Quantitative free-breathing 3T T₂-mapping of the heart designed for longitudinal studies

Background

Recently, T₂-weighted MRI for the characterization of edema after myocardial infarction has attracted considerable attention (Friedrich, NatRevCardiol2010). Furthermore, the recently proposed combination of bSSFP imaging and T₂Prep for T₂-mapping at 1.5T has enabled a rapid quantitative cardiac T₂ estimation (Huang et al., MRM2007). However, the accuracy of this method may still be limited due to the complex T₂/T₁ signal weighting. Especially for longitudinal studies designed for monitoring and/or guiding therapy, accurate and reproducible T₂ measurements will be critical. A novel quantitative 3T T₂-mapping protocol was therefore developed and tested in both healthy volunteers and patients.

Methods

An adiabatic T₂prep with 3 incremental TE values, affine coregistration, a navigator and 2D radial gradient echo imaging were combined for free-breathing T₂-mapping at 3T with a spatial resolution of 1.25mm. Bloch equation simulations of this sequence were used to optimize scan parameters and to determine an empirical equation that compensates for T₁ relaxation and which returns the “true” T₂. The T₂-mapping sequence and empirical equation were then validated in a series of 15 phantoms in which the true T₂ was determined with a 9-TE spin-echo sequence. Next, the myocardial short axis T₂ of 8 healthy volunteers was mapped in two different scan sessions while a reference phantom (T₂=43.1±0.7ms) was placed next to the thorax. The average myocardial T₂ for both sessions was computed with and without correction with the “true” reference phantom T₂. Finally, this validated protocol was used in 5 patients in the subacute phase after revascularization of acute ST-elevation myocardial infarctions and compared to T₂-weighted TSE imaging.

Results

As a result of both the simulations and phantom scans, optimized sequence parameters included: TE_T2prep=60/30/0ms, T_RR=3 heartbeats, TR/TE=5.3/2.4ms. The empirical equation to determine T₂ was S=S0[exp(-TE_T2prep/T₂)+0.06], where S and S0 are the measured and steady-state signal (Fig. 1a). Scans of the phantoms with known T₂ confirmed a 12±2%(p<0.001) improvement in T₂ estimation with the empirical equation as compared to the standard T₂ decay measurements (Fig. 1b). The myocardial T₂ in the volunteers was homogeneous (42±5ms over all volunteers) and on average showed a 5±2% difference between the two scan sessions. When compensated with the T₂ from the reference phantom, this difference decreased to 2±1% (p=0.02). In all patients, T₂maps could successfully be obtained and a clear demarcation of zones with elevated T₂ values was consistent with the findings on T₂-weighted MRI and X-ray coronary angiography as shown in the example in Fig. 2.

Figure 1

Single pixel T₂-mapping in a simulation and phantom scan. A) Simulated magnetization (black dots) for myocardium with input T₂=45ms at the T₂prep times (60, 30 and 0ms) and fitted curves with the standard (dashed line) and new, empirical (whole line) equation. The new equation leads to more accurate T₂ computations. B) Similar results are obtained in a pixel in a T₂map of a phantom where the T₂ was determined to be 45ms with a 9-TE spin echo scan.

Figure 2

Short-axis T₂map together with conventional T₂-weighted turbo spin-echo and X-ray coronary angiogram in a patient with a myocardial infarct. A) A clearly demarcated zone with elevated T₂ can be seen in the region of the black arrow, which might indicate myocardial edema. The non-infarcted tissue has a homogenous T₂, while the reference phantom adjacent to the thorax appears homogeneous with T₂ values similar to those in healthy tissue. B) The conventional T₂-weighted TSE image confirms the elevated T₂ in the region of the infarct (arrow). C) Consistent with these findings, the x-ray coronary angiogram shows a severe stenosis in an obtuse marginal artery (arrow).

Conclusions

The methodology presented in this study enables robust and accurate cardiac T₂-mapping at 3T, while the addition of a reference phantom improves reproducibility. Therefore, it may be well-suited for longitudinal studies in patients with ischemic heart disease.

Funding

N/A.