**4.1 T1 of hyperpolarized 129Xe**

For conventional MRI, the magnetization at thermal equilibrium is induced by the magnetic field, and the longitudinal relaxation time (T1) is the time for the magnetization, i.e. the magnetic resonance signal, to recover back to the thermal equilibrium. However, hyperpolarized noble gas magnetization produced by SEOP is unable to recover back to the hyperpolarized magnetization by itself, and the T1 of an hyperpolarized noble gas is the decay time of magnetization, because the thermal polarized magnetization is almost zero related to the hyperpolarized magnetization. The longitudinal relaxation time of hyperpolarized 129Xe in the brain is a critical parameter for developing hyperpolarized 129Xe brain imaging and spectroscopy and optimizing the pulse sequences, especially in the case of cerebral blood flow measurements. Various studies have produced widely varying estimates of hyperpolarized 129Xe T1 in the rat brain (Choquet, 2003; Wakai, 2005).

The hyperpolarized magnetization is generally tipped by a pulse with a small flip angle, and it is very challenging to make the T1 as long as possible in order to ensure the signal lasts long enough for the acquisition. Therefore, when considering hyperpolarized 129Xe as a marker for brain perfusion by MRI, evaluation of tissue characterization and pulse sequence optimization, the T1 of hyperpolarized 129Xe in the brain is a critical parameter. Previous attempts to measure T1 in the rat brain have yielded strikingly disparate results. Wilson et al. found that T1 measured in rat brain homogenates in vitro ranged from 18±1 to 22±2 s (mean ± SD) (Wilson et. al,2009) depending on the oxygenation level of the tissue, and T1 values from measurements in rat brain *in vivo* have ranged from 3.6±2.1 (Choquet et al, 2003) to 26±4 s (Wakai et. al, 2005). Part of the discrepancy is believed to be due to the protocols used in T1 determination. The attempt of Choquet et al. used a multi-pulse protocol during the uptake and washout process by injecting hyperpolarized 129Xe in a lipid emulsion, whereas the estimation of Wakai et al. used a two-pulse protocol during the washout process after the rat had breathed hyperpolarized 129Xe gas. Under the condition of typically achieved polarizations (5–21%) (Zook et. al, 2001), low signal-to-noise ratio (SNR) due to the low concentration of the dissolved hyperpolarized 129Xe in tissue is an important factor in making T1 measurements in the rat brain (Cherubinia et. al, 2003; Ruppert et. al, 2000). The maximum SNR in the above two measurements in vivo was only 30 (Choquet et. al, 2003) and 46 (Wakai et. al, 2005), and the noise effect was not considered in these studies. When the SNR is low, noise will dominate the measured signal and result in large differences between the true T1 and the measured T1. Thus, low SNR might be a large contributor of error in the published T1 values.

Hyperpolarized Xenon Brain MRI 133

T1 results independent of the measurement protocol and offer a resolution to the

T1( s) (group 2)

1 14.2 12.9 19.5 17.6

2 12.2 15.1 18.2 16.4

3 11.5 16.1 16.3 14.9

4 11.7 15.5 18.0 16.5

5 12.7 16.4 18.8 16.6

6 12.1 15.7 17.2 15.4

Mean 12.4 15.3 18.0 16.2

Std. Deviation 1.0 1.2 1.1 0.9

(conventional method) and after (improved method) setting a threshold of SNR=5.5 are also

Because there is no background signal from xenon in biological tissue, and because the inhaled xenon is delivered to the brain by the blood flow, we would expect a perfusion deficit, such as could be seen in stroke, to reduce xenon concentration in the region of the deficit. Thermal polarization yields negligible xenon signal relative to hyperpolarized xenon; therefore, hyperpolarized xenon can be used as a tracer of cerebral blood flow (CBF). This subsection will describe that hyperpolarized 129Xe MRI is able to detect, *in vivo* hypoperfused area of focal cerebral ischemia—i.e., the ischemic core area of stroke, by using a rat permanent right middle cerebral artery occlusion (MCAO) model (Zhou et al., 2011a). Stroke is the single most common reason for permanent disability and is the third leading cause of death in developed countries. During acute ischemic stroke, a core of brain cells at the center of the affected region dies quickly, and the damage subsequently spreads to surrounding tissue over the next few hours. Because they allow for the delineation of areas of ischemic neuronal injury and hypoperfusion within minutes after the induction of cerebral ischemia, conventional proton MRI, especially diffusion-weighted imaging (DWI) and perfusion weighted imaging (PWI), have been particularly useful in the diagnosis of acute ischemic stroke. The target of acute stroke therapy is the portion of the ischemic region

Table 2. T1 values of hyperpolarized 129Xe from six rat brains. The mean T1 value and standard deviation obtained from the multi-pulse and two-pulse protocols before

Multi-pulse protocol 2-pulse protocol Conventional Improved Conventional Improved

> T1( s) (group 3)

T1( s) (group 4)

discrepancy between previously reported values.

