Influence of temporal resolution on computed tomography feature-tracking strain measurements

Abstract

Purpose: Temporal resolution significantly affects strain values demonstrated by Magnetic resonance feature-tracking and speckle-tracking echocardiography. We investigated the influence of R-R interval reconstruction increments on left ventricular (LV) and left atrial (LA) strain measurements of Computed tomography feature-tracking (CT-FT).

Methods: Subjects who underwent retrospective electrocardiogram-gated coronary CT angiography (CCTA) were retrospectively included, and CCTA images were reconstructed in 5% and 10% steps throughout the entire cardiac cycle (0-100% R-R interval). LV global longitudinal strain (GLS), circumferential strain (GCS), radial strain (GRS), LA GLS, ejection fraction (EF), and left atrioventricular coupling indices were computed. We evaluated the consistency and variability of continuous variables between the two reconstruction increment groups, the demarcation between the LA conduit and contraction phases, and observer reproducibility in 20 randomly selected participants.

Results: Eighty-one participants with or without cardiac disease were included. The reconstruction increment of the R-R interval significantly affected the CT-FT-derived strain values. The 5% R-R increment resulted in significantly larger absolute strain values. LV GRS had the greatest difference between the two groups. In the subgroups with heart rates greater than 80 beats per minute or impaired cardiac function, group differences were attenuated, especially for LV GLS, LV GRS, and LA GLS. The prevalence of definite demarcation between the LA conduit and contraction phases was significantly higher in the 5% R-R reconstruction increment group than in the 10% R-R reconstruction increment group. The average heart rate during CCTA scanning was a strong risk factor for indefinite demarcation, which is independent of LVEF. As average heart rate increased, so did the incidence of indefinite demarcation between the LA conduit and contraction phases. The observer reproducibility of LV and LA strain values was independent of the R-R reconstruction increment.

Conclusion: Reconstruction increment of the R-R interval is an important source of variation in LV and LA CT-FT strain values, especially with low heart rate and preserved cardiac function. It is essential to control the heart rate and apply a narrow R-R reconstruction increment to quantify phasic LA strain.

Keywords: Cardiovascular diseases; Computed tomography; Feature-tracking; Reproducibility; Strain; Temporal resolution.