ABSTRACT
BACKGROUND: Three-dimensional (3D) speckle tracking echocardiography (STE) can overcome some of the inherent limitations of two-dimensional (2D) STE; however, clinical experience is lacking. We aimed to assess and compare the feasibility, agreement, and reproducibility of left ventricular (LV) global longitudinal (GLS), and regional strain by 3D vs 2D STE in normal children. METHODS: Healthy pediatric subjects (n = 105, age mean = 11.2 ± 5.5 years) were prospectively enrolled. Three-dimensional and 2D LV GLS, as well as regional strain in 16 myocardial segments were quantified. Bland Altman analysis, intra- class correlation coefficients (ICC), percent error and linear regression were used for agreement and correlation between the two techniques. Analysis and acquisition times were compared. Inter- and intra-observer reproducibility was assessed in 20 studies. RESULTS: There was good to excellent agreement for 2D and 3D global longitudinal strain (ICC =0.82) and modest agreement for regional strain (ICC range 0.43-0.71). Both methods had high feasibility (88.6% for 2D vs 85.7% for 3D, p = 0.21), although 3D STE required significantly shorter acquisition and analysis time than 2D STE (acquisition time 1 ± 1.2 mins vs 2.4 ± 1 mins; p = 0.03, analysis time = 3.3 ± 1 mins vs 8.2 ± 2.5 mins; p = 0.001, respectively). Inter and intra-observer reproducibility was excellent for GLS by the two techniques (ICC = 0.78-0.93) but moderate to poor for regional strain (ICC = 0.21-0.64). CONCLUSION: Three-dimensional global LV strain is as feasible and reproducible as 2D strain, with good agreement yet significantly more efficient acquisition and analysis. Regional strain is less concordant and 2D and 3D values should not be used interchangeably. 3D LV GLS may represent a viable alternative in evaluation of LV deformation in pediatric subjects.
