CMR-Flow

1351046 - 4D and M2D flow sequences show excellent agreement for non-invasive estimation of mean pulmonary artery pressure by CMR

Wednesday, January 25, 2023

Location: E-POSTER

Presenting Author(s)

PB

Pernilla Bergqvist

Undergraduate
Karolinska Institute
Solna, Stockholms Lan, Sweden

Primary Author(s)

PB

Pernilla Bergqvist

Undergraduate
Karolinska Institute
Solna, Stockholms Lan, Sweden

Co-Author(s)

GA

Goran Abdula, MD

MD
Karolinska Institute, Stockholms Lan, Sweden
JC

Jenny Castaings, MSc

Biomedical Scientist
Karolinska University Hospital, Stockholms Lan, Sweden
AF

Alexander Fyrdahl, PhD

Affiliated Researcher
Karolinska Institutet
Hagersten, Stockholms Lan, Sweden
NJ

Ning Jin, PhD

Senior Key Expert
Siemens Medical Solutions USA, Inc., Ohio, United States
DG

Daniel Giese, PhD

Research Scientist
Magnetic Resonance, Siemens Healthcare GmbH, Erlangen, Germany, Germany
FT

Frederik Testud, PhD

Collaboration Scientist Research Associate
Siemens Healthcare AB, Skane Lan, Sweden
PS

Peder Sörensson, MD, PhD

Associate professor
Karolinska Institute, Stockholms Lan, Sweden
AS

Andreas Sigfridsson, PhD

Researcher
Karolinska Institute
Stockholm, Stockholms Lan, Sweden
Martin Ugander, MD, PhD

Professor of Cardiac Imaging
University of Sydney
DM

David Marlevi, PhD

Postdoctoral Fellow
Karolinska Institutet, Massachusetts, Sweden

Background: Four-dimensional (4D) phase-contrast Cardiac Magnetic Resonance (CMR) image analysis can highly accurately and non-invasively estimate mean pulmonary artery pressure (mPAP) by quantifying the duration of persistence for a pathological vortex in the main pulmonary artery [1-2]. This has been achieved using multiple two-dimensional (M2D) flow sequences of time-resolved three-directional phase-contrast velocity encoding with free-breathing and averaging. In comparison, acquisition using a true 4D sequence bears a number of theoretical advantages including isotropic resolution and improved signal-to-noise ratios for out-of-plane flow motion [3]. However, a direct head-to-head comparison between previously utilized M2D and true 4D sequencing has yet to be performed. Therefore, the aim of this study was to validate a state-of-the-art 4D sequence compared to the previously utilized M2D sequence for estimating mPAP in a clinical setting.

Methods: Patients clinically referred for CMR were included (n=30). All patients underwent a CMR exam at 1.5T (MAGNETOM Aera or Sola, Siemens Healthcare, Erlangen, Germany), covering a volume encompassing the pulmonary artery using M2D (research application, spatial resolution: 1.8x1.8x6.0 mm³; temporal resolution: 75.2 ms; venc: 90 cm/s) and 4D (research application, spatial resolution: 2.0x2.5x2.5 mm³; temporal resolution: 58.7 ms; venc: 90 cm/s, with pencil beam respiratory navigator) sequences, both acquired with ECG retrospective gating. Both sequences were reconstructed using compressed sensing (R=8.8 and 7.6 for M2D and 4D, respectively), with a total of 20 cardiac phases generated. mPAP was quantified in a blinded fashion by manual detection of pulmonary artery vortex duration using multiplanar reformatted velocity vector arrows in a flow visualization research application (4D Flow, Siemens Healthcare, Erlangen, Germany), and using a previously determined empirical linear equation describing the relationship between mPAP and vortex duration expressed as percent of the cardiac cycle [1]. Agreement between methods was assessed using linear regression and Bland-Altman analysis.

Results: Pathological vortices were detected in 14/30 patients using both M2D and 4D (absence of any vortex indicating mPAP≤16 mmHg). M2D and 4D mPAP were correlated (R²=0.66, p< 0.001, Fig. 1), with negligible bias and minimal variability (mean±SD difference 0.7±3.2 mmHg, Fig. 2), and no group differences in mPAP (median [interquartile range] 23.9 [22.3–27.0] vs 20.7 [19.1–27.0] mmHg, p=0.74). Scan time was shorter for 4D compared to M2D (135±34 vs 175±36 seconds, p=0.04).

Conclusion: There was excellent agreement between the established M2D and a state-of-the-art 4D Flow sequence for non-invasive estimation of mPAP. For the first time, these results highlight the utility of full volumetric excitation for estimating mPAP, considering both theoretical advantages and apparent shorter acquisition time.