Generously hosted by consortium partner Aristotle University of Thessaloniki (AUTH), the COMFORT consortium met in Thessaloniki, Greece on 15th and 16th of May for the second progress meeting. These annual meetings are essential checkpoints, providing an opportunity for partners to share updates on the progress of their respective work packages (WPs). With the project now at its halfway point, this gathering held particular significance. Discussions centered on aligning interdependencies between work packages, ensuring that teams can move forward in a coordinated, efficient manner. The consortium achieved consensus on several key issues, laying the groundwork for strong momentum in the months ahead—an intense period marked by critical deliverables and milestones. Special attention was given to challenges surrounding data sharing and the multimodality of the AI models.

Key Exploitable Results

Within the framework of WP 6, an Exploitation workshop was chaired by Dr. Sonja Bergner (EURICE) who presented the outcomes of a recently carried out innovation questionnaire-survey. The goal of the workshop was to discuss and refine the project’s key exploitable results (KER), and to allocate a responsible owner for each. Across all identified main results, clear interdependencies emerge – e.g., data from clinical partners powers the models which need to be validated via the platform and require explainability to be trusted by end-users and patients, all of which must be communicated through education. Four KER’s were defined, and although several partners will collaborate on them, an owner was defined for each of them, who will take on the responsibility to carry them further. The most valuable results of the COMFORT project will be: • Large, annotated datasets for prostate and kidney cancer patients • Trustworthy and clinically validated decision support tools (AI models) for prostate and kidney cancer • Tested ‘End-user platform’ for multi-centric AI-based studies • Publication of educational materials, prospective study results and policy briefs

COMFORT End-User Working Group

The primary end users of the COMFORT decision support tool will be urologists and radiologists. However, patient needs must also be central to its design and functionality. To ensure that both clinical and patient perspectives are meaningfully integrated, an End-User Working Group will now be established. This group will support the consortium throughout the remainder of the project and will include representatives from partners involved in the prospective study, as well as patient advocacy groups.

Workshop on Data Management

During the meeting, our partner Quantib/Deephealth held a workshop on their cloud-based platform developed for prostate cancer diagnostics. The platform features advanced data management capabilities, including customisable access rights and user permissions. Its workflow includes a pre-processing phase with AI model deployment and image registration, followed by interactive steps that can be tailored to the user's needs—for example, allowing radiologists to edit AI-generated prostate segmentations and input PI-RADS scores. The final results are compiled into a PDF report. To ensure clinical reliability, the platform matches its training and test datasets with real patient distributions and conducts subgroup analyses to rule out bias and confirm statistical validity.

Collaborative exchange

In addition to sessions with external advisors, a General Assembly meeting and an Internal Ethics Board meeting, the progress meeting focussed on a collaborative exchange with related projects, showcasing the invaluable benefit of exchange and collaboration between projects and experts in this field.

EUCAIM - EUropean Federation for CAncer Images

EUCAIM aims to create a federated, pan-European research infrastructure to address the fragmentation and limited accessibility of cancer imaging data across Europe. By building on existing data repositories from the AI4HI initiative and integrating major European Research Infrastructures, EUCAIM will enable the development and validation of AI tools to support precision medicine in cancer care. The project emphasises ethical and safe deployment of AI in real-world clinical settings, while ensuring legal compliance and data sovereignty across participating countries.

Find more information on EUCAIM here.

RAISE - Research Analysis Identifier

RAISE aims to establish a distributed, crowdsourced data processing infrastructure that emphasizes transparency and adherence to the FAIR (Findability, Accessibility, Interoperability, and Reusability) principles. A key innovation of RAISE is its mechanism that allows algorithms to be sent to datasets, rather than transferring large datasets to processing sites, thereby enhancing efficiency and data security. The project plans to deliver several key outputs, including the RAI Cloud platform for orchestrating data sharing and processing, a network of RAI Certified nodes offering data storage and processing resources, the Research Analysis Identifier (RAI) for uniquely identifying analysis results without disclosing raw data or source code, services for dataset plagiarism detection and proof-of-origin, and a synthetic data generator. Through these developments, RAISE seeks to streamline data processing workflows and bolster trust in scientific data analysis. Find more information on RAISE here.

PREPARE - Personalised REhabilitation via Novel AI PAtient StRatification StratEgies

PREPARE’s mission is to enhance the lives of people with chronic noncommunicable diseases through personalised and holistic rehabilitation strategies. The project integrates real-world clinical data using advanced machine learning techniques while preserving patient privacy. By combining clinical, socio-behavioural, and public health data—including information on patients' living conditions and demographics—PREPARE aims to develop tools that support optimal therapy decisions. The project’s innovative approach will be demonstrated through pilot cases across nine major rehabilitation-relevant conditions, including Parkinson’s disease, scoliosis, and joint replacements. Find more information on PREPARE here.

iPROLEPSIS

iPROLEPSIS aims to enhance the management of Psoriatic Arthritis (PsA) through a personalised digital care ecosystem. The project employs explainable AI (xAI) models to analyse diverse data sources—including clinical records, wearable sensor data, and behavioural information—to identify key factors influencing the progression from psoriasis to PsA. The project integrates Internet of Things (IoT) technologies and mobile applications to collect real-world data, facilitating early diagnosis and tailored interventions. Clinical studies across five European countries are underway to validate digital biomarkers and assess the effectiveness of the developed tools. Ultimately, iPROLEPSIS seeks to empower patients and healthcare professionals with innovative solutions for proactive PsA management. Find more information on iPROLEPSIS here.