CN105699923A - Magnetic resonance imaging method for measuring R2, R2* and R2' parameter image of tissue in noninvasive and dynamic manner - Google Patents

Abstract

The invention discloses a magnetic sensitivity characteristic of different degrees by using different deoxygenated hemoglobin contents in tissues, by using an asymmetric spin echo magnetic resonance imaging sequence, and adopting a strategy of periodically moving 180 degrees to refocus the radio frequency pulse position, in each cycle Let the 180-degree refocusing pulses be symmetrically distributed on both sides of the TE/2 time to obtain multi-echo images with high time resolution; further, for the collected multi-echo signals, move with the corresponding 180-degree refocusing pulses The period length is the window width for motion estimation, and the dynamic parameter images of R2, R2* and R2' are simultaneously obtained according to the signal attenuation exponential model by using the least square estimation. Using this method, dynamic R2, R2* and R2' parameter images with high temporal resolution can be obtained, and dynamic curves of any region of interest can be displayed.

Description

A Magnetic Resonance Imaging Method for Noninvasive and Dynamic Measurement of Tissue R2, R2* and R2` Parameter Images

technical field

The invention belongs to the technical field of magnetic resonance imaging (MRI), in particular to a set of methods for dynamically and noninvasively and quantitatively measuring the intrinsic relaxation times R2, R2* and R2' of different tissues based on the magnetic resonance imaging technology.

Background technique

Susceptibility, as an important contrast mechanism, has a wide range of applications in magnetic resonance techniques. For example, blood oxygen level-dependent imaging (BOLD) is an imaging method developed based on the paramagnetic properties of deoxygenated hemoglobin. Blood) functional evaluation.

In magnetic resonance imaging, susceptibility effects affect the R2* parameter, and R2* is the free decay rate of the tissue signal, usually fitted using a single-exponential decay model. And R2* contains two components: (1) one is the irreversible component R2, and (2) the other is the reversible R2', which reflects the degree of phase dispersion in the pixel caused by magnetic sensitivity. In a single exponential model, R2*=R2+R2'. Typically, tissue susceptibility is measured using R2* rather than R2', although the latter is not affected by changes in R2 due to pathological causes. On the other hand, R2* and R2 imaging are combined in dynamic susceptibility perfusion-weighted imaging (DSC-PWI) and functional magnetic resonance imaging (fMRI) studies to quantitatively describe tissue hemodynamic information. For example, Prinster et al. used the BOLD technique to use the relative change ratio of relaxation time (ΔR2/ΔR2*) under hypoxic and hypercapnic conditions as an indicator to distinguish voxels containing large vessels from small vessels. Further, R2 and R2* imaging information after paramagnetic contrast agent injection was used to provide quantitative information on the mean size of cerebral vessels. To sum up, a sequence that can obtain accurate quantitative information of R2 and R2* at the same time is very necessary.

Generally speaking, R2 and R2* information can be obtained by using a multi-echo spin echo sequence and a multi-echo gradient echo sequence, respectively. In addition to the disadvantage of longer time for this method, the strategy of obtaining R2 and R2* from two scans may cause a large estimation error between R2 and R2* due to physiological changes between the two scans. Recently, a gradient echo sampled free decay echo (GESFIDE) sequence was proposed to simultaneously obtain R2 and R2* information. In this sequence, multiple gradient echoes are used to simultaneously acquire the free-decay magnetization after the 90° excitation pulse and the spin magnetization after the 180° refocusing pulse. R2* can be obtained by using the magnetization intensity during the free decay process, while R2 can be obtained by comparing the echo intensity at the left and right symmetrical positions of the 180° pulse. A sequence called gradient echo sampled spin echo (GESSE) similar to GESFIDE is also proposed to obtain R2 and R2* information simultaneously. In this sequence, multiple gradient echoes are used to simultaneously acquire spin echoes to form and magnetization during decay. However, the above two methods both adopt the traditional Cartesian K-space acquisition strategy, resulting in slow imaging speed, which cannot meet the requirements of high temporal resolution in applications such as DSC-PWI or fMRI. Recently, a multi-echo spin and gradient echo (SAGE) echo planar imaging (EPI) sequence that can dynamically obtain R2 and R2* information has been proposed. However, in this sequence, non-ideal slice-selection profiles caused by short radio frequency (RF) pulse times lead to a mismatch of image slice profiles acquired before and after the 180° refocusing pulse, which will be discussed later in R2 Large estimation error occurs in the quantitative calculation of R2* and R2*.

