CN103714521B - Liver R2* figure measuring method based on inquiry table - Google Patents
Liver R2* figure measuring method based on inquiry table Download PDFInfo
- Publication number
- CN103714521B CN103714521B CN201310742267.3A CN201310742267A CN103714521B CN 103714521 B CN103714521 B CN 103714521B CN 201310742267 A CN201310742267 A CN 201310742267A CN 103714521 B CN103714521 B CN 103714521B
- Authority
- CN
- China
- Prior art keywords
- liver
- represent
- interpolation
- pixel
- formula
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Landscapes
- Magnetic Resonance Imaging Apparatus (AREA)
Abstract
A kind of liver R2* figure measuring method based on inquiry table: including: (1) obtains magnetic resonance liver image, draw the area-of-interest comprising liver；(2) confluent hypergeometric function to given receiving coil port number carries out spline interpolation, sets up and inquires about table accordingly；(3) to each pixel in liver area-of-interest, its gray scale and echo time are fitted in Single-Index Model first moment model under non-central Chi influence of noise, obtain the R2* figure comprising liver.The R2* figure of energy fast and accurate measurement liver of the present invention.
Description
Technical field
The invention belongs to the technical field that in magnetic resonance, transverse relaxation rate R2* figure is measured, being specifically related to one can be quick
And accurately measure the R2* figure of liver to measure the liver R2* figure measuring method based on inquiry table of hepatic iron concentration.
Background technology
If occurring in human body, too much deposition of iron easily causes serious change when the endocrine organ such as liver, heart, special
Not being for sickle anemia and patients with thalassemia, long-term treatment of blood transfusion can cause too much deposition of iron phenomenon.Due to
In human body, 70%～90% too much ferrum will be deposited in liver, generally using liver concentration of iron as a weight of iron content in reactant
Want index.
Compared to the direct method of liver tissue bioptic, hepatic iron concentration measuring method based on nuclear magnetic resonance R2* figure
The complication such as liver is hemorrhage can be prevented effectively from, and certainty of measurement will not be by sample size, position and liver ferrum uneven distribution etc.
Factor affects.
Measure liver concentration of iron by nuclear magnetic resonance parameter R2* at present, mainly have two kinds of methods: one is based on multiple little
Liver area-of-interest, the gray scale of pixel in area-of-interest is first averaged by the method, then by obtain average after
Gray value be fitted according to suitable curve model with the corresponding echo time, obtain representing the R2* value of liver.Due to liver
Ferrum is not to be evenly distributed in liver, and the selection of the size and location of area-of-interest exists artificial difference, past
Toward final measurement result can be caused inaccurate.
Another kind of method is R2* based on nuclear magnetic resonance figure, the method to each pixel in liver, by its gray scale with
Echo time is fitted to suitable curve model, obtains the R2* figure of whole liver.Owing to each pixel is fitted, therefore
Amount of calculation is huge, calculates more time-consuming, is particularly especially apparent in some curve models.
Researcher is had to propose to use noise correction first moment model to measure R2* value (Feng Y, He T, Gatehouse
PD, Li X, Harith Alam M, Pennell DJ, Chen W, Firmin DN. Improved MRI R2*
relaxometry of iron-loaded liver with noise correction. Magn Reson Med 2013;
70:1765-1774.), obtain the R2* estimated value of preferable accuracy and precision.Owing to the method only accounts for multiple little
Liver area-of-interest, is averaged signal in region and then carries out once fitting, and artificial at area-of-interest selects,
Sometimes the impact that R2* is measured by some focus or artifact can not be got rid of；In addition, confluent hypergeometric function in matching
Calculating occupies most of the time, carries out single matching relatively time-consuming but still fall within the scope that can accept, and if carried out
The Fitting Calculation liver R2* figure pixel-by-pixel, estimates to need the most tens of hours, limits its application in clinical practice.
Therefore, not enough for prior art, it is provided that one quickly and accurately liver R2* figure measuring method is existing to overcome
The deficiency of technology is the most necessary.
