CN101856220B - Quantitative optical molecular tomography device and reconstruction method - Google Patents

Abstract

Translated fromChinese

本发明公开了一种定量光学分子断层成像装置和定量重建方法，该分子成像装置包括生物发光断层成像数据采集平台、MicroCT系统、定量校准系统和定量重建计算机，生物发光断层成像数据采集平台用于捕获小动物体内的生物发光光源溢出体表的分布情况，MicroCT系统用于获得小动物体的解剖结构信息，在定量重建方法中包括基于视场的接收光源的定量方法以及基于有限元细分网格的重建方法。利用本发明的光学定量分子断层成像装置和定量的重建方法可以通过二维的小动物体表的光源分布和定量能量信息反演出小动物体内的三维光源分布和细胞数量的定量信息。

The invention discloses a quantitative optical molecular tomography imaging device and a quantitative reconstruction method. The molecular imaging device includes a bioluminescence tomography data acquisition platform, a MicroCT system, a quantitative calibration system and a quantitative reconstruction computer. The bioluminescence tomography data acquisition platform is used for Capture the distribution of the bioluminescent light source overflowing the body surface in the small animal body, the MicroCT system is used to obtain the anatomical structure information of the small animal body, and the quantitative reconstruction method includes the quantitative method of receiving the light source based on the field of view and the subdivision network based on finite elements grid reconstruction method. Using the optical quantitative molecular tomography device and the quantitative reconstruction method of the present invention, the three-dimensional light source distribution and the quantitative information of the cell number in the small animal body can be retrieved from the two-dimensional light source distribution and quantitative energy information on the small animal body surface.

Description

Quantitative optical molecular tomographic device and method for reconstructing

Technical field

The present invention relates to the optical image technology field, specially refer to the structure and the quantitative method for reconstructing of light source of quantitative molecular imaging system.

Background technology

Nearly ten years, (Bioluminescence imaging, BLI) technology obtained development fast, is widely used in biological study, has obtained significant scientific achievement in the bioluminescence imaging.Yet the bioluminescence imaging technique can only generate the two-dimensional plane picture, can not show the three-dimensional space position information at petty action object inner light source place intuitively, and significant limitation is arranged on effect of visualization.Aspect quantitative analysis, the bioluminescence imaging can only be carried out relative quantitative analysis, the meaning that does not have absolute quantitation, and the three-dimensional imaging technology of tomography just has the meaning of absolute quantitation, can be referring to R.Weissleder and M.J.Pittet, " Imaging in the era of molecular oncology; " Nature 452,580-589 (2008) .2004, professor Wang Ge of U.S. University of Iowa has proposed bioluminescence fault imaging (Bioluminescence tomography, BLT) theoretical and based on the reconstruction uniqueness theory of prior information, can be referring to G.Wang, Y.Li, and M.Jiang, " Uniqueness theorems in bioluminescence tomography; " Medical Physics 31,2289-2299 (2004). in recent years, bioluminescence fault imaging theory has obtained development more rapidly and has paid close attention to widely, but in quantitative study, refer more particularly to gross energy quantitatively and the work of the quantitative analysis aspect of cell quantity very few, be the difficult point problem of a needs solution; In the research that also rests on substantially aspect the quantitative algorithm for reconstructing with imitative body such as polyethylene, polypropylene and nylon and pure algorithm, seldom have quantitative algorithm for reconstructing and imaging system are combined closely, and obtain robust, accurately location and quantitative algorithm for reconstructing.

Aspect quantitative computed tomography (SPECT) system, the Wang Ge professor's of U.S. University of Iowa research team has researched and developed first in the world cover prototype system, but aspect quantitative analysis, only obtain a changeless calibration factor, can be referring to D.Kumar, W.Cong, F.Bohenkamp, T.Kakaday, P.Taft, L.Wang, G.McLennan, E.Hoffman, and G.Wang, the variation meeting of " Development of lung tissue phantoms for bioluminescent imaging, " 2004,687. visual field in actual small animal imaging produces significant change to the power of light, if analysis meeting is not to the very big influence of quantitative result generation in such variation, this method has been brought very big inconvenience for the calibration of multiple image simultaneously in addition.At present, there is no document and patent report both at home and abroad and change the research report that enters the camera lens different-energy with the visual field about the CCD camera lens.

