Periodic Reporting for period 1 - iCARE4CVD (INDIVIDUALISED CARE FROM EARLY RISK OF CARDIOVASCULAR DISEASE TO ESTABLISHED HEART FAILURE)
Okres sprawozdawczy: 2023-10-01 do 2024-09-30
iCARE4CVD seeks to transform cardiovascular disease (CVD) management by targeting three major aims: 1) improving early and precise diagnosis as well as classification of patient subgroups; 2) advancing risk prediction models to better define the priority of interventions; and 3) implementing a precision medicine approach that targets treatments to individual patient needs, improving the balance between benefits and side-effects.
To achieve these goals, iCARE4CVD employs advanced AI models, federated databases, and interoperable data platforms, fostering a comprehensive and integrated framework for CVD care. The project empowers patients as active participants in medical decision-making. Through collaboration with public and private stakeholders, iCARE4CVD aims to deliver cost-effective solutions, align with European standards, and implement real-world pilot projects, as new benchmark for sustainable and equitable CVD care.
The overall objectives are:
1) To collect and describe large patient cohorts covering all CVD stages and healthy subjects
2) To create a federated database for sustainable collection and storage of data
3) To identify known and novel biomarkers
4) To ensure stakeholder engagement, especially patients, identify relevant PROMs and PREMs, and understand preferences and motivation
5) To improve individualized diagnostics, risk and therapy models using AI
6) To prospectively validate novel prediction models
7) To establish cost-effective, novel patient-oriented and stratified care pathways
8) To ensure implementation via a strong legal and policy framework
To achieve these goals, iCARE4CVD employs advanced AI models, federated databases, and interoperable data platforms, fostering a comprehensive and integrated framework for CVD care. The project empowers patients as active participants in medical decision-making. Through collaboration with public and private stakeholders, iCARE4CVD aims to deliver cost-effective solutions, align with European standards, and implement real-world pilot projects, as new benchmark for sustainable and equitable CVD care.
The overall objectives are:
1) To collect and describe large patient cohorts covering all CVD stages and healthy subjects
2) To create a federated database for sustainable collection and storage of data
3) To identify known and novel biomarkers
4) To ensure stakeholder engagement, especially patients, identify relevant PROMs and PREMs, and understand preferences and motivation
5) To improve individualized diagnostics, risk and therapy models using AI
6) To prospectively validate novel prediction models
7) To establish cost-effective, novel patient-oriented and stratified care pathways
8) To ensure implementation via a strong legal and policy framework
A fully functional federated database has been established, enabling secure and efficient access to harmonised data from several cohorts, ready for analysis. Data contribution and access agreements are in place, fostering collaboration and ensuring compliance with ethical and legal standards. AI-powered tools have been developed to support data mapping, enhancing the integration of additional cohorts into the project. Systematic reviews have advanced, with research questions defined and a large language model implemented to streamline the review process. Furthermore, initial AI-based analyses have been completed, serving as use cases to demonstrate the potential of data-driven approaches. These achievements lay a robust foundation for leveraging AI and big data to improve diagnosis, risk prediction, and personalised treatment in CVD care.
In RP1, iCARE4CVD has achieved 2 significant results that have a high impact on both the progress of the project and the key stakeholders of the project as well.
1. Establishment of the federated database infrastructure: A comprehensive and secure federated database was developed to enable AI-based analyses across multiple cohorts at the individual patient level, without the need for central data pooling. Key functionalities include:
o A cohort explorer that allows researchers to screen the content of participating cohorts and assess feasibility for joint analyses.
o An online platform that supports the remote upload and execution of code (e.g. Python scripts) for distributed AI-driven analyses, enabling cross-cohort research without compromising data privacy.
o Implementation of a sophisticated data protection framework, including an encrypted dataspace and access controls that comply with GDPR and other relevant data protection standards.
Impact: This infrastructure addresses key barriers in cardiovascular research by unlocking access to high-quality, diverse data sources across Europe. It fosters large-scale, collaborative, and reproducible AI research in CVD, laying the groundwork for personalised treatment strategies.
2. Development of an AI model to predict treatment response in heart failure: A first-generation predictive model was developed using federated data to estimate individual treatment responses to evidence-based heart failure therapies. This includes predictions on:
o The benefit of specific guideline-recommended drugs, and
o The adjustment of diuretic therapy based on individual patient profiles.
Impact: This model represents an important step towards clinical decision support tools that facilitate personalised therapy. It will be further developed into software that supports the implementation of the PREDICT-CVD trial in WP6, aiming to test the clinical utility of such tools in real-world settings.
1. Establishment of the federated database infrastructure: A comprehensive and secure federated database was developed to enable AI-based analyses across multiple cohorts at the individual patient level, without the need for central data pooling. Key functionalities include:
o A cohort explorer that allows researchers to screen the content of participating cohorts and assess feasibility for joint analyses.
o An online platform that supports the remote upload and execution of code (e.g. Python scripts) for distributed AI-driven analyses, enabling cross-cohort research without compromising data privacy.
o Implementation of a sophisticated data protection framework, including an encrypted dataspace and access controls that comply with GDPR and other relevant data protection standards.
Impact: This infrastructure addresses key barriers in cardiovascular research by unlocking access to high-quality, diverse data sources across Europe. It fosters large-scale, collaborative, and reproducible AI research in CVD, laying the groundwork for personalised treatment strategies.
2. Development of an AI model to predict treatment response in heart failure: A first-generation predictive model was developed using federated data to estimate individual treatment responses to evidence-based heart failure therapies. This includes predictions on:
o The benefit of specific guideline-recommended drugs, and
o The adjustment of diuretic therapy based on individual patient profiles.
Impact: This model represents an important step towards clinical decision support tools that facilitate personalised therapy. It will be further developed into software that supports the implementation of the PREDICT-CVD trial in WP6, aiming to test the clinical utility of such tools in real-world settings.