Periodic Reporting for period 1 - CVDLINK (A federated paradigm of real-world data sources utilization for the empowerment of diagnosis, prognosis and risk assessment of cardiovascular conditions)
Reporting period: 2024-01-01 to 2025-06-30
The development of effective preventive interventions, adopted for individuals and population, will reduce the overall cost of patient management and justify the assertion that CVDs can be prevented and controlled. The exploration of existing clinical, genetic, and real-world data sources can contribute to these goals. However, several challenges related to data availability, management, and processing adds complexity to the scope.
CVDLINK project aims to tackle these challenges by implementing a privacy-by-design European-wide federated platform-as-a-service (PaaS) to deliver effective data-driven human-centric interventions and advance research and management in the CVD domain.
CVDLINK is based on two major offerings:
1) the seamless and legally compliant integration, secure sharing, and automated curation of existing data
2) a set of AI and data-enabled precision medicine tools and pipelines for better diagnosis, risk stratification, and treatment of CVDs.
The project will examine different cardiovascular conditions, making use of retrospective datasets, cohort studies, and biobanks from seven countries, aiming to demonstrate how heterogeneous data sources can be effectively exploited to build comprehensive AI-driven tools to benefit health systems, patients, industry, and EU citizens. The developed tools will be validated in five countries, demonstrating the impact of the CVDLINK outputs.
Additionally, the project will generate a set of best practices which, alongside a cost-effectiveness analysis and awareness-raising campaigns, will promote the adoption of CVDLINK outputs in the mid-term.
CVDLINK project aims to tackle these challenges by implementing a privacy-by-design European-wide federated platform-as-a-service (PaaS) to deliver effective data-driven human-centric interventions and advance research and management in the CVD domain.
CVDLINK is based on two major offerings:
1) the seamless and legally compliant integration, secure sharing, and automated curation of existing data
2) a set of AI and data-enabled precision medicine tools and pipelines for better diagnosis, risk stratification, and treatment of CVDs.
The project will examine different cardiovascular conditions, making use of retrospective datasets, cohort studies, and biobanks from seven countries, aiming to demonstrate how heterogeneous data sources can be effectively exploited to build comprehensive AI-driven tools to benefit health systems, patients, industry, and EU citizens. The developed tools will be validated in five countries, demonstrating the impact of the CVDLINK outputs.
Additionally, the project will generate a set of best practices which, alongside a cost-effectiveness analysis and awareness-raising campaigns, will promote the adoption of CVDLINK outputs in the mid-term.
During the first reporting period, CVDLINK project has made steady progress towards its objectives, with activities advancing in line with the established plan.
CVDLINK project successfully completed a comprehensive gap analysis of cardiovascular disease (CVD) management and defined use case scenarios, legal framework, and a federated data-sharing architecture.
Scientifically, the project advanced the development of multimodal, AI-driven tools for early CVD risk identification by integrating clinical, genomic, and imaging data using an extended OMOP Common Data Model.
Technically, a robust federated infrastructure was established, incorporating privacy-preserving AI workflows, explainable AI models, a user-centric platform and early validation studies and patient recruitment were launched.
Communication and stakeholder engagement were driven through co-design activities, an active webpage and social media presence, and participation in over 10 scientific and industry events.
A dynamic exploitation strategy and sustainability roadmap were initiated, aligning key outputs with commercialization opportunities and EU initiatives like the 1M+ Genomes.
The expected impacts are relevant, and consortium partners are working together on implementing a federated platform-as-a-service (PaaS) for the delivery of effective data-driven interventions and the advancement of research and management in the cardiovascular domain.
These coordinated efforts lay the foundation for CVDLINK to scale federated, AI-powered cardiovascular care across Europe.
CVDLINK project successfully completed a comprehensive gap analysis of cardiovascular disease (CVD) management and defined use case scenarios, legal framework, and a federated data-sharing architecture.
Scientifically, the project advanced the development of multimodal, AI-driven tools for early CVD risk identification by integrating clinical, genomic, and imaging data using an extended OMOP Common Data Model.
Technically, a robust federated infrastructure was established, incorporating privacy-preserving AI workflows, explainable AI models, a user-centric platform and early validation studies and patient recruitment were launched.
Communication and stakeholder engagement were driven through co-design activities, an active webpage and social media presence, and participation in over 10 scientific and industry events.
