Diagnostic Accuracy of Global Longitudinal Strain and Speckle-Tracking-Derived Ejection Fraction for the Detection of Reduced Left Ventricular Ejection Fraction on Cardiac Magnetic Resonance: A Secondary Diagnostic Accuracy Analysis

Abstract

Left ventricular ejection fraction (LVEF) guides diagnosis and treatment across cardiovascular disease, yet cardiac magnetic resonance (CMR), the reference standard, is frequently unavailable in middle-income settings. In this study we determined optimal cut-off values of global longitudinal strain (GLS) and speckle-tracking-derived LVEF for detecting CMR-defined reduced LVEF, compared their diagnostic accuracy, and reported sensitivity, specificity, likelihood ratios, and predictive values. We believe the work is a good fit for PLOS ONE as a methodologically rigorous, hypothesis-generating diagnostic-accuracy study of potential value for settings with limited access to advanced imaging. It is reported following the STARD 2015 guideline (included as Supporting Information) and includes a QUADAS-2 assessment and internal validation of the derived thresholds. In the interest of full transparency, we wish to disclose that this manuscript is a secondary analysis of a patient cohort previously reported by our group in a concordance study among Simpson-derived LVEF, GLS, and CMR-derived LVEF (Herrera-Escandon A, et al. Arq Bras Cardiol Imagem Cardiovasc. 2026;39(2):e20250102). The present study addresses a distinct research question, namely the diagnostic accuracy and optimal cut-off of these echocardiographic markers against a dichotomized CMR reference standard, that was not examined in that report, and it is cited and discussed accordingly in the manuscript. No figures, tables, or text have been reproduced from the prior publication. We make this overlap explicit in accordance with ICMJE recommendations on overlapping publications and are happy to provide the prior article to the editorial office.