CN105699923A

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

Info

Publication number: CN105699923A
Application number: CN201510576993.1A
Authority: CN
Inventors: 王成彦; 张锐; 王鹤; 金立新; 王霄英; 张珏; 方竞
Original assignee: Peking University
Current assignee: Peking University
Priority date: 2015-09-12
Filing date: 2015-09-12
Publication date: 2016-06-22

Abstract

The invention discloses a magnetic resonance imaging method which utilizes the fact that a tissue with different deoxygenated hemoglobin content has magnetic sensitive chatacteristics of different degrees, uses an asymmetric spin echo magnetic resonance imaging sequence, adopts a strategy of moving a 180-degree echo refocusing RF pulse position periodically, allows 180-degree echo refocusing pulses to distribute on both sides of a TE/2 time position in each period, and obtains a high-time-resolution multiple-echo image. Further, aiming at acquired multiple-echo signals, the magnetic resonance imaging method performs mobile estimation by taking length of a corresponding 180-degree echo refocusing pulse mobile period as a window width, adopts least square estimation, and obtains dynamic parameter images of R2, R2* and R2' simultaneously according to a signal attenuation index model. By adopting the magnetic resonance imaging method, the high-time-resolution dynamic R2, R2* and R2' parameter images can be obtained, and a dynamic curves in an arbitrary region of interest can be displayed.

Description

A kind of MR imaging method of noinvasive kinetic measurement tissue R2, R2* and R2` parametric image

Technical field

The invention belongs to nuclear magnetic resonance (MRI) technical field, specifically a set of intrinsic relaxation time R2, R2* and R2 based on mr imaging technique dynamically non-invasively quantitative measurement different tissues ' method。

Background technology

Magnetic susceptibility, as a kind of important contrast mechanism, has a wide range of applications in mr techniques。For example, Blood oxygen level-dependent (BOLD) technology is namely based on the imaging means that the paramagnetic properties of deoxyhemoglobin grows up, and has been widely applied to the functional evaluation at different condition (anoxia, high carbon acid and ischemia) of head, kidney and skeletal muscle at present。

In nuclear magnetic resonance, magnetic susceptibility effect can affect R2* parameter, and R2* is the free damping rate of tissue signal, it is common to use single exponent ring-down model carrys out matching。And R2* contains two compositions: (1) one is irreversible composition R2, (2) another be reversible R2 ', reflect dephasing degree in the pixel that magnetic susceptibility causes。In Single-Index Model, R2*=R2+R2 '。Typically, tissue magnetic susceptibility uses R2* rather than R2 ' to weigh, but the latter will not be subject to the impact of the change of the R2 caused by causes for pathological in fact。On the other hand, R2* and R2 is imaged on combined in dynamic magnetic sensitivity Perfusion weighted imaging (DSC-PWI) and functional mri (fMRI) research making for the hemodynamic data of tissue is quantitatively described。Such as, Prinster et al. utilizes BOLD technology, is used for the voxel distinguished containing big blood vessel and thin vessels as a kind of index than (Δ R2/ Δ R2*) using the relative change in relaxation time under hypoxia with hypercapnic conditions。Further, adopt R2 and the R2* image-forming information after injection paramagnetic contrast medium to be used for the mean size to cerebral vessels and quantitative information is provided。In sum, a kind of sequence that can simultaneously obtain R2 and R2* accurate quantitative analysis information is very necessary。

As a rule, it is possible to be respectively adopted many echoes spin-echo sequence and many echo gradient echo sequence to obtain R2 and R2* information。This method is except time longer shortcoming, and twice sweep respectively obtains the strategy of R2 and R2* may cause the estimation difference between bigger R2 and R2* due to the physiological change between twice sweep。Recently, it is thus proposed that a kind of gtadient echo sampling free damping echo (GESFIDE) sequence is used for obtaining R2 and R2* information simultaneously。In this sequence, utilize multiple gtadient echo to gather the free damping intensity of magnetization after 90 ° of excitation pulses and the spin magnetization intensity after 180 ° of refocusing pulse simultaneously。R2* can utilize the intensity of magnetization in free damping process to obtain, and R2 then can obtain by comparing the echo strength of 180 ° of symmetrical positions of pulse。A kind of be called gtadient echo sampling spin echo (GESSE) sequence similar with GESFIDE is also proposed to obtain R2 and R2* information simultaneously, in this sequence, utilizes multiple gtadient echo to gather spin echo simultaneously and is formed and the intensity of magnetization in attenuation process。But, both the above method all have employed traditional Descartes's K space acquisition strategy, causes that image taking speed is slower, it is impossible to meets the requirement of high time resolution in such as DSC-PWI or fMRI application。Recently, a kind of many echoes spin that can dynamically obtain R2 and R2* information and gtadient echo (SAGE) echo-planar imaging (EPI) sequence are proposed again。But, in this sequence, the imperfect choosing layer profile caused by short radio frequency (RF) burst length can cause the unmatched problem of image layer facial contour collected before and after 180 ° of refocusing pulse, and this problem can produce bigger estimation difference in the quantitative Analysis of follow-up R2 and R2*。

