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
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Global longitudinal strain (GLS) and speckle-tracking-derived left ventricular ejection fraction (LVEF) are increasingly used to assess systolic function, but no consensus cut-off exists for identifying reduced LVEF as defined by cardiac magnetic resonance (CMR), the imaging reference standard. Our group previously reported excellent concordance between echocardiographic and CMR-derived LVEF in a cohort of 35 patients, but concordance statistics do not translate directly into clinically actionable diagnostic thresholds.
Objectives
To determine optimal cut-off values of GLS and speckle-tracking-derived LVEF for detecting CMR-defined reduced LVEF (<50%), and to compare their diagnostic accuracy, in a secondary analysis of this previously described cohort.
Methods
In this cross-sectional diagnostic accuracy study, 34 of the 35 originally enrolled patients (with complete strain and CMR data) were analyzed. Receiver operating characteristic (ROC) curve analysis estimated the area under the curve (AUC) for each index test against CMR-derived LVEF <50%. Optimal cut-offs were selected using the Youden index. Sensitivity, specificity, predictive values, and likelihood ratios were reported with 95% confidence intervals (CI). AUCs were compared using the DeLong test.
Results
Reduced LVEF (<50%) by CMR was present in 12 of 34 patients (35.3%). GLS showed an AUC of 0.892 (95% CI 0.729-0.996), with an optimal cut-off of -17.0% (sensitivity 75.0%, specificity 86.4%; positive likelihood ratio 5.50, negative likelihood ratio 0.29). Speckle-tracking-derived LVEF showed an AUC of 0.771 (95% CI 0.538-0.971), with an optimal cut-off of 52.8% (sensitivity 58.3%, specificity 95.5%). The difference in AUC was not statistically significant (DeLong p=0.192).
Conclusions
A GLS cut-off of -17.0% showed good discriminative accuracy for detecting CMR-defined reduced LVEF, numerically superior to but not statistically different from speckle-tracking-derived LVEF. These exploratory, hypothesis-generating findings require external validation in larger, multicenter cohorts before clinical application.
