Multimodal Prognostic Model for Predicting Chronic Coronary Artery Disease in Patients Without Obstructive Sleep Apnea Syndrome

Obstructive sleep apnea syndrome (OSAS), with metabolic disorders as a central feature, is closely correlated with coronary artery disease (CAD). Our goal was to develop a prediction nomogram that integrated multimodal data that could accurately predict the prognosis of patients with chronic coronary disease (CCD). We evaluated 393 patients with CCD with a low-to-intermediate pretest probability of OSAS based on polysomnography. A nomogram was constructed by means of least absolute shrinkage and selection operator (LASSO) and multiple Cox regression analyses to identify independent risk factors for major adverse cardiovascular events (MACEs). Two hundred seventy-seven patients were randomly assigned to the training set, and 116 to the verification set. The constructed nomogram consisted of seven clinical variables: age, previous CAD, current alcohol consumption, neck circumference, apnea-hypopnea index (AHI), and triglyceride-glucose index (TyG). The nomogram showed good discriminatory power, as evidenced by Harrell's C-index values of 0.79 (95% confidence interval [CI] 0.731–0.849) in the training set and 0.78 (95% CI 0.678–0.882) in the verification set. Moreover, a high correlation was observed between the predicted and actual incidence of MACEs in both the training and verification sets. Decision curve analysis demonstrated excellent clinical utility of the nomogram based on net benefit and threshold probabilities. We developed an integrated visualized prognostic nomogram that utilizes multi-modal data, including clinical characteristics, AHI, and TyG index, to predict MACEs in patients with CCD. This approach demonstrated excellent performance, highlighting the potential of combining different data sources to enhance prediction accuracy.