Abstract
BACKGROUND: VT (Ventricular Thrombus) is a serious complication of dilated cardiomyopathy (DCM). Our goal is to develop a nomogram for personalized prediction of incident VT in DCM patients. METHODS: 1267 patients (52.87 ± 11.75 years old, 73.8% male) were analyzed retrospectively from January 01, 2015, to December 31, 2020. A nomogram model for VT risk assessment was established using minimum absolute contraction and selection operator (LASSO) and multivariate logistic regression analysis, and its effectiveness was validated by internal guidance. The model was evaluated by the area under the receiver operating characteristic curve (AUC), calibration curves, and decision curve analysis (DCA). We compared the performance in predicting VT between nomogram and CHA2DS2, CHA2DS2- VASc or ATRIA by AUC, akaike information criterion (AIC), bayesian information criterion (BIC), net reclassification index (NRI), and integrated discrimination index (IDI). RESULTS: 89 patients (7.02%) experienced VT. Multivariate logistic regression analysis revealed that age, left ventricular ejection fraction (LVEF), uric acid (UA), N-terminal precursor B-type diuretic peptide (NT-proBNP), and D-dimer (DD) were important independent predictors of VT. The nomogram model correctly separates patients with and without VT, with an optimistic C score of 0.92 (95%CI: 0.90-0.94) and good calibration (Hosmer-Lemeshow χ(2) = 11.51, P = 0.12). Our model showed improved prediction of VT compared to CHA2DS2, CHA2DS2-VASc or ATRIA (all P < 0.05). CONCLUSIONS: The novel nomogram demonstrated better than presenting scores and showed an improvement in predicting VT in DCM patients.