Hsin-Jung yang1, Damini Dey1, Jane Sykes2, John Butler2, Xiaoming Bi3, Behzad Sharif1, Ivan Cokic1, Sotirios Tsaftaris4, Debiao Li1, Piotr Slomka1, Frank Prato2, and Rohan Dharmakumar1
1Cedars Sinai Medical Center, Los Angeles, CA, United States, 2Lawson Health Research Institute, 3Siemens Healthcare, 4IMT School for Advanced Studies Lucca
Synopsis
Despite the
advances to date, myocardial BOLD MRI continues to be plagued by imaging
confounders, which limit its reliability. We hypothesized that (a) the loss in
BOLD sensitivity is dependent on the magnitude of the change in heart rate (HR)
between rest and vasodilator stress; and (b) HR-insensitive T2 maps can enable
BOLD changes to be accurately captured. We tested our hypothesis by examining
the BOLD response to a HR-insensitive T2 mapping approach and conventional T2
mapping. Our results show that reliability of T2-based myocardial BOLD MRI
could be markedly improved through heart-rate-insensitive T2 acquisitions.
Introduction
Myocardial blood-oxygen-level-dependent (BOLD) MRI is a non-contrast approach
for examining myocardial perfusion. Despite the advances to date, myocardial BOLD
MRI continues to be plagued by imaging confounders, which can limit its reliability.
Unlike most applications that rely on cardiac T2 MRI, myocardial BOLD MRI is
acquired at rest and under vasodilator stress, which is often accompanied with
an increase in heart rate (HR). We
hypothesized that (a) the loss in BOLD sensitivity is directly dependent on the
magnitude of the change in HR (ΔHR)
between rest and vasodilator stress; and (b) HR-insensitive T2 maps can enable
BOLD changes to be accurately captured. We tested our hypothesis by examining
the BOLD response to a HR-insensitive T2 mapping approach and conventional T2
mapping. To assess whether ΔHR
leads to a loss in myocardial BOLD sensitivity, we performed numerical
simulations, ex-vivo imaging subjected to simulated ΔHR, and in-vivo imaging during adenosine infusion
in health dogs.Methods
A HR-insensitive
(saturation-recovery (SR) prepared), free-breathing 3D T2 mapping sequence at
3T with near perfect imaging efficiency was developed and studied in a hybrid
clinical PET/MR system. Computer
simulations employing Bloch equations, with parameters corresponding to
commercially available 2D T2 mapping sequence (bSSFP readouts with Cartesian
trajectory, TR/TE = 2.9 ms/1.1 ms, iPAT=2, partial Fourier = 3/4, FA = 35°, BW
= 1184 Hz/ pixel, trigger pulse = 5, FOV = 288x360 mm2, matrix size
=154x192, and voxel size = 2.5 x 1.7 x 6.0 mm3) and the proposed 3D T2 mapping approach (GRE
readout with stack-of-stars trajectory, TR/TE = 3.0 ms/1.5 ms, flip angle (FA)
= 15°, BW = 1100 Hz/ pixel, trigger pulse = 1, and voxel size = 2.0 x 2.0 x 6.0
mm3, SR recovery time= 130ms.), with and without SR preparation were
performed to assess the dependence of heart rate on T2 measurements. To
validate the simulation results and to experimentally determine the influence
of heart rate on T2 maps, freshly excised canine hearts (n=3) were immersed in
saline solution and individually scanned in a head coil with i) a commercially available 2D T2
mapping sequence, ii) the proposed 3D method without SR preparation and iii) proposed 3D sequence (which
includes SR preparation) by
artificially imposing heart rates of 40 – 110 beats/min (bpm) in increments of
10 bpm. 3D
sequences were prescribed with full LV coverage and 2D sequences were prescribed
to the match the 3D partitions. Heart rates were chosen to capture the
typically observed DHR
between conditions of rest and adenosine stress. In-vivo imaging was performed in
healthy dogs (n=10). Conventional 2D T2 mapping, and proposed sequence were prescribed
to investigate the influence of measured T2 by the HR elevation caused by
adenosine stress. Hyperemic response from the in-vivo studies were validated with
simultaneously acquired 13N-NH3 PET perfusion. Myocardial
BOLD Response (MBR) was computed as 100% x [T2(stress)–T2(rest)]/ T2(rest),
where T2(rest) and T2(stress) are the mean myocardial T2 pre- and post
adenosine infusion. BOLD response derived from the conventional T2 map and the
proposed HR-independent approach were compared to ass the loss in BOLD response
from HR elevation during stress scans.Results
Numerical
simulations and ex-vivo studies demonstrated that the proposed approach minimized
the HR-dependent changes in T2 between rest and stress compared to the T2 maps
acquired using conventional and proposed method without SR preparation (Fig. 1).
T2 values acquired using the proposed sequence under adenosine stress was
significantly greater than at rest (38.5±1.0 ms (rest) vs. 44.4±3.1 ms
(stress), p<0.05), which was consistent with the PET perfusion: (0.8±0.1
ml/mg/min (rest) vs 2.0±0.9 ml/mg/min (stress); p<0.05, Fig. 2A and 2B). The
discrepancy between MBR acquired with the proposed method and conventional T2
mapping, referred to here as Loss of Apparent BOLD Contrast [100% x (MBRprop
– MBRconv)], was highly correlated (R=0.7, p<0.05) as shown in
Fig. 2C. Conclusion
Conventional cardiac
T2-based MRI is highly sensitive to ΔHR between rest and
adenosine stress. These changes appear to decrease the expected BOLD response by
counteracting the increase in T2 from vasodilator-induced hyperemia. The
reliability of T2-based cardiac BOLD MRI could be improved through heart-rate-insensitive
T2 acquisitions. Acknowledgements
No acknowledgement found.References
No reference found.