Accurate myocardial T1 mapping at 5T using an improved modified Look-Locker inversion recovery method: A validation study

Figures

• 

  Fig. 1. The timing diagram of the coMOLLI sequence. The acquisition follows a 5-(3)-3 MOLLI scheme. Initially, five images (Img1-Img5) are acquired in successive mid-diastolic heartbeats after the first inversion pulse (orange arrow). A three-heartbeat rest period is then inserted to allow recovery of longitudinal magnetization before the second inversion (orange arrow), after which three additional images (Img6-Img8) are acquired. Colored bars denote GRE readouts, and the curve depicts the recovery of longitudinal magnetization (Mz) following each inversion pulse. coMOLLI combined-correction modified look locker inversion recovery, MOLLI modified Look-Locker inversion recovery, GRE radiofrequency spoiled gradient recalled echo

• 

  Fig. 2. Pre-scan for estimating the inversion efficiency δ. An ECG-gated GRE scan comprising two single-shot acquisitions first obtains a PDw single-shot image (I0). After a five-heartbeat rest period, an inversion pulse (orange arrow) is applied, and a second single-shot image (IIR) is acquired immediately. Myocardial ROIs are drawn on I0 and IIR to compute the mean signals. δ is defined as the ratio of the mean myocardial signals of IIR and I0. ECG electrocardiogram, GRE radiofrequency spoiled gradient recalled echo, PDw proton density weighted, ROI region-of-interest

• 

  Fig. 3. The inversion efficiency δ maps for the optimal HSn and Tan/Tanh designs across the B0 and B1 ranges (ΔB0: ±250 Hz, B1: 225–540 Hz). The white dashed contours denote lines of constant δ, and the red dashed rectangles mark the target operating window. HSn hyperbolic secant, Tan/Tanh tangent/hyperbolic tangent

• 

  Fig. 4. Bland-Altman analysis of vial-wise T1 measurements in phantoms, comparing coMOLLI and IR-FSE. The blue points show the relative difference (%) plotted against the mean T1 of the two methods. The dashed black line shows the mean bias (−0.72%), and the red dashed lines show the 95% limits of agreement (−5.89% to 4.45%). IR-FSE inversion recovery-fast spin echo, coMOLLI combined-correction modified look locker inversion recovery, SD standard deviation

• 

  Fig. 5. Native myocardial T1 maps acquired with coMOLLI in three healthy volunteers at apical, middle, and basal slices. These examples illustrate the expected healthy-subject appearance, with homogeneous myocardium and a uniform blood pool signal. coMOLLI combined-correction modified Look-Locker inversion recovery

• 

  Fig. 6. Representative native (left) and post Gd infusion (right) T1w images and corresponding T1 maps from three patients. The T1w images are provided for anatomical reference. Gd gadolinium