Summary of the invention
The present invention proposes a kind of liver R2* figure measuring method based on inquiry table, and the method can be prevented effectively from interested
In the improper R2* measurement error caused of area sampling and noise correction first moment model, confluent hypergeometric function calculates time-consuming
Problem, can obtain full liver R2* figure accurately in several minutes.
The above-mentioned purpose of the present invention is realized by techniques below means.
A kind of liver R2* figure measuring method based on inquiry table, in turn includes the following steps:
(1), gather magnetic resonance liver image, and it is emerging to draw the sense comprising liver on the magnetic resonance liver image obtained
Interest region；
(2), to known receiving coil port numberConfluent hypergeometric functionCarry out batten to insert
Value, and set up by interpolation knot and interpolating function coefficient is constituted one to one with interpolation subinterval inquiry table；
(3), the gray scale of each pixel and echo time are fitted to Single-Index Model under non-central Chi influence of noise
First moment model formula (I) obtains the R2* value corresponding with each pixel, is obtained by the R2* value of each pixel and comprise liver
R2* schemes；
Wherein, formula (I) is:
... (I);
In formula (I),Represent expectation,Represent observation signal value,Represent that the Gauss of each receiving coil passage makes an uproar
The standard deviation of sound,Represent double factorial (i.e.,Represent receiving coil port number；Table
Show the spline interpolation function of confluent hypergeometric function,Represent the echo time,RepresentTime muting true letter
Number value,Represent transverse relaxation rate；Due at muting image background regions signal, therefore the standard deviation in formula (I)Can be obtained by formula (II):
... (II).
Above-mentioned steps (1) specifically uses many echo gradient echo sequence to obtain magnetic resonance liver image.
Above-mentioned steps (2) specifically uses under not a node boundary condition cubic spline interpolation method to confluent hypergeometric function
Approximate, and table is inquired about in foundation accordingly.
In above-mentioned steps (2) interpolation knot to choose specifically chosen be non-equally spaced interpolation knot.
Preferably interpolation knot,；WhenTime interval select
It is 0.1, whenTime interval be chosen as 50, boundary condition selects the boundary condition of not a node.
Above-mentioned steps (3) specifically uses the noise correction first moment approximated confluent hypergeometric function based on inquiry table
The curve matching of model.
Present invention liver R2* figure measuring method based on inquiry table, the method can be prevented effectively from prior art interested
In the improper R2* measurement error caused of area sampling and noise correction first moment model, confluent hypergeometric function calculates time-consuming
Defect, can obtain the R2* figure of the most whole liver in several minutes.
Accompanying drawing explanation
The present invention is further illustrated to utilize accompanying drawing, but the content in accompanying drawing does not constitute any limit of the people to the present invention
System.
Fig. 1 is the schematic flow sheet of the inventive method；
Fig. 2 is the inquiry list structure schematic diagram that the cubic spline functions of confluent hypergeometric function is corresponding；
Fig. 3 is that the emulation data of different signal to noise ratio use the method for the present invention and use noise correction first moment model
M1The R2* that NCM method respectively obtains schemes and comparative result, comprises R2* value from 100 ~ 1000 s in emulating image-1；
Fig. 4 is method and the M of the emulation data use present invention of different signal to noise ratio and receiving coil port number1NCM method
Calculate the comparative result form of R2* figure time used respectively；
Fig. 5 is method and the M that two groups of true hepatic data use the present invention1The R2* that NCM method respectively obtains schemes and ratio
Relatively result；
Fig. 6 is method and the M that two groups of true hepatic data use the present invention1NCM method calculates the R2* figure time used respectively
Comparative result form.
Detailed description of the invention
The invention will be further described with the following Examples.
Embodiment 1.
A kind of liver R2* figure measuring method based on inquiry table, as it is shown in figure 1, in turn include the following steps:
(1), gather magnetic resonance liver image, and it is emerging to draw the sense comprising liver on the magnetic resonance liver image obtained
Interest region.