T1( s) (group 1)

Rat

given. (Zhou et al., 2008)

**5. Stroke MRI with hyperpolarized 129Xe** 

#### **4.2 Multi-pulse and two-pulse washout protocols for measurements of T1**

Hyperpolarized 129Xe transport in the brain has been modeled using appropriate adaptations of the Kety–Schmidt theory (Martin et al., 1997; Peled et al., 1996). Martin and co-workers derived the equation of the cerebral xenon concentration during hyperpolarized 129Xe delivery to the lungs as follow:

$$\frac{dC\_{brain}}{dt} = FC\_{cereb} - (\frac{F}{p} + \frac{1}{T\_{1brain}})C\_{brain} \tag{1}$$

where Ccereb is the xenon concentration in the cerebral artery, Cbrain is the xenon concentration in brain parenchyma, F is the tissue perfusion in units of (volume blood)/(volume tissue)/time, p is the brain/blood partition coefficient for xenon, and T1brain is the longitudinal relaxation time of xenon in the brain. In this equation, the first term on the right describes xenon transport to the brain, and the second term describes the loss of xenon signal due to both perfusion and T1 decay. Xenon signal observed from the brain is proportional to Cbrain. During the washout phase of the xenon signal, there is no transport of hyperpolarized 129Xe by the cerebral artery, and hence Ccereb is zero. Accordingly, the xenon concentration in the brain during washout (Cbrainwashout) is given by the following equation:

$$\frac{d\mathbf{C}\_{b\text{minuschout}}}{dt} = -(\frac{F}{p} + \frac{\mathbf{1}}{T\_{1b\text{minuschout}}})\mathbf{C}\_{b\text{minuschout}}\tag{2}$$

This equation can be solved to yield an analytical solution for the concentration of xenon in the brain during the washout of signal. The decay time constant (τ) of hyperpolarized 129Xe during the washout from the rat brain is given by:

$$\tau = \frac{1}{\left(\frac{F}{p} + \frac{1}{T\_{1b\text{min}}}\right)}\tag{3}$$

Thus, τ can be calculated from a series of pulse excitations (multi-pulse protocol) after compensating for the hyperpolarized xenon signal losses resulted from radio frequency (RF) excitation, as described below. To compare the results obtained using the multi-pulse protocol, a two-pulse protocol has also been adopted to measure τ (Wakai et al., 2005). Both protocols were performed on each rat during the washout phase of the 129Xe signal.

Table 2 shows individual T1 values of hyperpolarized 129Xe and their mean from the six rat brains using the two protocols, with and without the SNR threshold. The mean T1 value calculated using the improved two-pulse method is larger than that using its conventional counterpart, whereas the mean T1 value calculated using the improved multi-pulse method is less than that using its conventional counterpart. These T1 values were named 'group 1' to 'group 4' for easy reference during discussion.

In this subsection, we investigated the error in T1 measurement as a result of low SNR of the 129Xe signal *in vivo*. Correcting for these errors allowed us to more accurately measure the T1 of hyperpolarized 129Xe in the rat brain *in vivo*. Our calculations produced highly consistent

Hyperpolarized 129Xe transport in the brain has been modeled using appropriate adaptations of the Kety–Schmidt theory (Martin et al., 1997; Peled et al., 1996). Martin and co-workers derived the equation of the cerebral xenon concentration during hyperpolarized

<sup>1</sup> ( ) *brain*

where Ccereb is the xenon concentration in the cerebral artery, Cbrain is the xenon concentration in brain parenchyma, F is the tissue perfusion in units of (volume blood)/(volume tissue)/time, p is the brain/blood partition coefficient for xenon, and T1brain is the longitudinal relaxation time of xenon in the brain. In this equation, the first term on the right describes xenon transport to the brain, and the second term describes the loss of xenon signal due to both perfusion and T1 decay. Xenon signal observed from the brain is proportional to Cbrain. During the washout phase of the xenon signal, there is no transport of hyperpolarized 129Xe by the cerebral artery, and hence Ccereb is zero. Accordingly, the xenon concentration in the brain during washout (Cbrainwashout) is given

1

1

1 <sup>1</sup> ( ) *brain*

*F p T*

protocols were performed on each rat during the washout phase of the 129Xe signal.

Thus, τ can be calculated from a series of pulse excitations (multi-pulse protocol) after compensating for the hyperpolarized xenon signal losses resulted from radio frequency (RF) excitation, as described below. To compare the results obtained using the multi-pulse protocol, a two-pulse protocol has also been adopted to measure τ (Wakai et al., 2005). Both

Table 2 shows individual T1 values of hyperpolarized 129Xe and their mean from the six rat brains using the two protocols, with and without the SNR threshold. The mean T1 value calculated using the improved two-pulse method is larger than that using its conventional counterpart, whereas the mean T1 value calculated using the improved multi-pulse method is less than that using its conventional counterpart. These T1 values were named 'group 1' to

In this subsection, we investigated the error in T1 measurement as a result of low SNR of the 129Xe signal *in vivo*. Correcting for these errors allowed us to more accurately measure the T1 of hyperpolarized 129Xe in the rat brain *in vivo*. Our calculations produced highly consistent

This equation can be solved to yield an analytical solution for the concentration of xenon in the brain during the washout of signal. The decay time constant (τ) of hyperpolarized 129Xe

*brain*

*brainwashout*

(2)

(3)

<sup>1</sup> ( ) *brainwashout*

*dC <sup>F</sup> <sup>C</sup> dt p T*

*dC <sup>F</sup> FC <sup>C</sup> dt p T*

1

(1)

*cereb brain brain*

**4.2 Multi-pulse and two-pulse washout protocols for measurements of T1**

129Xe delivery to the lungs as follow:

by the following equation:

during the washout from the rat brain is given by:

'group 4' for easy reference during discussion.


T1 results independent of the measurement protocol and offer a resolution to the discrepancy between previously reported values.

Table 2. T1 values of hyperpolarized 129Xe from six rat brains. The mean T1 value and standard deviation obtained from the multi-pulse and two-pulse protocols before (conventional method) and after (improved method) setting a threshold of SNR=5.5 are also given. (Zhou et al., 2008)
