Gradiant Unveils First Prototype of TRUMPET: A Project Using AI with Privacy Guarantees to Enhance Cancer Patient Care
The Galician technology center presented the first pilot of this European project in Manchester, which applies AI to medical records to define personalized treatments and improve patient care
Gradiant, leading this initiative, is developing the knowledge and software necessary to enable the secure exchange of data between different European hospitals while fully respecting privacy
The technology center Gradiant has presented the European Commission with the first prototype of TRUMPET, a project that leverages Artificial Intelligence (AI) to learn from the most successful treatments and improve care for cancer patients.
During a recent meeting in Manchester, Jaime Loureiro, Head of Advanced Data Security at Gradiant, presented the results of the first phase of TRUMPET (Trustworthy Multi-site Privacy Enhancing Technologies), a three-year project that has demonstrated the tool’s ability to allow European hospitals to establish a collaborative model for securely sharing patient data while fully respecting privacy.
Gradiant’s role, as the leading partner in this project alongside European entities from six countries, is to generate the necessary knowledge and software to ensure data sharing occurs with full privacy guarantees. Applying AI to patient medical records will also enable the definition of personalized treatments. “The project arose from the need to generate AI models from various data sources, with the aim of creating more effective and useful tools that help doctors improve cancer treatments for their patients, while ensuring the privacy and confidentiality of clinical data,” explains Loureiro.
TRUMPET Project
TRUMPET is a European collaborative initiative involving entities from Belgium, Spain, France, Israel, Italy, and the United Kingdom. Nearly 40 highly qualified professionals, including 11 from Gradiant, are working on its development.
The project is part of Horizon Europe under its Civil Security for Society section and has received €5 million in funding.
The proposed concept for TRUMPET is innovative and unique, offering a platform that allows for the analysis of large amounts of clinical data in a protected federated learning environment, ensuring data privacy. TRUMPET applies advanced privacy methods that combine various techniques based on cryptography, encoding, and differential privacy, protecting this analysis from “curious” attackers trying to extract private information from patient clinical data. Additionally, it introduces a set of privacy metrics that allow both data owners and AI model developers to measure the privacy provided by the platform before using it, thereby ensuring total trust.
“A potential use of the TRUMPET platform is to provide a secure and unified service where clinical data owners can share medical records with AI model developers while ensuring patient privacy. This will allow for the construction of increasingly effective AI models that substantially improve current cancer treatments while always respecting privacy and complying with GDPR requirements,” says Loureiro. The second TRUMPET prototype, to be presented in late 2025, will enhance privacy technologies with real data.
TRUMPET comprises nine European entities, including technology centers, hospitals, and universities: Gradiant and Universidade de Vigo (Spain), Arteevo Technologies Ltd. (Israel), Institut National de Recherche en Informatique et Automatique – INRIA and Commissariat à l’Énergie Atomique et aux Énergies Alternatives (France), Instituto Romagnolo per lo studio dei tumori Dino Amadori (Italy), Centre Hospitalier Universitaire de Liège and Time.Lex (Belgium), Technovative Solutions Ltd (United Kingdom).
AI in Healthcare
According to data from the World Health Organization, there were an estimated 20 million new cancer cases worldwide in 2022. The application of AI-based techniques in healthcare will help improve people’s health; however, for AI to be truly effective and efficient, training processes need to be conducted on a large amount of clinical data.
Patient medical data is considered extremely sensitive personal information and is therefore subject to the requirements set forth by the General Data Protection Regulation (GDPR) for its processing.
TRUMPET is not Gradiant’s first innovation in the fight against cancer. The technology center, along with collaborators in the PERSIST project, has developed an AI-based system to prevent the relapse of colon and breast cancer patients.
Trumpet project has received funding from a Research and Innovation action activity under Horizon Europe Framework Programme with Grant Agreement No.101070038