For noctilcent cross sectional reconstruction method, only incorporate quantitative, the locating information of the unique definite light source of prior information ability during owing to bioluminescence light source cross sectional reconstruction, remain a problem demanding prompt solution but how to obtain enough prior informations.At present, generally acknowledge that effective prior information mainly contains: the spectral information of the anatomical structure in reconstructed object zone, optical characteristics, bioluminescence light source etc.On rebuilding, also can come the rough area of feasible solutions that marks off light source to strengthen the restriction of light source is improved reconstruction quality according to the light distribution on reconstructed object surface and the organizational structure information of target usually.But based on the light source reconstruction of area of feasible solutions priori, all be to keep area of feasible solutions changeless in the present document.In fact only can obtain very coarse primary light source feasible zone according to surperficial light distribution, this situation makes to separate and trends towards not uniqueness, other report then is very subjective chosen a more accurate area of feasible solutions and guarantee the uniqueness of rebuilding, and is to be difficult to realize during this petty action object inner light source in reality is rebuild.In the existing document of bioluminescence tomography technology, tend to unified grid segmentation is carried out in the overall goals zone in order to improve the reconstruction precision, thereby this precision raising is a cost to roll up computation burden, this there is article proposed adaptive grid divided method (referring to Y.Lv, T.Jie, C.Wenxiang, W.Ge, L.Jie, Y.Wei, and L.Hui, " A multilevel adaptive fimite element algorithm for bioluminescence tomography, " Optics Express 14 (2006) .) thus keeping to a certain degree reducing computation complexity in the precision optionally segmenting grid on the basis of coarse grid.Yet this segmentation only is the segmentation on how much, for the subdivision quality that keeps integral grid still needs a large amount of grids are segmented the degree that waste has also limited segmentation of calculating that causes.Present most of article report all is by adding the pathosis that the Tikonov regularization term reduces the reconstruction problem, document Y.Lu, X.Zhang, A.Douraghy, D.Stout, J.Tian, T.F.Chan, and A.F.Chatziioannou, " Source reconstruction for spectrally-resolved bioluminescence tomography with sparse a priori information; " Opt Express 17,8062-8080 (2009) also can improve reconstructed results with the total variation regularization term according to the sparse characteristic of light source.Yet these two kinds of regularization term all are unique selection in process of reconstruction, do not adjust according to the light source area of feasible solutions of rebuilding, and accurate description light source character are still had certain restricted.Iteration or optimization method have been used for carrying out the reconstruction of light source in a large number, and present algorithm is based on all that node in the numerical discretization method or discrete unit search for, and this way of search is extremely flexible, but also very responsive to noise simultaneously.

Summary of the invention

The objective of the invention is:, provide a kind of optical molecular imaging device and corresponding quantitatively method for reconstructing with better quantitatively effect at the problem that above-mentioned existing noctilcent cross sectional reconstruction method aspect exists.In quantitative method for reconstructing, propose a kind of brand-new based on hybrid adaptive finite element grid segmentation Strategy and adopt the Level Set Method of robust more to carry out the light source reconstruction of bioluminescence fault imaging.Effectively reduce the pathosis of light source reconstruction, improved reconstruction precision and speed.

For achieving the above object, the quantitative optical molecular imaging device that the present invention proposes, comprise bioluminescence fault imaging data acquisition platform, MicroCT system, quantitative calibration system and quantitative reconstructing computer, described bioluminescence fault imaging data acquisition platform is made of high-sensitive CCD camera, toy support and rotation platform.The high-sensitive CCD camera is used to catch the distribution situation that the intravital bioluminescence light source of toy overflows body surface, toy support fixedly petty action object bodily form in rotary course does not change, the toy support uses the lower polythene material of density to make, and avoids X ray is produced strong decay.Rotation platform is used to rotate measurand to obtain the surface source of light distributed intelligence of toy body surface different visual angles by the high-sensitive CCD camera; Described MicroCT system is made up of X-ray tube and flat panel detector, and it is that 5 microns to 35 microns, peak power are at 30-100 watt that the X ray emitter is selected little focal spot (being positioned at the position of light pipe X ray light-emitting window) size for use; The pixel size of the flat panel detector that adopts is between 10 μ m * 10 μ m to 50 μ m * 50 μ m pixel sizes, and the imaging region size is determined according to the toy size, generally adopts the detector image-forming zone of 5cm * 5cm to 15cm * 15cm size; The MicroCT system adopts the cone-beam backprojection algorithm to realize the three-dimension disclocation reconstruction, obtains the 3 D anatomy structure information of petty action object, and this 3 D anatomy structure information is identical with the actual size of toy; Described quantitative calibration system adopts the light-emitting window light source uniformity greater than 98% integrating sphere uniform source of light system, realizes the quantitative calibration to bioluminescence fault imaging data acquisition platform; Quantitatively reconstructing computer is realized the quantitative reconstruction of collection, control and the light source of optics and MicroCT system data.