As it has been described above, existing method also exists the shortcoming that temporal resolution is low or estimation difference is bigger。So, one can simultaneously and accurately obtain R2, R2* and R2 ' dynamic approach of information is to be badly in need of。

Summary of the invention

In this patent, we have proposed a kind of periodicity and move many echoes asymmetric spin echo (psMASE) sequence of 180 ° of pulse positions。In this sequence, owing to being formed at spin echo and attenuation process carries out signals collecting, it is to avoid layer profile unmatched problem。Further, in order to improve the temporal resolution of dynamic imaging, a kind of Moving Window estimate strategy be introduced in R2, R2* and R2 ' quantitative Analysis in, the length of Moving Window is identical with the Cycle Length of psMASE sequence。Illustrating with the cycle for 3TR, concrete method flow is as shown in Figure 1。

For traditional many echoes ASE sequence, the position of 180 ° of refocusing pulse can set flexibly。In theory, 180 ° of pulse position determined can obtain R2, R2 according to Single-Index Model for the data acquisition of many echoes (usual 2-6 echo) ' and R2* information。It practice, in order to obtain R2, R2 more accurately ' and R2* information, 180 ° of pulses typically require mobile position several times to obtain more echo data。This patent proposes a kind of periodicity and moves the asymmetric spin-echo sequence of many echoes of 180 ° of pulse positions, illustrates with the cycle for 3TR and fours' EPI echo, and sequence diagram is as shown in Figure 2。In psMASE sequence, 180 ° of pulses are not linear movements but periodically mobile。

For each subsequence determining 180 ° of pulse positions, multiple EPI echoes are positioned at the both sides of spin echo position for gathering imaging data, echo time (TE_i) it is defined as the interval at 90 ° of excitation pulse centers and echo k-space center。After first EPI echo data collection completes, phase code is reversed, and ensuing three echo datas are adopted and gathered successively in a like fashion。Often the interval between adjacent two echoes is defined as Δ TE。Different echo positions is owing to having different deviations from spin echo position, and the signal that corresponding different echo positions collect has different T2* weights。According to single exponent ring-down model, the signal intensity function collected at 90 ° of pulse different times of distance is as follows:

Wherein, S (t) represents at the acquisition time t signal obtained, S₀It is excite the intensity of magnetization obtained, TE under 90 ° of pulses_SEIt it is effective spin echo time interval。Formula [1] can simplify and is expressed as following form:

Formula [2] both sides are taken natural logrithm simultaneously, it is possible to be further simplified as:

Specifically, in psMASE sequence, it is possible to set the interval between 90 ° and 180 ° of pulses, 180 ° of pulse centers and TE flexibly₁Interval between/2 is defined as τ, keeps echo time TE by changing τ_iConstant, the signal that each echo acquirement obtains has the weighted of identical R2 weight and R2 ', and all R2 ' can effectively be estimated。More particularly, illustrating with the cycle for 3TR and fours' EPI echo, as shown in the step 1 of Fig. 1, τ is employings-Δ TE/2 within a cycle, 0 and tri-sizes of Δ TE/2 at utmost utilizing spin echo position to carry out data acquisition to improve signal to noise ratio (snr)。As shown in the step 3 of Fig. 1, within a cycle, R2, the R2 of each voxel ' and R2* size can be obtained by formula below matching:

Wherein, S_3n+1(TE_i),S_3n+2(TE_i) and S_3n+3(TE_i) represent in different τ situations at different echo time TE_iThe signal intensity collected, wherein TE₁Can between 20-80ms, TE_SE1,TE_SE2And TE_SE3Represent the effective spin echo time in different τ situation。

Additionally, in order to improve temporal resolution further, the strategy that a kind of Moving Window is estimated is introduced in R2, R2 ' and the estimation of R2* in the middle of, as shown in the step 4 of Fig. 1。Illustrating with the cycle for 3TR, the metering system of mobile estimation strategy is as it is shown on figure 3, being dimensioned to of Moving Window is in the same size with the cycle of psMASE sequence。Within each TR time, for each imaging layer, four images collected at four echo positions obtain an image sets。R2, the R2 of each time point ' and R2* size can be obtained according to Minimum Mean Square Error principle basis formula [4] matching by two adjacent image groups of the image sets of corresponding time point and time point around。As shown in the step 2 of Fig. 1, in order to improve signal to noise ratio, all images all utilize Gaussian filter to carry out pre-treatment (interior nuclear operator size=3 × 3；Standard deviation=1.5)。

