Using advanced machine learning and synthetic imaging data, Sant Pau is leading efforts in stroke aetiology prediction while safeguarding patient privacy.

As part of the Europe-wide SEARCH project, the Hospital de la Santa Creu i Sant Pau (Barcelona) is spearheading the development of a clinical study focused on improving diagnosis and prediction of acute cardioembolic stroke through federated artificial intelligence (AI) systems and synthetic data.

SEARCH (Synthetic hEalthcare dAta goveRnanCe Hub), launched under the Innovative Health Initiative, seeks to accelerate innovation in precision medicine by enabling healthcare partners to train and validate AI models on privacy-preserving, FAIR-compliant synthetic datasets.

Within this initiative, Sant Pau is coordinating one of the project’s four clinical validation studies. This study involves patients with acute ischemic stroke, including brain CT and CT angiography data. The goal is to define imaging biomarkers and anatomical risk signatures that distinguish cardioembolic strokes from other types.

“In cardioembolic stroke, early detection is crucial but remains challenging due to complex imaging patterns and variability in patient anatomy,” said Dr. Josep Munuera from the Radiology Department.

The curated dataset and study protocols developed by Sant Pau lay the foundation for AI model training across the SEARCH federated platform. By comparing models trained on real versus synthetic data, the team seeks to validate new synthetic image generation techniques that could enhance diagnosis, risk assessment, and personalized treatment planning.

The study also contributes to a broader multimodal database by identifying rare anatomical variants and comorbidity profiles that influence stroke outcomes. These insights will support the development of next-generation clinical decision support tools within the SEARCH ecosystem.

Sant Pau’s contribution marks a major milestone in proving how federated AI and synthetic imaging can be applied in real clinical settings to create robust, generalizable, and privacy-preserving diagnostic tools.

This groundbreaking work brings clinical AI research one step closer to transforming stroke care and delivering better outcomes for patients across Europe.

Using advanced machine learning and synthetic imaging data, Sant Pau is leading efforts in stroke aetiology prediction while safeguarding patient privacy.

As part of the Europe-wide SEARCH project, the Hospital de la Santa Creu i Sant Pau (Barcelona) is spearheading the development of a clinical study focused on improving diagnosis and prediction of acute cardioembolic stroke through federated artificial intelligence (AI) systems and synthetic data.

SEARCH (Synthetic hEalthcare dAta goveRnanCe Hub), launched under the Innovative Health Initiative, seeks to accelerate innovation in precision medicine by enabling healthcare partners to train and validate AI models on privacy-preserving, FAIR-compliant synthetic datasets.

Within this initiative, Sant Pau is coordinating one of the project’s four clinical validation studies. This study involves patients with acute ischemic stroke, including brain CT and CT angiography data. The goal is to define imaging biomarkers and anatomical risk signatures that distinguish cardioembolic strokes from other types.

“In cardioembolic stroke, early detection is crucial but remains challenging due to complex imaging patterns and variability in patient anatomy,” said Dr. Josep Munuera from the Radiology Department.

The curated dataset and study protocols developed by Sant Pau lay the foundation for AI model training across the SEARCH federated platform. By comparing models trained on real versus synthetic data, the team seeks to validate new synthetic image generation techniques that could enhance diagnosis, risk assessment, and personalized treatment planning.

The study also contributes to a broader multimodal database by identifying rare anatomical variants and comorbidity profiles that influence stroke outcomes. These insights will support the development of next-generation clinical decision support tools within the SEARCH ecosystem.

Sant Pau’s contribution marks a major milestone in proving how federated AI and synthetic imaging can be applied in real clinical settings to create robust, generalizable, and privacy-preserving diagnostic tools.

This groundbreaking work brings clinical AI research one step closer to transforming stroke care and delivering better outcomes for patients across Europe.