The present invention is merged MicroCT system and high-sensitive CCD camera to install and is constituted the bimodal system, and described bimodal system is bioluminescence computed tomography (SPECT) system and MicroCT system.The high-sensitive CCD camera of bioluminescence fault imaging data acquisition platform will make primary optical axis vertical mutually with the height of the formed cone-beam of emitter emission X ray of MicroCT system when installing, two dimension coordinates that obtain on the outline drawing that has spherical labels point that utilizes the high-sensitive CCD camera of bioluminescence fault imaging data acquisition platform to take add at the third dimension coordinate of primary optical axis direction by imaging formula calculating, three-dimensional coordinate is reverted to the actual size of object according to the image minification, three-dimensional coordinate coupling with the spherical labels point of MicroCT system, the bioluminescence view data that CCD is gathered is mapped on the body surface of toy, thereby gets to the end registration and mapping result.

The spherical labels point that on the toy support 4 diameters to be set be 0.5-1.0mm is so that directly read the centre coordinate of spherical labels point on the three-dimensional data of the MicroCT system that rebuilds.

For realize quantitatively, robust and accurate light source reconstruction algorithm, the present invention in quantitative method for reconstructing, proposes a kind of completely newly based on hybrid adaptive finite element grid segmentation Strategy and adopt the Level Set Method of robust more to carry out the light source reconstruction of bioluminescence fault imaging.The hybrid adaptive Finite Element Method has been considered among the position three of sizing grid h and interpolation basic function exponent number p and mesh point two or more factors simultaneously to the influence of reconstructed results, thereby has than traditional Finite Element Method and the higher convergence rate of h type self-adapting finite element method.Simultaneously, Level Set Method is by adopting the numerical value form to represent that the use of mode can increase the stability and the robustness of reconstructed results to ferret out.By on each grade grid, choosing new optimum feasible region, and adopt the combination of grid. policies such as self adaptation p segmentation, self adaptation h segmentation that the tetrahedron element in the new feasible region is segmented or adjusts, these measures have effectively reduced the pathosis of light source reconstruction, have improved reconstruction precision and speed.The invention provides a kind of quantitative algorithm for reconstructing has set up the visual field function light that enters the high-sensitive CCD camera has been carried out quantitatively; Utilize hp finite element grid segmentation Strategy that surface data is rebuild and obtain the intravital light source information of toy; The cell mass of the varying number by toy vivo carrying optical probe carries out the bioluminescence light source reconstruction and draws gross energy and verify, may further comprise the steps:

Step 1: utilize integrating sphere uniform source of light system to set up the one-to-one relationship of the gradation of image average of absolute irradiance value and high-sensitive CCD camera institute images acquired

A) measure the maximum noise of integrating sphere uniform source of light system and the sensitivity of high-sensitive CCD camera and determine the calibration system linear zone, image data and carry out the calibration of camera in linear zone is with precision and the reliability that guarantees to calibrate;

B) regulating integrating sphere uniform source of light system in the calibration system linear zone makes the light that enters equidistantly change n time between the minimum and maximum irradiance of irradiance intensity at linear zone on its light-emitting window surface, the value of n is not less than 10, and each irradiance value of n the variation that integrating sphere uniform source of light system reads and the gradation of image average that the high-sensitive CCD camera is caught all have unique corresponding relation;

C) when calibration high-sensitive CCD camera captured images the visual field will be in the scope of integrating sphere uniform source of light light-emitting window disc, and keep the light-emitting window center to overlap with the primary optical axis of high-sensitive CCD camera;

