USING SERUM BIOMARKERS TO PREDICT T1 AND T2 ON CARDIAC MRI IN PATIENTS PRESENTING WITH ACUTE MYOCARDITIS

Abstract

Introduction Contemporary diagnostics for myocarditis include cardiovascular magnetic resonance imaging (CMR). CMR techniques such as T1 and T2 mapping are essential for detecting myocardial inflammation and oedema in acute myocarditis. Serum biomarkers like troponin and C-reactive protein (CRP) could serve as complementary risk stratification tools to predict myocardial tissue abnormalities, such as T1 and T2 values. We aimed to quantify global T1 and T2 and detect associations between serum troponin, CRP and T1 and T2. Methods CMR reports of all patients enrolled to the biorepository of a larger tertiary referral centre (Barts BioResource) were interrogated with a natural language processing algorithm to identify individuals with elevated myocardial T2 signal, indicating oedema. Cases not in keeping with a clinical definition of acute myocarditis and those with a known diagnosis of cardiomyopathy were excluded. Age, sex, body mass index (BMI), peak high sensitivity Troponin T, and peak CRP were collated. CMR images were analysed using CVI42 software. Left ventricular ejection fraction was calculated and total late gadolinium enhancement quantified. A myocardial region of interest was drawn in short axis T1 and T2 images, and global T1 and global T2 were calculated. Data from studies acquired on a 1.5T CMR scanner were split into tertiles of Troponin T. Missing data was imputed using the MissForest algorithm. The Pearson correlation coefficient between global T1 and T2 was calculated. Multivariable linear regression analyses were performed to evaluate the associations of troponin and CRP with CMR T1 and T2 values adjusting for age, sex, time to CMR, BMI and cardiovascular risk factors. All statistical analyses were completed in Python 3.9. Results Cohort characteristics are summarised in table 1 and table 2. 127 studies were identified of which 96 had T1 and T2 mapping data acquired in a 1.5T scanner. Mean global T1 was 1088 +/- 75 ms and mean global T2 was 52 +/- 4.7ms. T1 and T2 were moderately correlated (Pearson correlation coefficient 0.59). In the multivariable analyses adjusting for BMI and cardiovascular risk factors, the highest tertile of troponin was associated with higher T1 (beta = 44, 95% CI 8.8 to 79.2 ms and P = 0.01) (figure 1) and a higher T2 (beta = 2.6, 95% CI 0.4 to 4.8 ms and P= 0.02) (figure 2). Additionally in these models, male sex was associated with lower global T1 and lower global T2 values. CRP was also associated with higher T1 values (beta= 16.2, 95% CI 2.2 to 30.2 ms and P = 0.02) but was not associated with global T2 values. Conclusions In acute myocarditis, peak troponin, a biomarker of myocardial injury was independently associated with markers of CMR inflammation and oedema manifested by higher global T1 and T2 values. In addition, the inflammatory biomarker CRP was associated with higher T1. Biomarkerbased risk stratification using troponin and CRP combined with CMR may enhance diagnostic accuracy in acute myocarditis.