BSC contributes to developing an AI tool to enhance cancer diagnosis and treatment

02 Dec 2024 | Network Updates | Update from Barcelona Supercomputing Center
These updates are republished press releases and communications from members of the Science|Business Network

Researchers at the Barcelona Supercomputing Center - Centro Nacional de Supercomputación (BSC-CNS) have played a key role in the EU-funded INCISIVE project, which has enabled the creation of AI-based decision support services and a federated data repository with images, as well as associated clinical data, to facilitate the diagnosis of lung, colorectal, breast and prostate cancer.

The BSC team has overseen the development of the core components of the INCISIVE platform, a prototype AI toolbox with 28 AI models that help healthcare providers better interpret AI-generated recommendations.

In addition, INCISIVE's federated data repository provides secure, RGDP-compliant access to more than 5.5 million anonymised cancer images across 14 clinical centres, enabling secure data sharing and model training across different locations.

The project, which was completed this year and involved 26 partners from nine European countries, has been commended by the European Commission for its contributions to cancer diagnosis and treatment through the development of innovative tools and data sharing systems. In addition to BSC, the Spanish representation is completed by the Hospital Clínic Barcelona-IDIBAPS, the Fundación TIC Salut Social and Medtronic Iberica.

BSC's main contributions include:

- Integration of the AI-based toolbox: The BSC has developed components that allow the AI models developed to operate as services in a robust and distributed infrastructure (such as data centres). This is essential given the resource-intensive requirements of these AI models, which rely on substantial computational power to operate effectively.

- Federated learning for data privacy: The BSC also implemented federated learning capabilities, allowing AI models to be trained on data distributed across different locations without centralising sensitive patient information, an approach that enables models to learn from multiple data sources while ensuring data privacy and aligning with GDPR standards.

The results of the INCISIVE project will be further extended as partners work to support interoperability at all levels. Many of the INCISIVE partners, including the BSC, are working towards this goal in the framework of initiatives such as Cancer Image Europe (EUCAIM). Further testing and validation of the tools developed in the project will seek to bring them closer to clinical practice.

This article was first published on 22 November by BSC.

Never miss an update from Science|Business:   Newsletter sign-up