Quick Fire Session
SCMR 22nd Annual Scientific Sessions
Mikael Kanski, MD, PhD
Clinical Researcher
NYU School of Medicine
Anders Nelsson, MD
MD
Lund University, Department of Clinical Sciences Lund, Clinical Physiology, Skane University Hospital, Lund, Sweden
Marcus Carlsson, MD, PhD
Associate Professor
Lund University, Department of Clinical Sciences Lund, Clinical Physiology, Skane University Hospital, Lund, Sweden
Håkan Arheden, MD, PhD
Professor
Lund University, Department of Clinical Sciences Lund, Clinical Physiology, Skane University Hospital, Lund
Background: Congestion is an important prognostic factor in heart failure (HF). The pulmonary blood volume (PBV) has been shown to correlate to left atrial pressure (1). PBV measured by cardiac magnetic resonance imaging (CMR) using pulmonary transit time (PTT) has been described and validated (2). Recently, the PROVE-HF study proposed a modified approach to quantify PBV using first-pass perfusion (FPP) imaging, and showed that HF outpatients with increased PBV are at risk of major adverse cardiac events (3). The PROVE-HF method, however, has not been compared to the PTT method. Therefore, the purpose was to perform a head-to-head comparison between PROVE-HF and new approaches of quantifying the PBV with FPP, using PTT as reference.
Methods: Thirty patients (60±13 years) underwent CMR on a MAGNETOM Aera 1.5T system (Siemens Healthcare, Erlangen, Germany). The CMR protocol included flow measurement in the pulmonary trunk (for right ventricular stroke volume [RVSV]) and cardiac output [CO]) as well as prototype sequences for PTT and FPP. PTT measures the time it takes for a contrast bolus to pass from the pulmonary artery to the left atrium with a high temporal resolution (2). PBVPTT is calculated using PTT× CO (Figure 1). PBV by the PROVE-HF method uses FPP to calculate the pulmonary transit beats between the maximal signal intensity in the most basal slice in the right ventricle and left ventricle, and calculated PBV as pulmonary transit beats × RVSV (3). FPP peak to peak and FPP center of gravity (CoG) used the same FPP slices and positioning but utilize peak values and CoG from the respective FPP signal intensity curves to determine the time difference, calculating PBV as time difference × CO. PBV was indexed to body surface area (PBVI).
Results: Results are shown in Figure 2 and 3. Highest correlation and precision as well as lowest interobserver variability was seen between FPPCoG vs PTT (correlation: R2=0.72, p2=0.95, p+8%). Corresponding values for PROVE-HF vs PTT were (correlation: R2=0.33, p=0.001, bias 14±30%, interobserver variability: R2=0.77, ppeak-peak vs PTT were (correlation: R2=0.38, p=0.0003, bias 19±29%, interobserver variability: R2=0.82, p<0.0001, bias 1+18%).
Conclusion:
Compared to the PBVI reference method (PTT), the FPPCoG approach showed higher correlation, higher accuracy, and lower interobserver variability when quantifying the PBVI. Utilizing the CoG method could lead to higher sensitivity and specificity when identifying patients with increased PBVI. We therefore suggest the CoG as the method of choice for quantification of PBVI when designing studies on congestive heart failure.
Acknowledgement: We thank Dr. Kelvin Chow, Siemens Healthcare, Erlangen, Germany for providing the SASHA sequence from which the PTT sequence was developed in-house and Dr. Peter Kellman, NIH, Bethesda, USA for providing the FPP sequence.