Strain-encoded cardiac magnetic resonance imaging: a new approach for fast estimation of left ventricular function

Background Recently introduced fast strain-encoded (SENC) cardiac magnetic resonance (CMR) imaging (fast-SENC) provides real-time acquisition of myocardial performance in a single heartbeat. We aimed to test the ability and accuracy of real-time strain-encoded CMR imaging to estimate left ventricular volumes, ejection fraction and mass. Methods Thirty-five subjects (12 healthy volunteers and 23 patients with known or suspected coronary artery disease) were investigated. All study participants were imaged at 1.5 Tesla MRI scanner (Achieva, Philips) using an advanced CMR study protocol which included conventional cine and fast-SENC imaging. A newly developed real-time free-breathing SENC imaging technique based on the acquisition of two images with different frequency modulation was employed. Results All parameters were successfully derived from fast-SENC images with total study time of 105 s (a 15 s scan time and a 90 s post-processing time). There was no significant difference between fast-SENC and cine imaging in the estimation of LV volumes and EF, whereas fast-SENC underestimated LV end-diastolic mass by 7%. Conclusion The single heartbeat fast-SENC technique can be used as a good alternative to cine imaging for the precise calculation of LV volumes and ejection fraction while the technique significantly underestimates LV end-diastolic mass.


Background
Currently, cine cardiac magnetic resonance (CMR) is the accepted standard of reference for quantification of ventricular volumes, mass and function [1]. According to recent recommendations quantification of left ventricular (LV) volumes is performed using manual contouring of the endocardial and epicardial surface from multiple short-axis planes and LV ejection fraction (EF) and mass are calculated. Although automated contouring algorithms are available the majority of dedicated analysis software still relies on human interaction at least in clinical practice. In recent years, important improvements in CMR techniques have significantly reduced scan time, necessary to image the entire heart, while the duration of semiautomated volumetric analysis did not change significantly and takes up to five minutes [2].
The aim of this pilot study was to assess the ability and accuracy of recently proposed fast-SENC technique to estimate LV volumes, ejection fraction and end-diastolic mass.

Study population
We prospectively invited 35 subjects (12 healthy volunteers and 23 patients with known or suspected coronary artery disease) to participate in this single center study. The coronary artery disease (CAD) was confirmed by previous quantitative coronary angiography. The study complies with the Declaration of Helsinki and was approved by the Ethics Committee of the Charité-Universitätsmedizin Berlin. All individuals were able to give written informed consent before entering the study.

Cardiac magnetic resonance
Study protocol and design details have been previously published [3]. All CMR examinations were performed on a 1.5 T MRI scanner (Achieva, Philips Healthcare, Best, The Netherlands) using a 5-channel phased array receiver coil in the supine position. All study participants were scanned using an identical imaging protocol.
The study protocol included an initial survey to define imaging planes. Cine images were derived using balanced steady state free precession (bSSFP) sequence with short periods of breath-holding in three LV long-axis planes. Short-axis cine images were acquired and used to calculate LV volumes, mass and ejection fraction. The following parameters were used: repetition time (TR) = 3.3 ms, echo time (TE) = 1.6 ms, flip angle = 60°, acquisition voxel size = 1.8 × 1.7 × 8.0 mm 3 , and 30 phases per cardiac cycle.

Data analysis
Before starting the CMR data analysis, the observers were similarly trained by a representative of the software All cine images were analyzed offline using Medis Suite, version 3.0 (Leiden, The Netherlands) or Virtue software (Morrisville, USA) in accordance to a recent consensus document for quantification of LV function and mass using CMR [1], while all fast-SENC images were uploaded from the scanner into a dedicated MyoStrain, version 4.2 software (Morrisville, USA). The enddiastolic and end-systolic cardiac phases were detected visually and after manual contouring of endocardial and epicardial borders LV end-diastolic (LVEDV), LV end-systolic (LVESV), LVEF and LV end-diastolic mass (LVEDM) were calculated. The LV longitudinal and circumferential strain was extracted from three LV short-axis and three LV long-axis fast-SENC images respectively. The global strain values were calculated by averaging measurements obtained from 16 segments for global longitudinal strain (GLS) and 18 segments for global circumferential strain (GCS). 82.00 ± 25.82 g, p < 0.001). Table 1 summarizes the demographic and LV functional parameters of the study population.

Intraobserver and interobserver reproducibility
There was excellent intraobserver agreement for LVEF:  Figure 4 demonstrates Bland-Altman analysis of intraobserver and interobserver agreement for LVEF and LVEDM.

Data acquisition and analysis time
The total scan time spent to acquire CMR images using fast-SENC was 15 s as only six heart beats without any breath-hold are necessary to derive 3 long-axis and 3 short-axis slices. In contrast, conventional cine imaging using bSSFP sequence in a patient with heart rate of 75 beats per minute lasted approximately 45-60 s per slice (30 phases per cardiac cycle including breathing commands and recovery period before the next breath-hold and following slice acquisition could be started). To acquire 3 long-axis and 3 short-axis images the same process was repeated 6 times (for every slice) with a total scan duration of 270-360 s. The time spent for data post-processing using automated contour detection algorithm or machine learning software for fast-SENC (90 s) or bSSFP (120 s) was comparable. However, total time required for data acquisition and analysis was shorter for fast-SENC technique.

Discussion
SENC is an advanced CMR technique for measuring regional myocardial function [4] as an alternative for time-consuming CMR tagging. The utility of SENC for the quantification of regional myocardial deformation has been previously validated in human [5,6] and animal [7] studies. In previous studies the total scan duration for cine and SENC imaging was 30-40 s and for cine and CMR tagging imaging was 56-74 s. Importantly, the time spent for strain analysis per patient was significantly lower using SENC (4.1 min) when compared to CMR tagging (9.2 min, p < 0.001) [8]. A further step to reduce image acquisition time was the development of the fast-SENC, which is a real-time version of SENC that shortened the scan duration to a single heartbeat [7]. These achievements provided several advantages such as elimination of breath-holds or ability to capture dynamic processes such as onset of myocardial ischemia during the stress study or cardiac arrhythmias [7]. In our study we found that using fast-SENC technique image acquisition and post-processing can be performed in less than 2 min. There are no studies comparing the assessment of LV volumes, EF and mass using fast-SENC to other CMR imaging modalities. We did not find any relevant differences between fast-SENC and conventional cine CMR imaging regarding the estimation of LV volumes and LVEF, while fast-SENC significantly underestimated LV EDM. Recently we demonstrated that fast-SENC is a highly reproducible method for assessing LV strain [3]. In this study we found that intraobserver and interobserver agreement for LVEF and LVEDM measurements derived using fast-SENC technique is also excellent.

Conclusion
The single heartbeat fast-SENC technique can be used as a good alternative to cine imaging for the precise calculation of LV volumes and ejection fraction in the interest of time-spent, especially in patients who are unable to perform breath-holds or with cardiac arrhythmias. However, LV mass is significantly underestimated with current fast-SENC technique. LVEF and LVEDM measurements derived using fast-SENC technique are highly reproducible.

Acknowledgements
We thank Anne Wölffel-Gale for editorial assistance.

Funding
This study was not supported with specific funding that could influence study design or results. SK, AP and KK received an unrestricted research grant by Philips Healthcare and a research grant from Myocardial Solutions.

Availability of data and materials
Supporting data and materials can be obtained from the corresponding author on request.