As mentioned above, existing methods have the disadvantages of low temporal resolution or large estimation errors. Therefore, a dynamic method that can simultaneously and accurately obtain R2, R2* and R2' information is urgently needed.

Contents of the invention

In this patent, we propose a multi-echo asymmetric spin-echo (psMASE) sequence that periodically shifts the pulse position by 180°. In this sequence, since the signal is acquired during spin echo formation and decay, the problem of slice profile mismatch is avoided. Further, in order to improve the temporal resolution of dynamic imaging, a strategy of moving window estimation is introduced into the quantitative calculation of R2, R2* and R2', and the length of the moving window is the same as the period length of the psMASE sequence. Taking the period as 3TR as an example, the specific method flow is shown in FIG. 1 .

For traditional multi-echo ASE sequences, the position of the 180° refocusing pulse can be flexibly set. Theoretically, for a certain 180° pulse position for data acquisition of multiple echoes (usually 2-6 echoes), R2, R2’ and R2* information can be obtained according to the single exponential model. In fact, in order to obtain more accurate R2, R2' and R2* information, the 180° pulse usually needs to move several times to obtain more echo data. This patent proposes a multi-echo asymmetric spin-echo sequence that periodically shifts the pulse position by 180°. Taking the cycle of 3TR and four EPI echoes as an example, the schematic diagram of the sequence is shown in Figure 2. In the psMASE sequence, the 180° pulses do not move linearly but periodically.

For each subsequence that determines the 180° pulse position, multiple EPI echoes are located on both sides of the spin echo position to acquire imaging data, and the echo time (TE _i ) is defined as the distance between the center of the 90° excitation pulse and the echo k-space center interval. When the first EPI echo data acquisition is completed, the phase encoding is reversed, and the next three echo data are sequentially acquired in the same way. The time interval between every two adjacent echoes is defined as ΔTE. Since different echo positions have different deviations from the spin echo positions, corresponding signals acquired at different echo positions have different T2* weights. According to the single exponential decay model, the signal strength function collected at different times from the 90° pulse is as follows:

Among them, S(t) represents the signal obtained at the acquisition time t, S ₀ is the magnetization obtained by excitation under the 90° pulse, and TE _SE is the effective spin echo time interval. Formula [1] can be simplified into the following form:

Taking the natural logarithm on both sides of formula [2], it can be further simplified as:

Specifically, in the psMASE sequence, the time interval between 90° and 180° pulses can be flexibly set, the time interval between the center of the 180° pulse and TE _1/2 is defined as τ, and the echo is maintained by changing τ The time TE _i is constant, the signal obtained by each echo acquisition has the same R2 weight but the R2' has different weights, and all R2' can be effectively estimated. More specifically, taking the period of 3TR and four EPI echoes as an example, as shown in step 1 of Fig. 1, τ uses three sizes of -ΔTE/2, 0 and ΔTE/2 in one period to maximize the use of Data acquisition is performed at the position of the spin echo to improve the signal-to-noise ratio (SNR). As shown in step 3 of Figure 1, within a period, the size of R2, R2' and R2* of each voxel can be fitted by the following formula:

Among them, S _3n+1 (TE _i ), S _3n+2 (TE _i ) and S _3n+3 (TE _i ) represent the signal strength collected at different echo times TE _i under different τ conditions, where TE ₁ It can be between 20-80 ms, TE _SE1 , TE _SE2 and TE _SE3 represent the effective spin echo time under different τ cases.

In addition, in order to further improve the time resolution, a strategy of moving window estimation is introduced into the estimation of R2, R2' and R2*, as shown in step 4 of Figure 1. Taking the period as 3TR as an example, the measurement method of the motion estimation strategy is shown in Figure 3, and the size of the moving window is set to be consistent with the period size of the psMASE sequence. In each TR time, for each imaging layer, four images collected at four echo positions form an image group. The size of R2, R2' and R2* at each time point can be obtained by fitting the image group at the corresponding time point and the two adjacent image groups at the left and right time points according to the principle of minimum mean square error and formula [4]. As shown in step 2 of Figure 1, in order to improve the signal-to-noise ratio, all images are pre-processed with a Gaussian filter (kernel operator size = 3 × 3; standard deviation = 1.5).

