A comparison of cine CMR imaging at 0.55 T and 1.5 T

Abstract

Background: There is a renewed interest in lower field magnetic resonance imaging (MRI) systems for cardiovascular magnetic resonance (CMR), due to their favorable physical properties, reduced costs, and increased accessibility to patients with implants. We sought to assess the diagnostic capabilities of high-performance low-field (0.55 T) CMR imaging for quantification of right and left ventricular volumes and systolic function in both healthy subjects and patients referred for clinical CMR.

Methods: Sixty-five subjects underwent paired exams at 1.5 T using a clinical CMR scanner and using an identical CMR system modified to operate at 0.55 T. Volumetric coverage of the right ventricle (RV) and left ventricles (LV) was obtained using either a breath-held cine balanced steady-state free-precession acquisition or a motion-corrected free-breathing re-binned cine acquisition. Bland-Altman analysis was used to compare LV and RV end-systolic volume (ESV), end-diastolic volume (EDV), ejection fraction (EF), and LV mass. Diagnostic confidence was scored on a Likert-type ordinal scale by blinded readers.

Results: There were no significant differences in LV and RV EDV between the two scanners (e.g., LVEDV: p = 0.77, bias = 0.40 mL, correlation coefficient = 0.99; RVEDV: p = 0.17, bias = - 1.6 mL, correlation coefficient = 0.98), and regional wall motion abnormality scoring was similar (kappa 0.99). Blood-myocardium contrast-to-noise ratio (CNR) at 0.55 T was 48 ± 7% of the 1.5 T CNR, and contrast was sufficient for endocardial segmentation in all cases. Diagnostic confidence of images was scored as "good" to "excellent" for the two field strengths in the majority of studies.

Conclusion: A high-performance 0.55 T system offers good bSSFP CMR image quality, and quantification of biventricular volumes and systolic function that is comparable to 1.5 T in patients.

Trial registration: Clinicaltrials.gov NCT03331380, NCT03581318.

Keywords: Cardiovascular magnetic resonance; Cine function; Low-field MRI; Ventricular volumes.

Conflict of interest statement

The authors are investigators on a US Government Cooperative Research and Development Agreement (CRADA) with Siemens Healthcare. Siemens participated in the modification of the CMR system from 1.5 T to 0.55 T.

Dr. Bandettini is principal investigator of a site involved in a multi-center trial sponsored by Bayer. The trial is unrelated to the current work.

Figures

Fig. 1

bSSFP parameter optimization for 0.55 T. (a) Simulations and (b) healthy subject imaging demonstrating parameter optimization for bSSFP cine imaging at 0.55 T by varying flip angle and receiver bandwidth (rBW). Simulated SNR and CNR are scaled relative to simulated 1.5 T SNR and CNR for our standard cine protocol. The yellow dots in (a) and yellow frame in (b) demonstrate the selected parameter combination

Fig. 2

Image quality of 0.55 T and 1.5 T breath-held cine. Examples of 0.55 T and 1.5 T breath-held cine bSSFP in (a) short axis and (b) long axis slices from a patient with a nonischemic cardiomyopathy

Fig. 3

Bland-Altman comparisons of left ventricular measurements at 0.55 T and 1.5 T. Bland Altman comparisons of (a) LVEDV, (b) LVESV, (c) LVED mass, (d) LVES mass, (e) LV stroke volume (SV), and (f) LVEF measured using both breath-held and free-breathing cine protocols. LV measurements are highly reproducibly between 0.55 T and 1.5 T

Fig. 4

Bland-Altman comparisons of RV measurements at 0.55 T and 1.5 T. Bland Altman comparisons of (a) RVEDV, (b) RVESV, (c) RVSV, and (d) RVEF measured using measured using both breath-held and free-breathing cine protocols. RV measurements are highly reproducible between the 0.55 T and 1.5 T scanners

Fig. 5

Example wall motion abnormality at 0.55 T and 1.5 T. Breath-held cine images from 0.55 T (top row) and 1.5 T (bottom row) are provided for a patient with a chronic myocardial infarction and apical aneurysm resulting in regional wall motion abnormality. Videos of wall motion abnormality are provided in Additional file 3

Fig. 6

Diagnostic Confidence scoring results. Histogram of scores of diagnostic confidence from two blinded expert readers for (a) breath-held cine and (b) free-breathing re-binned cine. The majority of the scores fall into the excellent category. A total of 130 measurements were collected (65 subjects × 2 readers)