Reproducibility and reliability of flow quantification using CMR 2D-phase contrast and 4D-Flow in secondary mitral valve regurgitation

TL;DR

This study compares the reliability of different MRI methods for measuring mitral valve regurgitation in older patients.

Contribution

The study identifies the 2D-PC standard method as the most reproducible for quantifying secondary mitral valve regurgitation using CMR.

Findings

The 2D-PC standard method showed the highest inter-reader and intra-reader reproducibility.

4D-flow methods had excellent intra-reader reliability but variable inter-reader agreement.

Standardized protocols and training could improve the clinical use of these MRI techniques.

Abstract

Accurate quantification of mitral valve regurgitation (MVR) is crucial for patient management. While different MVR quantification methods based on cardiac magnetic resonance imaging (CMR) exist, their reproducibility and reliability remain uncertain. This study aims to evaluate the reproducibility of different CMR 2D-phase contrast (PC) and 4D-flow MVR quantification methods. The inter-reader and intra-reader reproducibility were assessed using intraclass correlation coefficients (ICC). Seven methods were evaluated: 2D-PC standard (LVSV minus aortic flow), 2D-PC mitral-aortic (mitral inflow minus aortic flow), 2D-PC direct (quantifying mitral backflow), 4D-flow standard, 4D-flow mitral-aortic, 4D-flow direct, and volumetric method (LVSV minus RVSV) in 32 patients (74.8 ± 9.8 years, 28% females) with secondary MVR, analyzed independently by two experienced readers. A total of 26 patients were included in the analysis for 2D-PC and 15 for 4D-flow methods. Among all techniques, 2D-PC standard was the most reliable method with both good inter-reader (ICC = 0.85, p < 0.001) and intra-reader agreement (ICC = 0.87, p < 0.001). The 4D-flow standard (ICC = 0.97, p < 0.001) and the volumetric method (ICC = 0.81, p < 0.001) showed excellent and good intra-reader agreements, respectively, but only moderate inter-reader reproducibility (ICC = 0.52, p = 0.027 and ICC = 0.71, p < 0.001). In patients with secondary MVR, 2D-PC standard method demonstrated the highest reproducibility, while 4D-flow methods showed excellent intra-reader reliability but more variable inter-reader agreement. Standardized post-processing protocols and training would likely enhance the clinical application of these techniques. Future studies should investigate these methods in larger, diverse cohorts and correlate findings with clinical outcomes.

The online version contains supplementary material available at 10.1007/s10554-025-03421-x.

Figures