1353997 - 3D Cardiac Magnetic Resonance Multitasking: First Clinical Experience

Cardiac magnetic resonance (CMR) Multitasking is a novel acquisition and reconstruction method which can resolve cardiac/respiratory motions and acquire quantitative parametric mapping measurements across the entire left ventricle (LV) without the need for ECG gating or breath-holding (1,2). The aim of this study was to test the clinical utility of Multitasking in a larger clinical population by comparing LV ejection fraction (EF) and T2 values measured by Multitasking to reference cine and T2 mapping protocols in a mixed healthy volunteer (HV) and patient population.



Methods: HV and patients referred for a clinical CMR exam were recruited. Exams were conducted on a 3T MR system (Magnetom Skyra™, Siemens Healthineers, Erlangen, Germany). 2D balanced steady-state free precession (bSSFP) images were acquired in the in 2-chamber, 3-chamber, 4-chamber views, and as a short axis stack (SAX) through both ventricles. A standard T2-SSFP sequence was used to acquire 6 SAX T2 maps. The Multitasking sequence was used to acquire a 3D stack of SAX cine and T2 mapping images. Multitasking images were reconstructed using a low rank-tensor framework and then used for LVEF and T2 quantification (1). All data were analyzed using certified software (cvi42, Circle Cardiovascular Imaging Inc., Calgary, AB, Canada). The LVEF was calculated from SAX images. Regression and Bland-Altman analyses were used to compare LVEF measurements between Multitasking and bSSFP images. The HV were used to create sequence-, scanner-, and site-specific T2 normal values. A confusion matrix was created for patient T2 data to compare the number of American Heart Association (AHA) segments classified as having elevated or normal T2 values between T2-SSFP and Multitasking.

Results: 18 patients (mean age 55y., 11 male) with mixed pathologies (Table 1) and 21 HV (mean age 39y., 10 male) were recruited for this study. Multitasking underestimated LVEF with an average difference of 5% compared to bSSFP measurements (Fig. 1A) and demonstrated a moderate linear relationship with bSSFP (r² = 0.68) (Fig. 1B). Multitasking T2 measurements were lower than T2-SSFP on average (Fig.2A and 2B) with larger standard deviations in all AHA segments (Fig.2C and 2D) of the HV cohort. However, in patients Multitasking demonstrated strong agreement with T2-SSFP in classifying AHA segments as having elevated or normal T2 values as determined by a chi-squared test (p = 0.04) (Figure 2E).

Conclusion: This study shows the potential clinical use of Multitasking to quantitatively assess LVEF and T2 in various clinical pathologies at 3T without breath-holding or ECG triggering. Multitasking T2 values measured with higher variability compared to the standard T2-SSFP sequence but was able to correctly classify pathology in 85% of patient cases. Although further adjustments are required to decrease the variability in T2 measurements, this acquisition strategy may have clinical value in patients with trouble holding their breath or in pediatric populations.