This method is compared with GESSE or GESFIDE sequence, obtains R2 and R2* message context at the same time and has three advantages。First, ASE sequence can regulate and control the weight of R2 ' and keep the weight of R2 constant, and such R2 ' can directly obtain in matching, can obtain based on many echoes ASE method, R2, R2 ' and R2* information simultaneously。Second, in ASE sequence, echo number is no longer influenced by the TE time and reads the restriction of bandwidth, and the position of 180 ° of refocusing pulse can set flexibly。3rd, the acquisition strategies of EPI can, better in conjunction with flow rate attenuation gradient, be used for reducing the contribution of Ink vessel transfusing signal and will not introducing extra motion artifacts。

This method is compared with SAGE sequence, owing to being formed at spin echo and carrying out magnetization vector collection in attenuation process, it is to avoid gather the layer profile mismatch problem that image causes in SAGE sequence before and after 180 ° of refocusing pulse。

Generally speaking, the sequential design of psMASE makes R2, R2 ' and the dynamic imaging of R2* be possibly realized, meanwhile, Moving Window estimates that the introducing of strategy can improve the temporal resolution of dynamic imaging further。

Accompanying drawing explanation

Fig. 1 is the method flow diagram (illustrating for 3TR with the cycle) of noinvasive kinetic measurement tissue R2, R2* and R2' parametric image。

Fig. 2 is psMASE sequential design figure (illustrating for 3TR and fours' EPI echo with the cycle)。。

Fig. 3 is that Moving Window estimates strategy schematic diagram (illustrating with the cycle) for 3TR。

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

Detailed description of the invention

PsMASE sequence scanning should scan laggard row in the scanning of routine location and reference image, and in use, experimenter lies low on MR is scanning bed。Following primary minimum leg muscle is imaged as example, and a suitable psMASE sequential parameter can be provided that fieldofview (FOV)=150 × 150mm²,matrixsize=70×70,repetitiontime(TR)=2000ms,TE₁/TE₂/TE₃/TE₄=60/80/100/120ms, echospace=20ms, τ=-10/0/10ms, slicethickness=6mm, SENSEfactor=2, NSA=1, echoshiftnumber=12。Above parameter can be adjusted according to different demands。

Can kinetic measurement tissue R2 for embodying this method, the feature of R2* and R2 ' parametric image, at typical lower limb ischemia and experiment is recovered for a volunteer, experimenter faces upward and lies low on MR is scanning bed, use inflation sphygmomanometer cuff to tie up at right lower limb thigh place, use nylon tape to fix and prevent sphygmomanometer cuff from relaxing。In scanning process, keep the head of experimenter, foot and knee in same level。First carrying out tranquillization state 90s scanning, then sphygmomanometer cuff is manually inflated to 200mmHg pressure by us, and whole gas replenishment process completed in 5 seconds。After maintaining inflation pressure 8 minutes, we unclamp rapidly tourniquet。Carrying out the scanning (the employing cycle is 3TR and fours' EPI echo acquirements) of psMASE sequence in whole process, the change curve of gastrocnemius R2, R2* and the R2 that thus method obtains ' is as shown in Figure 4。

Claims

1. a noinvasive kinetic measurement tissue R2, the MR imaging method of R2* and R2 ' parametric image, it is characterized in that, by using asymmetric spin echo (ASE) MR imaging sequences, adopt the strategy of periodically mobile 180 degree times poly-radio-frequency pulse positions, obtain many echoes of high time resolution；And further by Moving Window method of estimation, quantitative predication goes out dynamic R2, R2* and the R2 of tissue ' parametric image。

2. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that the sequence period time of employing is 2-6 repetition time (TR)。

3. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that the sequence of employing 180 degree is returned poly-pulse and is symmetrically distributed in the position (TE/2) of echo time half。

4. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that the sequence of employing gathers 3-5 echo planar imaging (EPI) signal within each cycle and carries out imaging。

5. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that for the image obtained, calculate R2, R2* and R2 according to signal attenuation Single-Index Model ' Parameter Map。

6. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that imaging process can be Single Slice Mode or multilamellar scans simultaneously。

7. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that can chosen axis position, Coronal, sagittal plain or any loxosis scan。

8. a kind of noinvasive kinetic measurement tissue R2 according to claim 1, the MR imaging method of R2* and R2 ' parametric image, it is characterized in that, it is thus achieved that tissue R2, R2* and R2 ' method of value is one piece of area-of-interest (ROI) of selection from magnetic resonance image (MRI)。

9. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that obtain tissue R2, R2* and R2 ' method of value is to select monoblock tissue from magnetic resonance image (MRI)。

10. the MR imaging method of a kind of noinvasive kinetic measurement tissue R2, R2* and R2 according to claim 1 ' parametric image, it is characterised in that imaging region can be the positions such as muscle, kidney, liver, head。

11. the dynamic parameter image of a kind of noinvasive kinetic measurement tissue R2, R2* and the R2 according to claim 5 MR imaging method of the parametric image ', it is characterised in that use least-squares estimation obtains R2, R2* and R2 simultaneously '。