(2), to known receiving coil port numberConfluent hypergeometric functionCarry out batten to insert
Value, and set up by interpolation knot and interpolating function coefficient is constituted one to one with interpolation subinterval inquiry table, such as Fig. 2 institute
Show.
(3), the gray scale of each pixel and echo time are fitted to Single-Index Model under non-central Chi influence of noise
First moment model formula (I) obtains the R2* value corresponding with each pixel, is obtained by the R2* value of each pixel and comprise liver
R2* schemes.
Wherein, formula (I) is:
... (I);
In formula (I),Represent expectation,Represent observation signal value,Represent that the Gauss of each receiving coil passage makes an uproar
The standard deviation of sound,Represent double factorial (i.e.,Represent receiving coil port number；Table
Show the spline interpolation function of confluent hypergeometric function,Represent the echo time,RepresentTime muting true letter
Number value,Represent transverse relaxation rate.
Due at muting image background regions signal, therefore the standard deviation in formula (I)Formula can be passed through
(II) obtain:
... (II).
Wherein, step (1) can use many echo gradient echo sequence to obtain magnetic resonance liver image.
Wherein, step (3) specifically uses the noise correction single order approximated confluent hypergeometric function based on inquiry table
The curve matching of square model.
Present invention liver R2* figure measuring method based on inquiry table, the method can be prevented effectively from prior art interested
In the improper R2* measurement error caused of area sampling and noise correction first moment model, confluent hypergeometric function calculates time-consuming
Defect, can obtain the R2* figure of the most whole liver in several minutes.
Embodiment 2.
A kind of liver R2* figure measuring method based on inquiry table, as it is shown in figure 1, with human liver for measuring object, specifically
Comprise the steps.
Step 1, obtains 12 echo liver magnetic resonance images, draws out full liver area-of-interest along liver image edge.Its
Imaging parameters is set to: the echo time chooses 0.93 respectively, 2.27,3.61,4.95,6.29,7.63,8.97,10.30,
11.64,12.98,14.32 and 15.66 ms, the repetition time is 200 ms, and matrix size is 64 × 128, and flip angle is 20o, layer
Thickness is 10 mm, and receiving coil channel number is 8.
It should be noted that the method drawing full liver area-of-interest can use manual drawing, it would however also be possible to employ other
Mode is drawn, and automatically draws such as programme-control or is drawn by other equipment.
Step 2, to known receiving coil channel numberConfluent hypergeometric functionEnter
Row cubic spline interpolation, set up by interpolation knot and with the inquiry that interpolating function coefficient is constituted one to one of interpolation subinterval
Table.
The parameter that cubic spline interpolation is used is set to: interpolation knot,
；WhenTime interval be chosen as 0.1, whenTime interval be chosen as 50, boundary condition selects non-
The boundary condition of node.Herein cubic spline functions be calculated as general knowledge known in this field, do not repeat them here.
Step 3, the inquiry table obtained by step 2 replaces confluent hypergeometric function in noise correction first moment model of fit
Calculate, each pixel in the liver area-of-interest that step (1) is obtained, its gray scale and echo time are fitted to list
Exponential model obtains the R2* value of each pixel in the first moment model formula (I) under non-central Chi influence of noise, further according to often
The R2* value of individual pixel obtains the R2* figure comprising liver.
... (I);
In formula (I),Represent expectation,Represent observation signal value,Represent that the Gauss of each receiving coil passage makes an uproar
The standard deviation of sound,Represent double factorial (i.e.,Represent receiving coil port number,Table
Show the spline interpolation function of confluent hypergeometric function,Represent the echo time,RepresentTime muting true letter
Number value,Represent effect transverse relaxation rate.
Due at muting image background regions signal, standard deviation in formula (I)It is to be obtained by formula (II)
Arrive:
... (II).
