Myocardial Strain Measurements Derived From MR Feature-Tracking: Influence of Sex, Age, Field Strength, and Vendor

Abstract

Background: Cardiac magnetic resonance feature tracking (CMR-FT) is a novel technique for assessing myocardial deformation and dysfunction. However, a comprehensive assessment of normal values of strain parameters in all 4 cardiac chambers using different vendors is lacking.

Objectives: This study aimed to characterize the normal values for myocardial strain in all 4 cardiac chambers and identify factors that contribute to variations in FT strain through a systematic review and meta-analysis of the CMR-FT published reports.

Methods: The investigators searched PubMed, Embase, and Scopus for myocardial strains of all 4 chambers measured by CMR-FT in healthy adults. The pooled means of all strain parameters were generated using a random-effects model. Subgroup analyses and meta-regressions were performed to identify the sources of variations.

Results: This meta-analysis included 44 studies with a total of 3,359 healthy subjects. The pooled means of left ventricular global longitudinal strain (LV-GLS), LV global radial strain, and LV global circumferential strain (GCS) were -18.4% (95% CI: -19.2% to -17.6%), 43.7% (95% CI: 40.0%-47.4%), and -21.4% (95% CI: -22.3% to -20.6%), respectively. The pooled means of left atrial (LA)-GLS (corresponding to total strain, passive strain, and active strain) were 34.9% (95% CI: 29.6%-40.2%), 21.3% (95% CI: 16.6%-26.1%) and 14.3% (95% CI: 11.8%-16.8%), respectively. The pooled means of right ventricular (RV)-GLS and right atrial global longitudinal total strain were -24.0% (95% CI: -25.8% to -22.1%) and 36.3% (95% CI: 15.5%-57.0%), respectively. Meta-regression identified field strength (P < 0.001; I² = 98.6%) and FT vendor (P < 0.001; I² = 98.5%) as significant confounders contributing to heterogeneity of LV-GLS. The variations of LA-GLS_active were associated with regional distribution (P < 0.001; I² = 97.3%) and FT vendor (P < 0.001; I² = 97.4%). Differences in FT vendor were attributed to variations of LV-GCS and RV-GLS (P = 0.02; I² = 98.8% and P = 0.01; I² = 93.8%).

Conclusions: This study demonstrated the normal values of CMR-FT strain parameters in all 4 cardiac chambers in healthy subjects. Differences in FT vendor contributed to the heterogeneity of LV-GLS, LV-GCS, LA-GLS_active, and RV-GLS, whereas sex, age, and MR vendor had no effect on the normal values of CMR-FT strain measurements.

Keywords: cardiac magnetic resonance; feature tracking; healthy subjects; strain.