Towards Standardization of Flow Component Analysis: A 4D Flow CMR Study in Healthy Subjects

Background: Altered intracardiac blood flow patterns may provide novel insights into ventricular function beyond conventional volumetric parameters. 4D flow cardiac magnetic resonance (CMR) allows for comprehensive quantification of left ventricular (LV) flow components, whereas cine and 2D flow techniques remain limited to surrogate measures.

Methods: In this prospective single-center study, 50 healthy volunteers (39.6 ± 11.2 years, 52% female) underwent 3T CMR including cine bSSFP, 2D flow, and contrast-enhanced 4D flow acquisitions. LV net forward volumes (NFV) were quantified across the aortic and mitral valves, and LV flow components (direct flow [DF], retained inflow [RI], delayed ejection flow [DEF], residual volume [RV]) were analyzed using two independent software packages (cvi42 and CAAS MR). Inter- and intra-observer variability as well as agreement between techniques were assessed using intraclass correlation coefficients (ICC), Bland–Altman analysis, and correlation statistics.

Results: 4D flow–derived NFV demonstrated excellent interobserver agreement (ICC 0.91–0.98). Software-dependent differences were observed, particularly at the mitral valve (Δ14.6 ml, p< 0.001). Compared with cine-derived stroke volume, 4D flow NFV was slightly lower (mean bias −6.4 ml). Flow component analysis revealed DF as the largest contributor (≈38%), followed by RI, DEF, and RV. Interobserver reproducibility for flow components was strong (ICC 0.81–0.92). EF derived from DF+DEF correlated well with cine EF (r=0.84, p< 0.001).

Conclusion: 4D flow CMR enables reproducible quantification of LV flow components and NFV in healthy subjects. Software-related systematic differences underscore the need for standardization. Flow component analysis provides functional information closely related to LV ejection and may complement conventional volumetric assessment in future clinical applications.