Research shows, confluent hypergeometric function computationally intensive and also time-consuming, confluent hypergeometric function is used it by the present invention
Spline interpolation function replace, use Single-Index Model first moment model (formula (I)) under non-central Chi influence of noise carry out by
Pixel fit, experiment show carried model can obtain almost with the noise correction single order moments method directly calculating confluent hypergeometric function
(it is called for short M1NCM) the R2* figure of identical liver, and the time that calculates accelerates nearly two orders of magnitude.
First liver image is drawn out area-of-interest by the method for the present invention, then for given receiving coil passage
Number, carries out cubic spline interpolation in advance to confluent hypergeometric function, sets up the corresponding inquiry table quick approximation for this function
Calculate, in conjunction with the first moment model of fit i.e. Single-Index Model of the noise correction first moment mould under non-central Chi influence of noise
Type (formula (I)), accelerates the speed of curve matching, shortens the time calculating liver R2* figure.
In order to verify method (the hereinafter referred to as M of the present invention further1NCM-LUT method) effect, by M1NCM-LUT side
Method and the noise correction first moment directly calculating confluent hypergeometric function (are called for short M1NCM) method carries out experiment and compares, result
As follows:
Fig. 2 shows the inquiry list structure figure that confluent hypergeometric function that second step of the present invention is set up is corresponding.
Fig. 3 be receiving coil port number be 8, the emulation data that are respectively under 15,30 and 60 of signal to noise ratio use the present invention
Method and M1The R2* that NCM method obtains schemes and comparative result, it can be seen that for different signal to noise ratios, two kinds of methods obtain
Almost identical R2* figure, error is 10-5s-1The order of magnitude.
Fig. 4 show the emulation data under different signal to noise ratio and receiving coil port number use the method for the present invention with
M1NCM method calculates the comparative result table of R2* figure time used respectively.Table shows different signal to noise ratios (signal to noise ratio is 15,
30,60) make with the lower emulation data of different receiving coil port number (coil channel number is 1,2,4,8,16,32,64,128) combination
With method and the M of the present invention1The accuracy comparison diagram of two kinds of methods of NCM method, it can be seen that the method for the present invention accelerates 95
～418 times, R2* figure can be obtained within the time of a few minutes.
Fig. 5 is method and the M that two groups of true hepatic data use the present invention1The R2* that NCM method respectively obtains schemes and ratio
Relatively, one group is serious liver ferrum overload, and one group is slight liver ferrum overload.It can be seen that two kinds of methods obtain almost identical R2*
Figure, error is 10-4 s-1The order of magnitude.
Fig. 6 gives method and the M that 6 groups of true hepatic data use the present invention1The time of NCM method compares, the most right
Should transship by liver ferrum in various degree.It can be seen that the method for the present invention can accelerate 120～162 times equally.
From result above, present invention liver based on inquiry table R2* figure measuring method can obtain and M1NCM is identical
Liver R2* schemes, but improves nearly two orders of magnitude in speed, it is possible to obtain the R2* figure that a width is complete in several minutes.
It should be noted that the measuring method of present invention liver based on inquiry table R2* figure is applicable not only to human liver
The measurement of R2* figure, can equally be well applied to the measurement of other animal liverss R2* figure.
It is last it should be noted that, the present invention is only protected by above example in order to technical scheme to be described
The restriction of scope, although being explained in detail the present invention with reference to preferred embodiment, those of ordinary skill in the art should manage
Solve, technical scheme can be modified or equivalent, without deviating from technical solution of the present invention essence and
Scope.
