinician AI concordance, the project aimed to understand whether such an approach could feasibly support more transparent oversight of deployed AI tools. The results were encouraging, but the intention was always exploratory, laying the groundwork for what robust and ongoing AI assurance could look like in the near future.

Presenting this work at RSNA generated thoughtful conversations with clinicians, researchers and industry colleagues, many of whom were particularly interested in how federated approaches might support cross-site learning while protecting patient privacy and how governance frameworks can evolve to keep pace with AI deployment. These discussions reinforced the importance of multidisciplinary collaboration which has been central to this project from the outset.

The partnership with Newton’s Tree has been especially meaningful. Bringing together clinical insight from LTHT with Newton’s Tree’s expertise in AI governance and system design has allowed us to develop an approach that is technically sound while remaining grounded in real clinical contexts. RSNA provided a valuable platform to share this early stage work and to learn from others shaping the next steps in safe and responsible AI.

As AI becomes more embedded in radiology, the need for thoughtful and ongoing monitoring will only grow. FAMOS is an early step in this direction, offering insight into what future assurance frameworks might look like and highlighting the potential of collaborative and patient-centred approaches to AI safety.

- Dr. Simran Singh Dhesi, Specialty Radiology Registrar, Leeds Teaching Hospitals NHS Trust

Newton’s Tree offers dissemination bursaries to existing customers to support the presentation of collaborative clinical or research work at conferences. For more information, contact info@newtonstree.com.

‍