As we approach the end of the year, the COMFORT consortium is proud to reflect on the significant progress achieved during the third year of the project. Over the past months, our partners have continued to push the boundaries of multimodal artificial intelligence in healthcare, while keeping a strong focus on trust, transparency, and patient perspectives.

Advancing Multimodal AI in Medical Imaging

One of the key scientific milestones this year was the release of MRSegmentator, our open-source tool for multi-structure MRI segmentation. Designed to support a wide range of clinical and research applications, MRSegmentator has already been adopted by researchers and institutions worldwide. Its growing user community highlights the value of open, reproducible tools in accelerating innovation across the medical imaging field.

In parallel, the consortium made substantial progress in automating the structuring of radiology reports using large language models. These efforts aim to reduce the administrative burden on clinicians, improve data consistency, and enable downstream AI applications. We also advanced cross-modality transfer learning techniques, helping to minimize the need for extensive manual annotations and making AI development more scalable across different imaging modalities.

Understanding Patients’ Views on AI

A major highlight of 2025 was the successful completion of our large-scale multinational patient survey. With responses from 13,806 patients across 43 countries, this study represents one of the most comprehensive efforts to date to understand patient attitudes, expectations, and concerns regarding the use of AI in healthcare.

These insights provide a crucial foundation for COMFORT’s work. By integrating patient perspectives early and systematically, we aim to ensure that future AI systems are not only technically robust, but also aligned with societal values, trusted by users, and accepted by the people they are ultimately designed to support.

Strengthening Trustworthy and Fair AI

Trustworthiness remains a central pillar of the COMFORT project. Throughout the year, our teams advanced research on fairness, robustness, and transparency in multimodal AI systems. This work directly informs the development of the COMFORT trustworthiness model, which seeks to provide practical guidance for building AI solutions that are reliable, ethical, and resilient in real-world clinical settings.

By combining technical innovation with ethical reflection and stakeholder engagement, COMFORT continues to contribute to a more responsible adoption of AI in healthcare.

As the year comes to a close, we would like to thank all consortium partners, collaborators, and participants for their commitment and contributions. We wish everyone happy holidays and look forward to an equally productive and impactful year ahead.