Claims (1)
1. a liver R2* figure measuring method based on inquiry table, it is characterised in that: in turn include the following steps:
(1), gather magnetic resonance liver image, and on the magnetic resonance liver image obtained, draw the region of interest comprising liver
Territory；
(2), to known receiving coil port number NRCConfluent hypergeometric function1F1(-1/2；NRC；-z) carry out spline interpolation, and
Set up by interpolation knot and interpolating function coefficient is constituted one to one with interpolation subinterval inquiry table；
(3) each pixel in the liver area-of-interest, to step (1) obtained, during by the gray scale of each pixel with echo
Between be fitted to Single-Index Model first moment model formula (I) under non-central Chi influence of noise obtains corresponding with each pixel
R2* value, obtained the R2* comprising liver by the R2* value of each pixel and scheme；
Wherein, formula (I) is:
In formula (I), E () represents expectation, SMRepresent observation signal value, σgRepresent the mark of the Gaussian noise of each receiving coil passage
It is accurate poor,！！Represent double factorial, NRCRepresent receiving coil port number,Represent the spline interpolation function of confluent hypergeometric function,
TE represents echo time, S0Representing muting true signal value during TE=0, R2* represents transverse relaxation rate, due to without making an uproar
The image background regions signal S of sound0=0, therefore the standard deviation sigma in formula (I)gFormula (II) can be passed through obtain:
Described step (1) specifically uses many echo gradient echo sequence to obtain magnetic resonance liver image；
Described step (2) specifically uses cubic spline interpolation method under not a node boundary condition to carry out confluent hypergeometric function
Approximate, and table is inquired about in foundation accordingly；
In described step (2) interpolation knot to choose specifically chosen be non-equally spaced interpolation knot；
Interpolation knot 0≤zm≤zmax=50000, m=1,2 ..., M；As 0≤zmWhen≤1000, interval is chosen as 0.1, when 1000
＜ zmWhen≤50000, interval is chosen as 50, and boundary condition selects the boundary condition of not a node；
Described step (3) specifically uses the noise correction first moment model approximated confluent hypergeometric function based on inquiry table
Curve matching.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310742267.3A CN103714521B (en) | 2013-12-30 | 2013-12-30 | Liver R2* figure measuring method based on inquiry table |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310742267.3A CN103714521B (en) | 2013-12-30 | 2013-12-30 | Liver R2* figure measuring method based on inquiry table |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN103714521A CN103714521A (en) | 2014-04-09 |
| CN103714521B true CN103714521B (en) | 2016-08-17 |
Family
ID=50407465
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310742267.3A Active CN103714521B (en) | 2013-12-30 | 2013-12-30 | Liver R2* figure measuring method based on inquiry table |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN103714521B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110338832A (en) * | 2019-06-19 | 2019-10-18 | 上海全景医学影像诊断中心有限公司 | Simulation cerebral blood perfusion parameter imaging method based on 18F-FDG |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105608681B (en) * | 2016-02-14 | 2018-12-21 | 电子科技大学 | A kind of liver magnetic resonance R2*Parameter drawing drawing method |
| CN113240078B (en) * | 2021-04-26 | 2024-03-19 | 南方医科大学 | Magnetic resonance R2 based on deep learning network * Parameter quantization method, medium and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102439474A (en) * | 2009-03-25 | 2012-05-02 | 皇家飞利浦电子股份有限公司 | Quantification of intracellular and extracellular spio agents with r2 and r2* mapping |
| CN102737392A (en) * | 2012-06-07 | 2012-10-17 | 南方医科大学 | Non-partial regularization prior reconstruction method for low-dosage X-ray captive test (CT) image |
| CN103218788A (en) * | 2013-04-24 | 2013-07-24 | 南方医科大学 | Method for measuring liver magnetic resonance crosswise relaxation rate R2* parameter |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8824766B2 (en) * | 2012-02-10 | 2014-09-02 | Duke University | Systems and methods for automated magnetic resonance imaging |
-
2013