Compared with GESSE or GESFIDE sequence, this method has three advantages in simultaneously obtaining R2 and R2* information. First, the ASE sequence can adjust the weight of R2' while keeping the weight of R2 unchanged, so that R2' can be directly fitted. Based on the multi-echo ASE method, R2, R2' and R2* information can be obtained simultaneously. Second, in the ASE sequence, the number of echoes is no longer limited by the TE time and readout bandwidth, and the position of the 180° refocusing pulse can be flexibly set. Third, the acquisition strategy of EPI can better combine the flow rate attenuation gradient to reduce the contribution of intravascular signals without introducing additional motion artifacts.

Compared with the SAGE sequence, this method avoids the problem of layer profile mismatch caused by acquiring images before and after the 180° refocusing pulse in the SAGE sequence because the magnetization vector is acquired during the spin echo formation and decay process.

In general, the sequence design of psMASE makes the dynamic imaging of R2, R2' and R2* possible, and at the same time, the introduction of the moving window estimation strategy can further improve the temporal resolution of dynamic imaging.

Description of drawings

Fig. 1 is a flowchart of a method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images (taking the cycle as 3TR as an example).

Figure 2 is a diagram of psMASE sequence design (take the cycle of 3TR and four EPI echoes as an example). .

Fig. 3 is a schematic diagram of a moving window estimation strategy (taking the period as 3TR as an example).

Figure 4 is a schematic diagram of the dynamic measurement results of R2, R2* and R2' under a typical lower limb ischemia and recovery model of a volunteer.

detailed description

The psMASE sequence scan should be performed after the routine positioning scan and reference image scan. During the use, the subject lies flat on the MR scan table. Taking the imaging of lower extremity calf muscles as an example, a suitable psMASE sequence parameter can be set as follows: fieldofview(FOV)=150×150mm ² , matrixsize=70×70, repetitiontime(TR)=2000ms,TE ₁ /TE ₂ /TE ₃ /TE ₄ =60/80/100/120ms, echospace=20ms, τ=-10/0/10ms, slicethickness=6mm, SENSEfactor=2, NSA=1, echoshiftnumber=12. The above parameters can be adjusted according to different requirements.

In order to reflect the characteristics of this method that can dynamically measure tissue R2, R2* and R2' parameter images, take a volunteer in a typical lower limb ischemia and recovery experiment as an example. The belt is tied to the thigh of the right leg and fixed with a nylon strap to prevent the blood pressure belt from loosening. During the scan, keep the subject's head, feet, and knees at the same level. First, a resting state 90s scan was performed, and then we manually inflated the blood pressure belt to a pressure of 200mmHg, and the entire inflation process was completed within 5 seconds. After maintaining the inflation pressure for 8 minutes, we quickly released the tourniquet. During the whole process, the psMASE sequence was scanned (acquisition with a period of 3TR and four EPI echoes), and the change curves of gastrocnemius R2, R2* and R2' obtained by this method are shown in Figure 4.

Claims (11)

1. A magnetic resonance imaging method for noninvasive dynamic measurement of tissue R2, R2* and R2' parameter images, characterized in that, by using an asymmetric spin echo (ASE) magnetic resonance imaging sequence, a periodical movement of 180 degrees of refocusing is adopted The strategy of radio frequency pulse position is used to obtain multi-echo images with high time resolution; and the moving window estimation method is used to quantitatively estimate the dynamic R2, R2* and R2' parameter images of the tissue. 2. A kind of non-invasive dynamic measurement tissue R2 according to claim 1, the magnetic resonance imaging method of R2 * and R2 ' parameter image, is characterized in that, the sequence cycle time that adopts is 2-6 repetition time (TR). 3. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that the sequence of 180-degree refocusing pulses used is symmetrically distributed at half the echo time position (TE/2). 4. A kind of magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that, the adopted sequence collects 3-5 planar echoes in each cycle (EPI) signal for imaging. 5. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that, for the acquired image, R2, R2 are calculated according to the single exponential model of signal attenuation * and R2' parameter plots. 6. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that the imaging process can be single-layer scanning or multi-layer simultaneous scanning. 7. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that axial, coronal, sagittal or any oblique scan can be selected . 8. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that the method for obtaining tissue R2, R2* and R2' values is from magnetic resonance Select a region of interest (ROI) in the image. 9. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, characterized in that the method for obtaining tissue R2, R2* and R2' values is from magnetic resonance Select the entire tissue in the image. 10. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 1, wherein the imaging area can be muscles, kidneys, liver, head and other positions. 11. A magnetic resonance imaging method for non-invasive dynamic measurement of tissue R2, R2* and R2' parameter images according to claim 5, characterized in that the dynamics of R2, R2* and R2' are simultaneously obtained using least squares estimation parameter image.