Objective To construct and internally validate a diagnostic model for angiographic obstructive coronary artery disease (obstructive CAD) (defined as ≥50% stenosis on coronary angiography) incorporating the cholesterol, high-density lipoprotein, and glucose (CHG) index and diagonal earlobe crease (DELC) alongside other traditional risk factors. Methods The study employed a cross-sectional design, involving a total of 1,645 patients, who were divided into two groups: those diagnosed with obstructive CAD (n = 1,298) and those without (n = 347). Independent risk factors were screened using least absolute shrinkage and selection operator (LASSO) regression and subsequently incorporated into a binary logistic regression model to construct the diagnostic model. The dose-response relationship between CHG and obstructive CAD risk was examined using restricted cubic spline analysis (RCS). The incremental diagnostic value was examined through DeLong’s test for area under the receiver operating characteristic curve (AUC) comparisons and through integrated discrimination improvement (IDI) and net reclassification improvement (NRI) for risk reclassification and discrimination improvement. Internal validation was performed using the Bootstrap method (B = 500 resamples). The model’s discriminative ability, calibration, and clinical utility were comprehensively assessed through a nomogram, AUC, calibration curve, decision curve analysis (DCA), and clinical impact curve (CIC). Results Lasso regression ultimately identified six independent risk factors: male sex, hypertension, age, serum creatinine (Scr), CHG, and DELC. RCS revealed a linear positive correlation between the CHG index and obstructive CAD risk (P for nonlinear = 0.865). The constructed model yielded an apparent AUC of 0.692 (95% CI: 0.661–0.724) on the full dataset, with an optimistically corrected AUC of 0.683 (95% CI: 0.650–0.715) following internal validation via bootstrapping. Conclusion CHG and DELC represent independent risk factors for obstructive CAD, with CHG levels exhibiting a linear relationship with obstructive CAD risk. The diagnostic model constructed based on these factors could assist in guiding subsequent diagnosis and treatment in high-risk populations.
Lin et al. (Fri,) studied this question.