“Multimodal and Personalized Artificial Intelligence for Opportunistic Cardiovascular Risk Screening in Thoracic Tomography”
Description: Clinical reports often omit relevant cardiovascular findings that are visible on chest computed tomography (CT) images, such as calcifications, epicardial fat, and others. This underreporting represents a missed opportunity to intervene in these patients and improve prognosis. Artificial intelligence tools have recently been proposed for the automatic detection of cardiovascular findings in CT, which could reduce the burden on radiologists and improve reporting. However, it is unclear if a comprehensive reporting of these findings can improve cardiovascular risk prediction.
This project proposes the development of an explainable, multimodal, and interactive artificial intelligence system for cardiovascular risk prediction. The approach focuses on the analysis of opportunistic CT exams through the quantification of cardiovascular findings using deep learning and radiomics, as well as clinical reports using natural language processing tools. In addition to risk prediction, interpretability and patient clustering analyses will be applied to identify risk trajectories and modifiable clinical factors.
Amount: €19,620
Funding type: grant
Funder scheme: FCT
Start date: 2025-09-01
End date: 2029-08-31
Investigator: Geanderson Nascimento de Sá
Funded outputs count: 0
Funder award ID: 2025.05961.BDANA
OpenAlex ID: G742876963