Observer- and sequence variability in personalized 4D flow MRI-based cardiovascular models.

Area of Science:

• Cardiovascular physiology
• Biomedical engineering
• Medical imaging analysis

Background:

• Estimating subject-specific parameters in lumped hemodynamic models is crucial for cardiovascular system analysis.
• Parameter estimation can be challenging due to variability in input data from experimental measurements.
• Previous methods exist for model-based analysis of 4D Flow MRI and cuff pressure data.

Purpose of the Study:

• To investigate the influence of inter-sequence, intra-observer, and inter-observer variability on subject-specific cardiovascular model parameter estimation.
• To assess the reproducibility of parameters describing left ventricular time-varying elastance and aortic compliance.
• To evaluate a previously developed model-based approach using 4D Flow MRI and cuff pressure data.

Main Methods:

• Utilized a model-based approach for analyzing data from 4D Flow MRI and cuff pressure measurements.
• Assessed parameter reproducibility concerning variability in MRI input measurements in ten healthy subjects.
• Investigated variability from inter-sequence, intra-observer, and inter-observer sources.

Main Results:

• Subject-specific parameters showed coefficients of variation between 2.6% and 35% in intra- and inter-observer analyses.
• Comparing parameters from two MRI sequences yielded coefficients of variation between 3.3% and 41%.
• The diastolic time constant of the left ventricle exhibited the lowest variability, while ascending aorta compliance showed the highest.

Conclusions:

• The modeling approach enables estimation of left ventricular elastance and aortic compliance from non-invasive measurements.
• The method demonstrates good to moderate reproducibility concerning intra-user, inter-user, and inter-sequence variability in healthy individuals.
• This approach offers a reliable way to obtain subject-specific cardiovascular parameters non-invasively.