Left ventricular global myocardial strain assessment comparing the reproducibility of four commercially available CMR-feature tracking algorithms

• Published in: European radiology Vol. 28; no. 12; pp. 5137 - 5147
• Main Authors: Barreiro-Pérez, Manuel, Curione, Davide, Symons, Rolf, Claus, Piet, Voigt, Jens-Uwe, Bogaert, Jan
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.12.2018; Springer Nature B.V
• Subjects: Algorithms; Cardiac; Cardiomyopathy; Cardiovascular Diseases - diagnosis; Cardiovascular Diseases - physiopathology; Coefficient of variation; Computer programs; Correlation coefficient; Correlation coefficients; Diagnostic Radiology; Diastole; Dilated cardiomyopathy; Female; Heart; Heart attacks; Heart diseases; Heart Ventricles - diagnostic imaging; Heart Ventricles - physiopathology; Humans; Image acquisition; Imaging; Internal Medicine; Interventional Radiology; Magnetic resonance; Magnetic Resonance Imaging, Cine - methods; Male; Medicine; Medicine & Public Health; Middle Aged; Myocardial Contraction - physiology; Myocardial infarction; Myocardium - pathology; Neuroradiology; Radiology; Reproducibility; Reproducibility of Results; Software packages; Tracking; Ultrasound; Variance analysis; Ventricle; Ventricular Function, Left - physiology; Myocardial infarction; Myocardium; Magnetic resonance imaging; Dilated cardiomyopathy
• ISSN: 0938-7994; 1432-1084
• DOI: 10.1007/s00330-018-5538-4

Objectives To compare the reproducibility of cardiovascular magnetic resonance feature-tracking (CMR-FT) packages to assess global left ventricular (LV) myocardial strain. Methods In 45 subjects (i.e. 15 controls, 15 acute myocardial infarction, 15 dilated cardiomyopathy patients), we determined inter-vendor, inter-observer (two readers) and intra-observer reproducibility of peak systolic global radial, circumferential and longitudinal strain (GRS, GCS and GLS, respectively) comparing four commercially available software packages. Differences between vendors were assessed with analysis of variance (ANOVA), between observers and readings with intraclass correlation coefficient (ICC) and coefficient of variation (CV). Results The normalised end-diastolic volume was 91, 77 and 119 ml/m 2 (median, Q1, Q3) and ejection fraction was 41 ± 14%, range 12-67%. Global longitudinal strain (GLS), global circumferential strain (GCS) and global radial strain (GRS) values were 13.9% ± 5.4% (3.9-23.8%), 12.2% ± 5.8% (1.0-25.1%) and 32.0% ± 14.7 (3.6-67.8%), respectively. ANOVA showed significant differences between vendors for GRS ( p < 0.001) and GLS ( p = 0.018), not for GCS ( p = 0.379). No significant bias was found for both intra- and inter-observer variability. The ICC for inter- and intra-observer reproducibility ranged 0.828-0.991 and 0.902-0.997, respectively. The CV, however, ranged considerably, i.e. 4.0-28.8% and 2.8- 27.7% for inter- and intra-observer reproducibility, respectively. In particular, for GRS differences in CV values between vendors were large, i.e. 5.2-28.8% and 2.8-27.7%, for inter- and intra-observer reproducibility, respectively. Conclusions In a cohort of subjects with a wide range of cardiac performances, GRS and GLS values are not interchangeable between vendors. Moreover, although intra- and inter-observer reproducibility amongst vendors is excellent, some vendors encounter problems to reproducibly measure global radial strain. Key Points • Different software packages are currently available for myocardial strain assessment using routinely acquired cine CMR images. • Global myocardial strain values are not interchangeable between vendors for global longitudinal and global radial strain. • Inter- and intra-observer reproducibility for global strain assessment is excellent. However, some vendors encounter problems to reproducibly measure global radial strain.