A medical AI- project t aims to anticipate clinical risks using data s from more than 15,000 patients

The University has partnered with Siemens Healthineers to launch a project research in medical artificial intelligence that will analyze more than two decades of clinical data from 15,429 patients. The goal of the project, in which the Health Checkup Unit of the Clínica Universidad de Navarra and the Institute of data and Artificial Intelligence (DATAI) at the university are participating, aims to develop algorithms capable of early identification of potential health risks and to provide healthcare professionals with additional information to improve patient prevention and follow-up.

The study is based on Julen Rodríguez Meneses’s doctoral thesis at DATAI, co-supervised by Dr. Manuel Landecho, a specialist in Internal Medicine at the Clinic, and Rubén Armañanzas, head of the laboratory of Digital Medicine at DATAI. “The value of this project lies in the fact that it allows us to view clinical information in a more comprehensive and longitudinal manner. It is not a matter of analyzing data a single clinical test , but rather of integrating data over time to detect patterns that can help us anticipate risks and make better clinical decisions,” notes Landecho.

The study cohort consists of patients treated at the Clinic’s Health Screening Unit. Unlike other initiatives focused on a single source data , the project integrates multimodal information, such as CT scans, electrocardiograms, lab tests, medical reports, and retinal photographs.

Artificial Intelligence for Early Detection

The team will launch a unique data platform capable of linking different clinical modalities, such as medical images, physiological signals, lab results, and medical reports. This integration enables artificial intelligence algorithms to analyze each test comprehensively, cross-referencing complementary information and taking into account the patient’s progression over time.

AI thus makes it possible to identify relationships among data that, taken separately, offer limited information but, when considered together, can reveal relevant signals about future risks. “The main technical challenge is to build an efficient and flexible architecture capable of organizing, harmonizing, and connecting highly heterogeneous information—ranging from medical images to electrocardiogram signals or clinical text,” explains Rodríguez Meneses.

Based on this comprehensive approach, the research aims to identify complex patterns of clinical data that have gone unnoticed in conventional evaluation . “These patterns could help detect early warning signs of risk, improve patient stratification, or support physicians in making preventive and follow-up decisions,” notes Armañanzas. The “ project ” explores the use of advanced deep learning techniques and high-performance computing in the “ design ” of new predictive algorithms that are more accurate than traditional risk scales. According to Armañanzas, “the ‘ goal ’ is to develop algorithms and decision- financial aid systems that complement the clinical professional’s judgment by providing additional information for decision-making.”

Privacy and Ethics from the design

The project is conducted in accordance with strict criteria regarding security, privacy, and the ethical use of health information. The data are pseudonymized and processed in a local, controlled environment, without relying on cloud services and in compliance with the current rules and regulations on the protection of data. Additionally, the research includes mechanisms designed to promote models that are transparent, interpretable, and evaluated based on equity criteria. With this partnership, Armañanzas concludes, “the University and Siemens Healthineers are reinforcing their commitment to precision medicine based on data, which is safe and manager.”