The Vall d’Hebron Research Institute (VHIR), in collaboration with the CIBER Network for Cardiovascular Diseases (CIBER-CV) and researchers from Emory University (United States), has developed a new artificial intelligence application that integrates advanced cardiovascular risk prediction algorithms into the commercial Cardiac ToolBox (Emory ToolBox) platform. This advance makes it possible to automatically calculate the risk of serious cardiovascular complications during the routine analysis of cardiac nuclear medicine studies such as gated SPECT and gated PET.

This innovation represents a pioneering step in the field of nuclear cardiology and cardiac imaging, as it is the first time that cardiovascular prediction models combining clinical data and imaging parameters have been directly integrated into commercial software used internationally for cardiac tomography. Until now, these models required external calculations or independent tools, which made their application in routine clinical practice more difficult.

“The objective is to bring advanced cardiovascular risk prediction tools into real clinical practice, in an automated manner and integrated within the specialists’ usual workflow,” explains Dr. Guillermo Romero-Farina, cardiologist, researcher at VHIR and CIBER-CV, and principal investigator of the project.

From the predictive model to real clinical application

This development continues the line of research promoted by the Vall d’Hebron group over the last fifteen years in coronary risk stratification through the combination of clinical variables and advanced cardiac imaging. Based on this research, the investigators have developed several prognostic algorithms capable of identifying the risk of experiencing serious cardiovascular events such as heart attacks, heart failure, the need for revascularization, or cardiac death.

Now, thanks to integration with artificial intelligence technologies, these algorithms can be incorporated directly into the software used to process cardiac studies. This makes it possible to automatically and immediately generate different risk levels (very low, low, moderate, or high) during the routine analysis of tests, with virtually no additional steps required from healthcare professionals.

“The distinguishing value of this advance is that we are not only developing a predictive model, but also turning it into a tool that can be applied in the healthcare setting,” highlights Dr. Romero-Farina.

The incorporation of these algorithms into the Cardiac ToolBox (Emory ToolBox) platform may facilitate their gradual implementation in real clinical environments. According to the researchers, this automation could help improve the accuracy of cardiovascular risk stratification and assist specialists in making more personalized clinical decisions.

The models integrated into the software were developed and validated using cohorts of nearly 10,000 patients studied in various research projects led by Vall d’Hebron between 2020 and 2024. The results demonstrated high predictive capability across different cardiovascular scenarios, establishing a solid methodological and clinical foundation prior to technological integration.

Unlike other experimental tools, this technology was created with the intention of real clinical application and the capacity to be integrated into the routine operations of nuclear cardiology services. Nevertheless, researchers will continue evaluating the performance of the algorithms in different populations and healthcare settings in order to consolidate their clinical usefulness.

Artificial intelligence applied to cardiac imaging

Cardiovascular medicine is one of the fields in which artificial intelligence is showing the greatest potential to transform clinical practice. The ability to analyze large volumes of clinical and imaging data in an automated manner makes it possible to develop tools capable of identifying risk patterns with greater precision and speed.

In this context, the integration of the algorithms developed at Vall d’Hebron into commercial cardiac imaging software represents a significant step toward more predictive, personalized, and accessible medicine.

“Our goal is to provide professionals with objective and easily accessible tools that help detect patients at high cardiovascular risk at an early stage and facilitate better clinical decision-making,” concludes Dr. Guillermo Romero-Farina.