A novel prediction model for myocardial fibrosis in patients suspected of myocarditis

Abstract

Purpose: This study aimed at establishing a predictive method that consists of clinical, electrocardiographic (ECG), and laboratory parameters for myocardial fibrosis, especially as detected on cardiac magnetic resonance imaging (CMRI), in patients examined with suspicion of myocarditis.
 Materials and Methods: This study is a retrospective, single-centre study that includes patients admitted to our centre with suspected myocarditis between March 2020 and November 2023. Participants were categorised into two groups (myocardial fibrosis positive and myocardial fibrosis negative), and a detailed comparison of comorbidities, ECG changes, and laboratory parameters was performed. Multivariate analysis was conducted to identify independent predictors of myocardial fibrosis. A nomogram was constructed using the coefficients from the multivariate analysis to estimate the probability of myocardial fibrosis presence based on key predictors. 
 Results: This study included 98 participants with a median age of 30 years, predominantly male (80.6%), with 14.3% having hypertension, 8.2% having diabetes mellitus, and 10.2% being smokers. The myocardial fibrosis-negative group exhibited higher levels of left ventricular ejection fraction and lymphocyte count. Conversely, the myocardial fibrosis-positive group showed higher levels of ECG changes at admission, peak C-reactive protein (CRP), CRP velocity, peak troponin, N-terminal pro-brain natriuretic peptide (NT-proBNP), monocytes, and platelets (PLT). In multivariate analysis, PLT, lymphocyte, monocyte, peak troponin, and ECG changes were identified as independent predictors of myocardial fibrosis. Receiving operating characteristic (ROC) curve analysis showed the model's diagnostic accuracy for predicting myocardial fibrosis (area under the ROC (AUC): 0.959, 95% confidence interval (CI), and p