Use of artificial intelligence as a method of detecting cardiovascular risk in patients with obesity: a scoping review

Obesity is a chronic, progressive, and multifactorial disease characterized by an excess of adipose tissue that disrupts metabolic homeostasis and promotes low-grade systemic inflammation. Its prevalence has steadily increased in recent decades, becoming one of the major public health problems worldwide. According to the World Health Organization (WHO), the global prevalence of obesity in adults has more than doubled since 1990, and in 2022 approximately 16% of the world's adult population lived with obesity (1). In absolute terms, it is estimated that more than 1 billion people were obese in 2022, including adults, adolescents, and children (2). This increase has had significant health, social, and economic repercussions. Obesity is associated with increased premature mortality, a decline in quality of life, and greater use of health services. Recent projections suggest that, if current trends continue, the prevalence will continue to rise over the next few decades, especially in low- and middle-income countries (3). In this context, obesity is not only an isolated clinical entity but also a key determinant of multiple chronic non-communicable diseases, most notably cardiovascular diseases. Cardiovascular diseases (CVDs) are the leading cause of death worldwide. According to the WHO, in 2022 they were responsible for approximately 19.8 million deaths, representing about one-third of all global deaths (4). Most of these events are related to modifiable risk factors, such as hypertension, dyslipidemia, type 2 diabetes mellitus, smoking, and obesity. Cardiovascular risk assessment is a central component of primary and secondary prevention. International guidelines, such as those of the European Society of Cardiology (ESC), recommend systematic risk stratification to guide individualized preventive interventions (5). Tools such as SCORE 2 allow for the estimation of the 10-year risk of fatal and non-fatal cardiovascular events, integrating traditional clinical variables (6). However, these scales were developed for general populations and have limitations when applied to subgroups with complex metabolic profiles, such as patients with obesity. Obesity increases cardiovascular risk through multiple interrelated pathophysiological mechanisms. Excess adiposity promotes insulin resistance, activation of the sympathetic nervous system, endothelial dysfunction, a prothrombotic state, and chronic inflammation, contributing to the development and progression of atherosclerosis (7). Consequently, patients with obesity have a higher risk of ischemic heart disease, heart failure, atrial fibrillation, and cerebrovascular disease. Furthermore, obesity rarely occurs in isolation. The coexistence of hypertension, type 2 diabetes mellitus, and dyslipidemia, frequently grouped under the metabolic syndrome, generates a cumulative and dynamic cardiovascular risk. Studies based on theGlobal Burden of Disease Studies have shown that a high body mass index is responsible for a significant proportion of cardiovascular mortality and disability globally (8). Furthermore, the distribution of body fat, especially visceral and ectopic adiposity, is associated with a higher cardiovascular risk, even in individuals with similar body mass index values ​​(9). These characteristics make estimating cardiovascular risk in obese patients particularly complex. Traditional tools may underestimate or fail to adequately capture the true risk, limiting the early identification of high-risk individuals and the timely implementation of intensive preventive strategies. In recent years, artificial intelligence (AI), especially machine learning and deep learning methods, has emerged as a promising tool for improving cardiovascular risk prediction. Unlike classical statistical models, AI algorithms allow for the analysis of large volumes of heterogeneous clinical data and the modeling of complex nonlinear relationships between variables (10). Several systematic reviews have shown that AI models applied to electronic health records can achieve comparable or superior performance to traditional scales for predicting cardiovascular events, although with methodological heterogeneity and limitations in external validation (11). Additionally,Recent scoping reviews highlight the potential of AI to integrate dataclinical, biochemical, imaging and physiological signals, such as the electrocardiogram, in the assessment of cardiovascular risk (12). In the specific context of obesity, initiatives have emerged focused on developing AI tools for predicting obesity-related vascular disease, integrating advanced anthropometric variables, metabolic biomarkers, and lifestyle data (13). However, the available evidence is fragmented and scattered, with substantial differences in the types of models, populations studied, cardiovascular outcomes assessed, and validation methods. The aim of this scoping review is to map and characterize the existing evidence on the use of artificial intelligence as a method for detecting cardiovascular risk in patients with obesity. A scoping review will allow us to identify the scope, characteristics, and gaps in current knowledge, as well as guide future research and facilitate the eventual clinical implementation of these tools in obese populations.