Skip to main content
2025 SCMR Conference Main banner
Final Program


Schedule at a Glance
Back

Rapid Fire Abstracts

Rapid Fire 06- Post-Processing and Workflow

Confounders in CMR-feature tracking derived left atrial strain analysis: A multi-vendor, multi-centre comparison (RF_FR_346)

Friday, January 31, 2025
11:30 AM – 11:40 AM East Coast USA Time
Location: Blue Room PreFunction
Background: Cardiovascular Magnetic Resonance-feature tracking (CMR-FT) strain analysis enables the assessment of cardiac deformation and function [1]. Left atrial (LA) strain analysis has been proven to have prognostic relevance, even in early cardiovascular disease [2,3]. Identifying confounders is crucial for comparing results across centres and distinguishing between health and disease. This study aims to evaluate the impact of field strength, post-processing software and subject-related confounders on LA strain results.


Methods: This study included 149 healthy participants from one centre (n = 123 examined at 1.5T, n = 26 at 3T). Of the 149 participants, 80 (54%) were female, ages ranged from 19 to 77 years (mean ± SD: 39 ± 14 years). A subgroup of n = 18 travelling volunteers underwent CMR at four different centres (one with 1.5T and three with 3T, one vendor) within a median of 8 days to assess inter-centre reproducibility. Three different post-processing software (1: CVI42, 2: TrufiStrain3.0 Research prototype[4], 3: Medis) were used to evaluate LA reservoir, conduit and contractile strain. Examples for each software are shown in Figure 1. FT-strain analysis was conducted on standardized bSSFP cine 2- and 4-chamber views with 30 phases. A mixed model with repeated measures was applied, and the following fixed factors included to assess their relation to strain components: software, scanner, field strength, centre, age, sex, BMI, heart rate, and diastolic and systolic blood pressure.


Results: The values of all LA strain components differed across post-processing software vendors (p< .0001). Specifically, reservoir and contractile strain values were lowest in CVI, followed by TrufiStrain and Medis. In contrast, conduit strain values were lowest in TrufiStrain followed by CVI and Medis. Detailed results for each strain component across the software are presented in Table 1. Field strength did not significantly affect strain results. A negative association was observed between age and both reservoir and conduit strain, indicating that strain values decreased with increasing age (r=-0.327, p=.0001 and r=-0.418, p=.0001, respectively). A positive association was found between age and contractile strain, indicating that strain values increased with age (r=0.268, p=.001). Differences were observed in all three strain components between age groups ≤50 and >50 years (all p< .01). Sex was not found to influence LA strain results. Within the traveling volunteer group, strain values did not differ between centers when derived from the same software (see Figure 2).


Conclusion:

Post-processing software was found to be a significant confounder in CMR-FT based LA strain analysis. In addition, age had a significant influence on LA strain values.


Therefore, age- and software-specific normal values should be used. CMR-derived LA strain demonstrates excellent inter-centre reproducibility, providing a valuable foundation for future multicentre studies.