A dynamic exploitation strategy and sustainability roadmap were initiated, aligning key outputs with commercialization opportunities and EU initiatives like the 1M+ Genomes.
The expected impacts are relevant, and consortium partners are working together on implementing a federated platform-as-a-service (PaaS) for the delivery of effective data-driven interventions and the advancement of research and management in the cardiovascular domain.
These coordinated efforts lay the foundation for CVDLINK to scale federated, AI-powered cardiovascular care across Europe.
CVDLINK has defined several innovations that go significantly beyond the current state of the art in digital health and cardiovascular AI. They are clustered into 5+1 domains that, when integrated, form a novel, multimodal, federated, and user-centric cardiovascular innovation ecosystem:
(1) CARDIAC AI MODELS. The use of multimodal self-supervised learning (ECG + ECHO + Angio) and lightweight architectures for explainable, real-time predictions sets CVDLINK system apart from existing tools that are typically modality-specific and lack interpretability. This enables deployment in clinical environments with constrained computational resources.
(2) CLINICAL DECISION SUPPORT & IMAGING. The integration of AI-driven ejection fraction (EF) prediction, data harmonisation, and imaging workflows within interoperable systems moves beyond current offerings by combining structured data, imaging biomarkers, and UI-enhanced decision-making in one platform.
(3) FEDERATED INFRASTRUCTURE & DATA SHARING. CVDLINK establishes a working, GDPR-compliant federated learning ecosystem, complete with federated storage, search, and a Common Data Model - interoperability based on an extended OMOP surpassing conceptual frameworks like Gaia-X or MedCo in terms of practical usability and integration readiness.
(4) GENOMIC RISK & PREDICTIVE MODELLING. The development of graph-based individualized risk models using polygenic risk scores, omics, and unstructured clinical data represents a novel methodological leap, far beyond the static or siloed risk models currently in use.
(5) USER-CENTRIC INTEGRATION. Advanced XAI methods, semi-automated annotation, and intuitive modular interfaces support not only transparency but clinical engagement, promoting real-world integration and trust.
(+1) ECOSYSTEM GOVERNANCE & CO-CREATION. By embedding sustainability, stakeholder co-creation, and ethical frameworks into the design, CVDLINK ecosystem ensures long-term innovation transferability. This cluster acts as the innovation backbone, enabling continuous feedback, regulatory alignment, and long-term sustainability.
These innovations combine methodological novelty, high clinical relevance, and technical excellence, creating a next-generation ecosystem for cardiovascular disease management.
(1) CARDIAC AI MODELS. The use of multimodal self-supervised learning (ECG + ECHO + Angio) and lightweight architectures for explainable, real-time predictions sets CVDLINK system apart from existing tools that are typically modality-specific and lack interpretability. This enables deployment in clinical environments with constrained computational resources.
(2) CLINICAL DECISION SUPPORT & IMAGING. The integration of AI-driven ejection fraction (EF) prediction, data harmonisation, and imaging workflows within interoperable systems moves beyond current offerings by combining structured data, imaging biomarkers, and UI-enhanced decision-making in one platform.
(3) FEDERATED INFRASTRUCTURE & DATA SHARING. CVDLINK establishes a working, GDPR-compliant federated learning ecosystem, complete with federated storage, search, and a Common Data Model - interoperability based on an extended OMOP surpassing conceptual frameworks like Gaia-X or MedCo in terms of practical usability and integration readiness.
(4) GENOMIC RISK & PREDICTIVE MODELLING. The development of graph-based individualized risk models using polygenic risk scores, omics, and unstructured clinical data represents a novel methodological leap, far beyond the static or siloed risk models currently in use.
(5) USER-CENTRIC INTEGRATION. Advanced XAI methods, semi-automated annotation, and intuitive modular interfaces support not only transparency but clinical engagement, promoting real-world integration and trust.
(+1) ECOSYSTEM GOVERNANCE & CO-CREATION. By embedding sustainability, stakeholder co-creation, and ethical frameworks into the design, CVDLINK ecosystem ensures long-term innovation transferability. This cluster acts as the innovation backbone, enabling continuous feedback, regulatory alignment, and long-term sustainability.
These innovations combine methodological novelty, high clinical relevance, and technical excellence, creating a next-generation ecosystem for cardiovascular disease management.