Step 2: bimodal data registration

Utilize bioluminescence fault imaging data acquisition platform to gather the bioluminescence signal of a plurality of visual angles toy body surface under the airtight fully dark situation and the outline drawing picture of petty action object respectively, obtain the spherical labels dot information on the toy support simultaneously, and can read the coordinate of spherical labels point; Gather the image of MicroCT system, and carry out the three-dimensional reconstruction of MicroCT system view data, can on the volume data after the reconstruction, read the three-dimensional coordinate of spherical labels point simultaneously; Two dimension coordinates that obtain on the outline drawing that has spherical labels point that utilizes the high-sensitive CCD camera of bioluminescence fault imaging data acquisition platform to take, and at the basic image-forming principle 1/u+1/v=1/f of primary optical axis direction utilization (wherein u represents object distance, v represents image distance, f represents focal length) can obtain the third dimension coordinate of outline drawing picture, three-dimensional coordinate is reverted to the actual size of object according to the image minification, with the spherical labels point data coupling of MicroCT system, the data reconstruction of MicroCT system is cut apart the 3 D anatomy structure information that obtains each organ;

Step 3: set up the quantitative visual field function of toy body surface light source, the light that enters the high-sensitive CCD camera is carried out quantitatively

Is the bottom surface by a visual angle of toy surface source of light center of distribution to the formed circumferential plane in high-sensitive CCD camera lens front end face edge, with the surface source of light center of distribution is the centre of sphere, is that the spherical crown that radius is intercepted on camera lens is the viewing area with this centre of sphere to the length of high-sensitive CCD camera lens edge line segment; Foundation is by the visual field function of the energy size that enters the high-sensitive CCD camera lens that the visual field determined:

$E=\frac{R_0(R-d)}{R(R_0-d_0)}---\left(1\right)$

(1) in the formula, R₀, when R represents with the system calibration of integrating sphere uniform source of light respectively the light-emitting window centre of surface and when measuring the distribution of light sources center of toy body surface to the distance at high-sensitive CCD camera lens front end face edge, d₀, when d represents with the system calibration of integrating sphere uniform source of light respectively the light-emitting window centre of surface and when measuring the distribution of light sources center of toy body surface to the distance at high-sensitive CCD camera lens front end face center, E represents to enter the unit energy of high-sensitive CCD camera lens, unit energy nW/mm in the bioluminescence computed tomography (SPECT) system²Represent;

Step 4: the definite and initial mesh subdivision of petty action object inner light source feasible region

The optical parametric of each organic region in the anatomical structure data that determiningstep 2 obtains carries out the initial mesh subdivision to finding the solution the zone, and obtains the energy density of surperficial node, distributes according to surface energy, delimits priori light source feasible region;

Step 5: set up the linear relationship between the inner unknown light source and surperficial known measurements on the k level grid based on hp mixed finite element grid segmentation Strategy;

AS_k＝Φ_k (2)

S wherein_kBe the unknown light source in inside on the k level grid, Φ_kBe the surperficial light distribution on the k level grid, A is S_kAnd Φ_kCoefficient matrix;

Step 6: the optimization objective function on the k level grid is found the solution

$\mathrm{\Θ}\left(S_k^{\mathrm{per}}\right)={\left|\right|A_k{S}_k-{\mathrm{\Φ}}_k^m\left|\right|}_{L^2\left(\mathrm{\Ω}\right)}+{\mathrm{\λ}}_k{\mathrm{\η}}_k\left(S_k\right)---\left(3\right)$

(3) in the formula

It is the zone boundary

The measured value of last pharosage, λ_kThe expression regularization parameter, η_kThe expression penalty; Utilize Level Set Method to find the solution this object function, obtain the distribution of light sources on the k level grid, L²(Ω) the subscript implication is the L2 norm of finding the solution on the territory;

Step 7: judge and satisfy the precision conditions of finding the solution

$\left|\right|{\mathrm{\&Phi;}}_k^m-{\mathrm{\&Phi;}}_k^c\left|\right|<{\mathrm{\&epsiv;}}_{\mathrm{\&Phi;}},k>L_{\max }---\left(4\right)$

If satisfy then stop, obtaining distribution of light sources S_k, otherwise change step 8, in (4) formula

Be value of calculation, L_MaxIt is maximum grid segmentation number of times;

Step 8: determine k+1 level light source feasible region and grid segmentation Strategy

According to the distribution of light sources S that reconstructs on the k level grid_kDetermine light source feasible region new on the k+1 level grid and grid segmentation Strategy;

Step 9: change step 5, dwindle the feasible region, make k increase progressively 1, rebuild, obtain the gross energy of light source reconstruction at the enterprising line light source of k+1 level grid.

