Purpose

Cardiovascular magnetic resonance tissue mapping has the potential to quantify changes in the myocardial blood volume in response to physiological exercise (Ex-CMR) for assessment of myocardial ischemia(1). However, rapid heart rate and deep breathing during Ex-CMR challenge the general breath-holding T₁ mapping techniques(1). This study aims to develop a free-running cardiac T₁^* mapping sequence to quantify the Dynamic changEs in myocardiaL T₁ stAr (DELTA) in response to physiological exercise.

Methods

Sequence: Figure 1A shows the diagram of the proposed prototype sequence. DELTA first performs a saturation pulse to renew the whole-heart longitudinal magnetization (Mz). Then, three Look-Locker inversion-recovery experiments are performed. Each Look-Locker experiment continuously acquires golden-angle radial data using low flip angle spoiled gradient-echo. Radial lines acquired in each Look-Locker experiment is fixed to drive the M_z to the steady state (M_ss). The evolution of M_z could be formulated as $$$M_{z}(t)=M_{ss}\left \{1 - \left [2-exp\left ( \frac{-T_{SAT}}{T_1^*} \right ) \right ]exp\left ( \frac{-t}{T_1^*} \right ) \right \}$$$, where T_SAT is the delay after saturation. During the free-running imaging, self-navigation is acquired every 30 readout in the kz direction. ECG signal synchronous with acquisition is recorded to retrospectively extract cardiac motion. Finally, radial lines collected at mid-diastolic and end-expiration phases (Figures 1B and 1C) are selected for the reconstruction of T₁ weighted images using a low rank plus sparsity constraint method(2).
Phantom experiments: Nine phantom vials(3) were scanned by inversion-recovery spin-echo for reference values and DELTA on a 3T scanner (MAGNETOM Vida, Siemens Healthcare, Germany) with body and spine phased-array coils. DELTA performed T₁^* mapping at three slices with a slice-gap of 16 mm (4). For each slice, DELTA performed 10 times for ten T₁^* measurements with the following parameters: FOV=200×200mm², resolution=1.8×1.8mm², thickness=8mm, points per radial spoke=224, TR/TE/flip angle=2.9ms/1.03ms/4º. The radial line is rotated with an angle of 111.25º. Every Look-Locker acquires 1500 spokes. Different measurements of three slices were performed in an interleaved fashion. In vivo experiments: We first investigated the performance of DELTA on nine healthy subjects (4 males, age 25±3 years), which was HIPAA compliant and approved by our Institutional Review Board. Written informed consent was obtained from each subject prior to CMR imaging. MOLLI5(3)3 and DELTA performed T₁^* maps at the base, middle, and apex for three left-ventricle (LV) slices. For each slice, DELTA was performed ten times for ten T₁^* measurements. Measurements of three slices were scanned in an interleaved fashion. Imaging parameters of DELTA were the same as those used in phantom studies. MOLLI5(3)3 was performed with the following parameters: bSSFP, FOV=320×320mm², resolution=1.8×1.8mm², thickness=8mm, TR/TE/flip angle=3.9 ms/1.95 ms/35°, partial Fourier factor =7/8, Acceleration factor=2(4,5). We recruited another five healthy volunteers (2 males, 25±5 years) for Ex-CMR study. Two rest scans were performed before exercise for baseline. Each subject participated in two Ex-CMR studies (1^st and 2^nd) using an MR-compatible bicycle ergometer. The interval between two Ex-CMR was 10 minutes. After the 1^st exercise, DELTA performed four scans. The interval between the two scans was <1 minute. One scan was performed after the 2^nd exercise. In each scan, DELTA only imaged one mid-ventricular slice with ten repetitions.
Data analysis: The mean and standard deviation (SD) of T₁^* for each phantom vial, LV, and septal myocardium were calculated by manually delineating the region of interest. SD was used to assess the T₁^* precision. T₁^* reactivity is calculated as:$$${T_{1}^*}_{reactivity}\%=\frac{{T_1^*}_{post-stress}-{T_1^*}_{rest}}{{T_1^*}_{rest}}\times100$$$, where T₁^*_post-stress represents T₁^* measured after stress and T₁^*_rest is the mean value of two rest scans.

Results

Phantom: In Figure 2A, when spokes increased from 20 to 80, the coefficient of variation of all vials decreased from 2.64% to 1.63% and the overall coefficient of variation was smaller than 3%. Figure 2B show the mean and SD of T₁^* of each vial from 10 measurements by DELTA at the simulated heart rate of 60 bpm. The corresponding relative ΔT₁*%($$$\Delta{T_{1}^*}_{\textit{}i}\%=\frac{{T_1^*}_{\textit{i}}-{T_1^*}_{\textit{avg}}}{{T_1^*}_{avg}}\times100$$$, where T₁*_avg was mean of ten measurements) ranged from -1.5% to 1.8% (Figure 2C). In Figure 2D, the T₁^* was nearly constant at different heart rates with low variation.
In vivo: Figures 3 shows representative T₁^* maps of one healthy volunteer. The mean LV and septal T₁^* measured by MOLLI5(3)3 and DELTA were 835±78ms and 929±65ms (P=0.01), and 863±87ms and 946±74ms (P=0.03), respectively. SD of LV and septal T₁^* was 62±9ms and 67± 13ms (P=0.25), 44±6ms and 46±8 ms (P=0.42), respectively. For DELTA, SD of ΔT₁* of LV and septal T₁* ­were 4.4% and 4.5% respectively. Figure 4 shows T₁^* maps of two healthy subjects imaged at rest and after two exercises. In Figure 5A, the average septal T₁^* was 887±13ms and 901±22ms measured at rest, and 983±35 ms, 946±21ms, 917±24ms, and 920±26ms measured after the 1^st exercise. The corresponding T₁^*_reactivity was 10.0±3.5%, 5.8±2.6%, 2.6±0.9%, and 2.9±1.1%. After the 2^nd exercise, the septal T₁^* and its reactivity were 971±64 ms and 8.1±5.1% (Figure 5B).

Discussion and Conclusion

Phantom T₁^* by DELTA was heart-rate insensitive and had comparable precision to that from MOLLI. The measured T₁^*_reactivity of healthy subjects was correlated with time after exercise-stress. In conclusion, DELTA can quantify myocardium T₁^* and reveal changes in myocardial tissue properties with stress using physiological exercise. Further studies are warranted to establish the diagnostic accuracy of DELTA in detecting myocardial ischemia.

Acknowledgements

No acknowledgement found.