An echocardiography and ferritin-based screening model to optimise cardiac T2* MRI use for detecting myocardial iron overload in transfusion-dependent thalassemia patients

Abstract

Abstract Background: Myocardial iron deposition drives heart failure in transfusion-dependent β-thalassemia major, yet cardiac T2* magnetic resonance imaging(CMR), the gold standard, is often unavailable in many centres. We evaluated whether two-dimensional echocardiographic measures, together with serum ferritin, could be used as a screening model to identify CMR-defined myocardial siderosis (T2* < 20 ms) with high sensitivity and acceptable specificity. Methods: We enrolled 64 patients (mean age 15.8 ± 4.2 years; 36 male) who underwent two-dimensional echocardiography and T2* CMR within six months, with ferritin measured at the time of echocardiography. Myocardial siderosis was present if T2* < 20 ms on CMR (n = 12). 2 D echo with tissue Doppler parameters and speckle tracking was done in all patients. We used data for normal values of 2D echocardiography and TDI parameters laid out in recent American Society of Echocardiography guidelines. Variables demonstrating p < 0.05 or AUC ≥ 0.65 in univariate and receiver operating characteristic (ROC) analyses were modelled into a multivariate logistic regression alongside ferritin. Results: Cardiac siderosis was seen in 12 out of 64 patients (19%). Patients with iron overloadhad larger right ventricular basal diameter (3.54 cm vs. 3.24 cm; p = 0.024), lower mitral E/A ratio (0.89 vs. 1.06; p = 0.044), and higher septal E/E′ ratio (10.62 vs. 8.47; p < 0.001). Septal systolic velocity (S′) and TAPSE also showed moderate ability to differentiate (AUCs 0.69 and 0.66). However, left ventricular ejection fraction (LVEF) and global longitudinal strain (GLS) did not show a significant difference between groups and had poor predictive value (AUC 0.50 and 0.49, respectively). A multivariable model using five parameters—RV basal diameter, septal E/E′, septal S′, TAPSE, and ferritin showed accuracy with an AUC of 0.81, giving 92% sensitivity and 80% specificity at a 0.30 probability cutoff and negative predictive value of 96%. Aninternal validation was performed using a 5-fold cross-validation model. At the same cutoff, the model maintained high discrimination, with a mean sensitivity of 90% (SD 5%), a mean specificity of 78% (SD 7%), and an average AUC of 0.80 across folds. For practical use in outpatient clinics, a simple rule using just two parameters, septal S′ < 8 cm/s or ferritin >1,850 ng/mL ,was able to identify 92% (11 out of 12) of patients with cardiac siderosis, with a reasonable specificity of 65%. Conclusion: In settings where access to T2* MRI is limited, a combination of basic echocardiographic parameters and ferritin levels can help identify patients at high risk of cardiac iron overload. A simple two-parameter screen (septal S′ < 8 cm/s or ferritin > 1,850 ng/mL) offers a practical, high-sensitivity tool to decide who needs further cardiac MRI evaluation.