Abstract:
Cardiovascular diseases remain the leading cause of global
mortality and increasingly affect young adults due to early exposure to modifiable
risk factors. Conventional cardiovascular risk assessment tools demonstrate limited
sensitivity in identifying subclinical vascular dysfunction in individuals under 35
years of age. The purpose of this study was to develop and validate an artificial
intelligence–based predictive model for early cardiovascular risk stratification in
asymptomatic young adults.
diagnosed
A prospective analytical study included 328 participants aged 18–35 years without
previously
cardiovascular
pathology.
Clinical,
biochemical,
inflammatory, and lifestyle parameters were integrated into a gradient boosting
machine learning algorithm with five-fold cross-validation. Model performance was
evaluated using accuracy, sensitivity, specificity, F1-score, and area under the
receiver operating characteristic curve.
The developed model achieved an accuracy of 90.2%, sensitivity of 88.7%,
specificity of 91.5%, and AUC of 0.95. Feature importance analysis identified LDL
cholesterol, systolic blood pressure variability, body mass index, high-sensitivity C
reactive protein, physical inactivity, and positive family history as dominant
predictors. Notably, 23.4% of individuals classified as high-risk by the AI model had
low conventional risk scores, indicating improved early detection capability.
Predictive simulation modeling demonstrated a potential 18–21% reduction in
projected 10-year cardiovascular event probability following AI-guided preventive
intervention.
The findings confirm the high discriminative capacity and clinical relevance of
artificial intelligence–based predictive modeling for early cardiovascular screening
in young populations.
