Abstract
BACKGROUND: This study aimed to create a nomogram to predict a double-zero score for obstructive coronary artery disease (CAD) in a hospital-based cohort initially scoring zero. We compared its diagnostic performance with the Framingham risk score (FRS) and models for atherosclerotic cardiovascular disease (ASCVD). METHODS: We retrospectively reviewed the clinical features and laboratory profiles of 634 participants with baseline zero coronary artery calcium scores. The target population consisted of individuals with a double-zero score. The primary endpoint was the diagnosis of obstructive CAD, defined as CAD-RADS ≥3 or vulnerable plaque formation on the second cardiac CT. The control group had a double-zero score, with no or less than 50% coronary stenosis. We developed a nomogram using a least absolute shrinkage and selection operation-derived logistic model. We assessed the models' discrimination and calibration abilities using the Hosmer-Lemeshow test. RESULTS: Participants were monitored for an average period of 4.26 ± 2.30 years and were randomly allocated to training and validation sets at a ratio of 2.8:1. The study results indicated that 5.13% (24 of 467) in the training cohort and 4.19% (7 of 167) in the validation cohort developed a double-zero score with obstructive CAD progression. This nomogram incorporated four predictors: "systolic blood pressure," "hypertension," "body fat percentage," and "HbA1c." The nomogram demonstrated superior diagnostic performance compared to the FRS and ASCVD models, with lower values of Akaike information criterion and Bayesian information criterion values. The nomogram's discriminative ability, measured by the area under the curve, was 0.792 in the training cohort and 0.824 in the validation cohort. CONCLUSIONS: The validated nomogram provides valuable predictive potential for identifying high-risk subclinical coronary atherosclerosis, thereby supporting personalized primary prevention and education strategies.