Periodic Reporting for period 1 - CGI-Clinics (Data-driven cancer genome interpretation for personalised cancer treatment)
Okres sprawozdawczy: 2022-11-01 do 2024-04-30
Implementing new Next Generation Sequencing (NGS) technologies in cancer aids informed clinical decisions and targeted therapies. However, accurately interpreting genomic alterations in tumors remains a challenge for broad clinical implementation. While many cancer mutations with relevance in cancer treatment have been annotated in the past decades, the impact of most mutations is still unknown. This underscores the need for advanced systems capable of systematising the interpretation of genomic data to assist and improve cancer diagnostics and treatment. These systems must be self-explainable for clinicians and patient-centric.
CGI-Clinics is a community-driven project that aims to enhance cancer mutation interpretation within European healthcare systems. The project develops a professional Cancer Genome Interpreter (CGI), a bioinformatics tool to systematise tumor genome interpretation for clinical decision-making, enabling doctors to choose effective treatments while ensuring data is safe and interoperable for research. The CGI tool uses data-driven methodologies to interpret known mutations and variants of unknown significance (VUS), and is being adapted as a clinical decision-support tool, while keeping the data safe, structured, and interoperable for its re-used in research.
The tool includes expert meetings for oncologists to discuss CGI-generated mutation interpretation and biomarker reports, and it is patient-centric, promoting knowledge exchange and facilitating treatment management. The consortium, comprising 17 European organizations, tackles the challenge of interpreting cancer mutations and emphasizes the value of data sharing for future cancer research.
CGI-Clinics is a community-driven project that aims to enhance cancer mutation interpretation within European healthcare systems. The project develops a professional Cancer Genome Interpreter (CGI), a bioinformatics tool to systematise tumor genome interpretation for clinical decision-making, enabling doctors to choose effective treatments while ensuring data is safe and interoperable for research. The CGI tool uses data-driven methodologies to interpret known mutations and variants of unknown significance (VUS), and is being adapted as a clinical decision-support tool, while keeping the data safe, structured, and interoperable for its re-used in research.
The tool includes expert meetings for oncologists to discuss CGI-generated mutation interpretation and biomarker reports, and it is patient-centric, promoting knowledge exchange and facilitating treatment management. The consortium, comprising 17 European organizations, tackles the challenge of interpreting cancer mutations and emphasizes the value of data sharing for future cancer research.
During the first 18 months of CGI-Clinics, a preliminary version of the CGI tool was deployed in 10 hospitals, with clinical partners providing feedback for its improvement. This feedback guided the development of a more professional tool, comprising the CGI-Platform and CGI-Pipeline. Additionally, both internal and external CGI databases were updated, including the CGI biomarkers database with new drug treatments in solid tumors and hematological malignancies. The CGI’s Automatic Learning Platform is being improved with enhancements in data-driven methods that will be included in the new CGI.
The consortium mapped the data to the OSIRIS data model, aiming to standardize data and facilitate interoperability with relevant EU infrastructures and promoting data sharing. The new tool is being developed following a regulatory roadmap informed by a market access study, with plans for clinical validation as a decision-support tool for Cancer Genome Interpretation.
We have designed a virtual molecular tumor board (vMTB) to facilitate discussion among experts and its implementation roadmap. The expert database is now actively inscribing experts. The vMTB will facilitate multidisciplinary decision-making and expert exchange in oncology.
For the patient engagement, we conducted interviews to understand patients' experiences, needs, and expectations regarding biomarker testing and data sharing in cancer genomics. This effort is complemented by the development of two educational videos aimed at empowering both patients and clinicians.
The technical advancements achieved during the first 18 months pave the way for significant impacts in clinical implementation and cancer research by the end of the project, while maintaining a strong focus on the patient's perspective in all developments and goals.
The consortium mapped the data to the OSIRIS data model, aiming to standardize data and facilitate interoperability with relevant EU infrastructures and promoting data sharing. The new tool is being developed following a regulatory roadmap informed by a market access study, with plans for clinical validation as a decision-support tool for Cancer Genome Interpretation.
We have designed a virtual molecular tumor board (vMTB) to facilitate discussion among experts and its implementation roadmap. The expert database is now actively inscribing experts. The vMTB will facilitate multidisciplinary decision-making and expert exchange in oncology.
For the patient engagement, we conducted interviews to understand patients' experiences, needs, and expectations regarding biomarker testing and data sharing in cancer genomics. This effort is complemented by the development of two educational videos aimed at empowering both patients and clinicians.
The technical advancements achieved during the first 18 months pave the way for significant impacts in clinical implementation and cancer research by the end of the project, while maintaining a strong focus on the patient's perspective in all developments and goals.
Sequencing genomes is now affordable in clinics and used in oncology to match therapies with a tumor's molecular profile. To fully realise the potential of genomic data in cancer precision medicine, automation and standardisation of interpretation using data-driven decision support tools are crucial. In this context, CGI-Clinics is making significant progress in employing computational and machine learning methods within the CGI tool to classify tumor mutations, including variants of uncertain significance. CGI has the potential to drive new clinical trials and biomarker-driven therapies for the treatment of cancers. Furthermore, the project is building new features into the CGI tool which will promote healthcare community engagement by sharing expertise and patient outcomes. Overall, the project expects to provide the European research community with a new, safe and effective way to reuse clinical and genomic data for future research.
At this stage of the project, the following key needs have been identified to ensure further uptake and success:
- A favourable legal framework that fosters data sharing for research and for public interest, including a detailed regulation for cancer genomic data sharing (the transition of GDPR to EHDS may impact the implementation of CGI).
- Strategies to effectively navigate potential barriers in new IVDR regulation that hinders early development and clinical implementation for data-based medical devices.
- Early adoption and integration of CGI-Clinics by EU-based stakeholders within emerging key infrastructures (uncan.eu EOSC, Cancer Data Hubs) to ensure sustainability within the evolving EU strategy on health data.
- Deployment of CGI-Clinics as an ideal use case to test the upcoming AI Act regulation, given its foundation in explainable AI algorithms that benefit from cancer genomic data.
At this stage of the project, the following key needs have been identified to ensure further uptake and success:
- A favourable legal framework that fosters data sharing for research and for public interest, including a detailed regulation for cancer genomic data sharing (the transition of GDPR to EHDS may impact the implementation of CGI).
- Strategies to effectively navigate potential barriers in new IVDR regulation that hinders early development and clinical implementation for data-based medical devices.
- Early adoption and integration of CGI-Clinics by EU-based stakeholders within emerging key infrastructures (uncan.eu EOSC, Cancer Data Hubs) to ensure sustainability within the evolving EU strategy on health data.
- Deployment of CGI-Clinics as an ideal use case to test the upcoming AI Act regulation, given its foundation in explainable AI algorithms that benefit from cancer genomic data.