The quantitative method for reconstructing of molecular imaging of the present invention is rebuild the gross energy that obtains light source to the equivalent light source that cell mass produced that varying number in the petty action object carries optical probe, with the relation of the different gross energies of demarcation reconstruction, comprise following steps with different cell quantities:

Step [1]: the cell mass that carries the optical probe varying number is carried out light source reconstruction, draw the gross energy of reconstruction;

Step [2]: the light energy that the quantity of pair cell and cell are produced is rebuild the gross energy that obtains and is carried out linear fit, demarcates the gross energy of reconstruction light source and the relation of cell quantity; The selection of cell quantity is carried the different of optical probe luminous intensity according to the difference of cell mass degree of depth in the petty action object with cell and difference to some extent in each cell mass, and the data of degree of depth cell mass in 1cm the time are usually 1 * 10⁴To 1 * 10⁷Quantity, the light source model quantity that the cell mass of each quantity makes up is no less than 5, the light source gross energy that same quantity cell is rebuild carries out statistical analysis, to results averaged, and analysis deviation; The gross energy w that the light source reconstruction that different cell quantity constituted is obtained_nWith pairing cell quantity c_nCarry out linear fit:

c_n＝a·w_n+b (5)

In the formula (2), a and b are the cell quantity C of the cell mass of formation light source_nThe gross energy w that obtains with its reconstruction_nThe linear fit coefficient;

Step [3]: the result to match does the data accuracy checking

5 cell masses to known same cell quantity carry out light source reconstruction, obtain gross energy and utilize resulting The Fitting Calculation formula (5) in the step [2] to calculate the total quantity of cell according to the meansigma methods of rebuilding the light source gross energy, the deviation that compares result of calculation and known results is if value of calculation is at e₀Then satisfy accuracy requirement in (1 ± 5%) scope, otherwise forward step [2] to again to many groups quantity, a plurality of cell mass gross energy and the cell quantity of rebuilding light source carries out linear fit on the same group.

Compared with prior art, beneficial effect of the present invention is:

Quantitative optical molecular tomographic device of the present invention made up bioluminescence fault imaging data acquisition platform under same system, merge and the MicroCT system of registration and to the quantitative calibration system of bioluminescence fault imaging data acquisition platform.This bimodal system can guarantee that toy bodily form in imaging process does not change, and can realize that gross energy quantitatively and the cell quantity in the living small animal quantitative (only obtaining the average information of petty action object) by quantitative calibration system.

The present invention also has tangible beneficial effect for quantitative optical molecular imaging reconstruction method: for quality that improves light source reconstruction and the pathosis that reduces light source reconstruction, the present invention proposes a kind of light source reconstruction algorithm based on hybrid adaptive Finite Element Method and level set.This method effectively reduces the pathosis of bioluminescence fault imaging problem, has improved the quality of light source reconstruction.By mixed self-adapting grid segmentation Strategy, when improving reconstruction light source space resolution, reduce the time of rebuilding cost; The pathosis that strategy has further reduced light source reconstruction is chosen in posterior optimum feasible region, has improved the precision of rebuilding.Represent the distribution of unknown light source by level set function, strengthened robustness, stability and the anti-noise ability of light source reconstruction.

Other characteristics of the present invention and advantage can be in conjunction with the accompanying drawings become clearer from the description below by the preferred implementation that for example principle of the present invention is made an explanation.

Description of drawings

The composition frame chart of Fig. 1 quantitative optical molecular tomographic device

Fig. 2 quantitative optical molecular tomographic device constitutes sketch map

Fig. 3 quantitative optical molecular tomographic method for reconstructing flow chart

Fig. 4 integrating sphere uniform source of light system visual field quantitative relationship sketch map

The gross energy of Fig. 5 A PC3 prostate gland cancer cell three-dimensional reconstruction and the relation of cell quantity

The gross energy of Fig. 5 B A549 lung carcinoma cell three-dimensional reconstruction and the relation of cell quantity

The specific embodiment

Describe each related detailed problem in the technical solution of the present invention in detail below in conjunction with accompanying drawing.Described embodiment only is intended to be convenient to the understanding of the present invention.

Fig. 1 is the main composition part in the quantitative optical molecular tomographic device of setting forth in according to the present invention.

