Artificial Intelligence in Cardiovascular Imaging: Advances and Future Challenges

Abstract

Artificial intelligence (AI) has transformed cardiovascular medicine, offering new possibilities for the diagnosis, prognosis, and treatment of cardiovascular diseases (CVD), which remain the leading cause of global mortality. Traditionally, risk stratification has been based on clinical factors such as hypertension, diabetes, and smoking, but the different manifestations of these factors suggest the need for more accurate and personalized approaches. Cardiovascular imaging, including computed tomography (CT), magnetic resonance imaging (MRI), and echocardiography, has emerged as a key provider of detailed information on cardiac structure and function, as well as coronary atherosclerosis. CT, for example, allows the quantification of coronary calcium and the identification of vulnerable plaques, such as low-attenuation and positive remodeling, which are associated with increased risk of cardiovascular events. The integration of clinical and imaging data through multimodal fusion models has shown promise for improving the prediction of adverse events. These models combine information from multiple sources, such as electronic patient records, imaging data, and electrocardiograms, to create more accurate and personalized predictions. However, manual extraction of these data is time-consuming and subject to bias, limiting their clinical application. Automation of this process through artificial intelligence techniques has been explored to facilitate data extraction, allowing the creation of more efficient and widely applicable fusion models. Furthermore, AI has been applied to the automatic analysis of cardiac structures, quantification of ventricular volumes and function, and detection of phenotypic features of high-risk atherosclerotic plaques. These applications not only improve the efficiency of clinical workflow but also have the potential to identify subtle patterns that may go unnoticed in traditional human analysis. This article explores the current state of AI in cardiovascular imaging, highlighting the application of multimodal fusion models and future prospects for the clinical implementation of these technologies, aiming at more personalized and precise medicine.