CMR-Analysis (including machine learning)
.jpg "Felicia Seemann, PhD photo")
Felicia Seemann, PhD
Postdoctoral Fellow
National Heart, Lung, and Blood Institute, National Institutes of Health
Bethesda, Maryland, United States
Felicia Seemann, PhD
Postdoctoral Fellow
National Heart, Lung, and Blood Institute, National Institutes of Health
Bethesda, Maryland, United States
Einar Heiberg, PhD
Associate Professor
Lund University
Lund, Skane Lan, Sweden
Christopher G. Bruce, MD
Staff Clinician
National Heart, Lung, and Blood Institute, National Institutes of Health, District of Columbia, United States
Jaffar M. Khan, MD, PhD
Staff Clinician
National Heart, Lung, and Blood Institute, National Institutes of Health
Washington, District of Columbia, United States
Rajiv Ramasawmy, PhD
Staff Scientist
National Heart, Lung, and Blood Institute, National Institutes of Health
Bethesda, Maryland, United States
Amanda Potersnak, RT (R) (MR)
MRI Technologist
National Heart, Lung, and Blood Institute, National Institutes of Health, United States
Marcus Carlsson, MD, PhD
Professor, Head of Department
Karolinska Institute, Clinical Physiology, United States
Håkan Arheden, MD, PhD
Professor
Lund University
Lund, Sweden
Robert J J. Lederman, MD
Principal Investigator
National Heart, Lung, and Blood Institute, National Institutes of Health
Bethesda, Maryland, United States
Adrienne E E. Campbell-Washburn, PhD
Principal Investigator
National Heart, Lung, and Blood Institute, National Institutes of Health
Bethesda, Maryland, United States
Pressure-volume (PV) loop analysis have potential utility in heart failure evaluation, but have limited adoption due to the invasiveness of pressure measurements. We recently proposed a noninvasive method to derive PV loops. In short, left ventricular (LV) pressure was estimated by multiplying LV volume quantified from CMR short-axis (SAX) cines with a time-varying elastance model that couples pressure and volume1–4 (Fig 1A). The elastance model amplitude is scaled with a validated transfer function using brachial cuff pressures as input to estimate peak aortic pressure which, in the absence of aortic stenosis, is approximately equal to peak LV pressure5. LV end diastolic pressure (EDP) requires a user-estimation, an inherent limitation with a previously shown small impact on the PV loop1.
The method found differences between healthy subjects and heart failure, and between rest and stress1,2. Validation was performed in pigs with CMR and pressure measurements acquired on the same day, but not simultaneously1. Here, we expand this validation with simultaneous pressures and volume measurements, concurrent to a preload reduction.
Methods:
We performed dynamic PV loop experiments under 0.55T MRI-guidance in 15 pigs (n=7 naïve, n=8 five months post induction of ischemic cardiomyopathy by multi-vessel chemoablation). We altered preload by inferior vena cava balloon occlusion at end-expiratory breath-hold6. Simultaneously to occlusion, we measured invasive LV and aortic pressures, and LV volumes from long-axis (LAX) real-time CMR to allow dynamic imaging (bSSFP, TE/TR/θ=1.4ms/77ms/80°, 2.3x2.3x8mm resolution, 360x270mm FOV, acceleration rate 3, 76ms temporal resolution, 326 timeframes). Center line rotation of time-resolved 2D LAX segmentations were used to derive 3D volumes6. End diastolic and end systolic LAX 3D volume pre occlusion were validated with SAX cine volumes. Invasive PV-loops were derived by combining pressure and volume signals6.
Model-based PV loops were derived using real-time LV volumes, peak aortic pressures and invasive LV-EDP to scale LV pressure magnitude, with the approximation V0≈0. Stroke work, mechanical potential energy, energy efficiency, stroke work/(stroke work + potential energy), and peak elastance were quantified and compared (Fig 1B).
Results:
We recorded pressure and volume during IVC occlusion in 15 pigs, totaling 330 PV loops. There was an excellent agreement and low bias between the invasive and model-based stroke work (0.04±0.04J), potential energy (0.01±0.02J), energy efficiency (2.5±2.7%) and peak elastance (0.02±0.07mmHg/ml) (Fig 2). End diastolic and end systolic LAX and SAX volume differed by 7.2±17ml and 2.4±6.0ml, respectively.
Conclusion: An elastance model-based estimation of PV loops and associated hemodynamic parameters at transient loading conditions was validated in pigs. Further human validation of the brachial cuff to peak aortic pressure approximation4, and subject-specific estimations of LV-EDP, and V0 are warranted.