- 2013-12-30 CN CN201310742267.3A patent/CN103714521B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102439474A (en) * | 2009-03-25 | 2012-05-02 | 皇家飞利浦电子股份有限公司 | Quantification of intracellular and extracellular spio agents with r2 and r2* mapping |
| CN102737392A (en) * | 2012-06-07 | 2012-10-17 | 南方医科大学 | Non-partial regularization prior reconstruction method for low-dosage X-ray captive test (CT) image |
| CN103218788A (en) * | 2013-04-24 | 2013-07-24 | 南方医科大学 | Method for measuring liver magnetic resonance crosswise relaxation rate R2* parameter |
Non-Patent Citations (3)
| Title |
|---|
| Estimating Non-Gaussian Diffusion Model Parameters in the Presence of Physiological Noise and Rician Signal Bias;Anders Kristoffersen;《JOURNAL OF MAGNETIC RESONANCE IMAGING》;20111003;第35卷(第1期);第183页左侧第2段，第188页左侧第2段 * |
| Improved MRI R2* Relaxometry of Iron-Loaded Liver with Noise Correction;Yanqiu Feng 等;《Magnetic Resonance in Medicine》;20130820;第70卷(第6期);1765-1774 * |
| 磁共振 -R2*值无创评估铁过载肝脏铁含量的初步研究;胡粟 等;《临床放射学杂志》;20120520;第31卷(第5期);733-737 * |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110338832A (en) * | 2019-06-19 | 2019-10-18 | 上海全景医学影像诊断中心有限公司 | Simulation cerebral blood perfusion parameter imaging method based on 18F-FDG |
Also Published As
| Publication number | Publication date |
|---|---|
| CN103714521A (en) | 2014-04-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kellman et al. | Myocardial perfusion cardiovascular magnetic resonance: optimized dual sequence and reconstruction for quantification | |
| Uribe et al. | Whole‐heart cine MRI using real‐time respiratory self‐gating | |
| Jalnefjord et al. | Optimization of b‐value schemes for estimation of the diffusion coefficient and the perfusion fraction with segmented intravoxel incoherent motion model fitting | |
| Gras et al. | Diffusion‐weighted DESS protocol optimization for simultaneous mapping of the mean diffusivity, proton density and relaxation times at 3 Tesla | |
| CN101359040A (en) | Method and device for optimization of imaging parameters | |
| CN107271937A (en) | A kind of synchronous acquisition and calibration method of three-dimensional multi-parameter weighted magnetic resonance imaging | |
| US8386214B2 (en) | Method and system for simulation of an MR image | |
| CN111721795B (en) | Material measurement method and system based on nuclear magnetic resonance system | |
| US10290103B2 (en) | Method, device and non-transitory digital storage medium for non-aqueous tissue volume estimation | |
| CN103714521B (en) | Liver R2* figure measuring method based on inquiry table | |
| Gao et al. | Arterial spin labeling‐fast imaging with steady‐state free precession (ASL‐FISP): a rapid and quantitative perfusion technique for high‐field MRI | |
| Xie et al. | Functional evaluation of transplanted kidneys with reduced field-of-view diffusion-weighted imaging at 3T | |
| Donker et al. | Quantitative 1H-NMR imaging of water in white button mushrooms (Agaricus bisporus) | |
| JP2017517325A5 (en) | ||
| Hermann et al. | Magnetic resonance fingerprinting for simultaneous renal T1 and T2* mapping in a single breath‐hold | |
| CN105699923A (en) | Magnetic resonance imaging method for measuring R2, R2* and R2' parameter image of tissue in noninvasive and dynamic manner | |
| DE102014206398B4 (en) | Magnetic resonance imaging method for at least two separate radio-frequency transmitting coils with time-delayed slice-selective excitation pulses | |
| CN105997074A (en) | Magnetic resonance quantitative magnetic susceptibility imaging multi-echo phase fitting method | |
| Hilbert et al. | Comparison of turbo spin echo and echo planar imaging for intravoxel incoherent motion and diffusion tensor imaging of the kidney at 3 Tesla | |
| DE102019215046B4 (en) | Method for compensating for eddy currents when recording measurement data using magnetic resonance | |
| Echeverria‐Chasco et al. | Multiparametric renal magnetic resonance imaging: A reproducibility study in renal allografts with stable function | |
| CN106659420A (en) | Magnetic resonance imaging device | |
| CN105608681B (en) | A kind of liver magnetic resonance R2*Parameter drawing drawing method | |
| Lee et al. | Improved 3‐Tesla cardiac cine imaging using wideband | |
| DE19616403C2 (en) | Method for determining the time course of the basic field of a magnetic resonance imaging device under switched gradients |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |