P5282Comparison of cardiovascular magnetic resonance ejection fraction and left ventricular volumes by long axis feature tracking strain analysis versus standard short axis stack contour tracing

Abstract

Abstract Introduction Cardiovascular magnetic resonance (CMR) feature tracking (FT) is based on the recognition of endocardial features obtained during standard CMR cine imaging to be tracked and followed throughout the cardiac cycle. Global longitudinal strain (GLS) has been proposed as a superior measure for diagnosis and prognosis than ejection fraction (EF). However, EF remains an important primary parameter to describe cardiac function. A rapid determination of GLS based on three long axis views (LAX) allows for a simultaneous calculation of EF without additional imaging or post-processing promising a significant reduction of scan and post-processing time. Purpose The purpose of this work is to compare the LV volumes and EF obtained during assessment of GLS based on CMR feature tracking with standard analysis of a short axis (SAX) stack used as the reference standard. Methods 75 consecutive patients underwent a routine clinical scan obtaining a full SAX stack as well as 3 standard LAX views using either 3-Tesla or 1,5-Tesla clinical scanners. We determined LV volumes and EF based on the reference standard as well as feature tracking analysis with additional GLS. A p value <0.01 was considered statistically significant. Results Mean EF was 45.9% using standard SAX (range, 13%-72%) and 51.1% using triplanar feature tracking (r=0.950; p<0.0001, figure 1A). Bland-Altman analysis showed a systematic bias of 5,27%; without proportional bias (figure 1B). End-diastolic volumes (r=0,975; p<0.0001) and end-systolic volumes (r=0.985; p<0.0001) demonstrated similar results. Mean GLS was −17.3% (range: −30,7% to −3,3%) and was significantly correlated with standard EF (r=−0,884; p<0.0001). Classification of EF into categories: reduced, mid-range or preserved (<40%, 40–49%, ≥50%) remain unchanged in 79% of patients when using EF by feature tracking analysis. Twelve of 16 reclassifications occurred in the mid-range category. Figure 1 Conclusion There is a good correlation between EF obtained by rapid post-processing of GLS with EF based on a full SAX stack resulting in an identical categorization in 79% of patients. Reduction of EF within the mid-range might be best assesses by the standard SAX stack.