As shown in Figure 1, quantitative optical molecular tomographic device is mainly by bioluminescence fault imaging data acquisition platform and MicroCT system, and quantitatively calibration system and quantitative reconstructing computer constitute.Wherein bioluminescence fault imaging data acquisition platform is realized bioluminescence light source in the petty action object is overflowed data acquisition in the body surface distributed intelligence; The MicroCT system realizes the visual of petty action object 3 D anatomy structure, and the resulting petty action object of light source and the MicroCT system 3 D anatomy structure after the data three-dimensional that while bioluminescence fault imaging data acquisition platform is gathered is rebuild carries out registration and merges; Quantitatively calibration system realizes the quantitative calibration to bioluminescence fault imaging data acquisition platform; Quantitatively reconstructing computer is realized the quantitative reconstruction of data acquisition, control and the bioluminescence light source of optics and MicroCT system.

Quantitative optical molecular tomographic device shown in Figure 2 constitutes sketch map and has introduced the formation and the basic electric connecting relation of device and bioluminescence fault imaging data acquisition platform 1, MicroCT system 2, the quantitatively calibration system 3 and the quantitative installation site of reconstructing computer 4 have been described.Wherein bioluminescence fault imaging data acquisition platform 1 is connected by high-sensitive CCD camera power supply cable between high-sensitive CCD camera 11 and the high-sensitive CCD camera power supply case 12 mainly by high-sensitive CCD camera 11, high-sensitive CCD camera power supply case 12.High-sensitive CCD camera data cable is connected between high-sensitive CCD camera 11 and the quantitative reconstructing computer 4, connect by serial ports control cable between electric translation and turntable 13, electric translation and turntable control chamber 14, electric translation and turntable 13 and electric translation and the turntable control chamber 14, connect with USB control cable between electric translation and turntable control chamber 14 and the quantitative reconstructing computer 4, bioluminescence fault imaging data acquisition platform 1 also comprises the toy support 15 that is fixed on the turntable in addition.High-sensitive CCD camera 11 adopts Scientific Grade high sensitivity cameras, can catch faint bioluminescence light signal in the organism, at the wave-length coverage internal quantum efficiency of 450nm to 800nm up to more than 75%.MicroCT system 2 is mainly by little focal spot X ray emitter 21 of 35 microns, transmitter power 22, be connected by feed cable between little focal spot X ray emitter 21 and the transmitter power 22, transmitter power 22 with quantitatively be connected by power supply control cable between the reconstructing computer 4, MicroCT system 2 also comprises the flat panel detector 23 that receives X ray, the pixel size of this flat panel detector 23 is 50 μ m * 50 μ m, imaging region is of a size of 12cm * 12cm, quantitatively connect the data of gathering with the MicroCT data cable between reconstructing computer 4 and the flat panel detector 23 with transmission, MicroCT system 2 adopts the cone-beam backprojection algorithm that the object of scanning is rebuild, and the spatial resolution of MicroCT system 2 is 65 μ m.Quantitatively calibration system 3 mainly comprises integrating sphere uniform source of light system 31, this system traces American National Standard and (the National Institute for Standards and Technology of Institute for Research and Technology, the low light uniform source of light system of NIST) demarcating, the result who finally is used for scientific research can compare in the world.quantitatively also comprise integrating sphere uniform source of light system power supply and control chamber 32 in the calibration system 3, be connected between integrating sphere uniform source of light system 31 and integrating sphere uniform source of light system power supply and the control chamber 32 be feed cable, be connected in the control cable between integrating sphere uniform source of light system power supply and control chamber 32 and the quantitative reconstructing computer 4.Quantitatively reconstructing computer 4 is realized bioluminescence fault imaging data acquisition platform 1, MicroCT system 2 and the quantitatively control and the data acquisition of calibration system 3, main temperature control, data acquisition control to high-sensitive CCD camera 11, to the translation and the Spin Control of electric translation and turntable 15, to the control of the voltage and current of little focal spot X ray emitter 21 and to the control of integrating sphere uniform source of light system 31.In quantitative optical molecular tomographic device, the formed cone-beam of little focal spot X ray emitter 21 little focal spots projection X ray is high vertical mutually in the primary optical axis of bioluminescence fault imaging data acquisition platform 1 and the MicroCT system 2, can guarantee the data registration accuracy of subsequent optical data and MicroCT system like this.

Fig. 3 is the flow chart that quantitative optical molecular tomographic method for reconstructing of the present invention is quantitatively rebuild.Its process step comprises:

Step 301: quantitatively rebuild beginning

Step 302 and 303: to high-sensitive CCD camera calibrated and data acquisition

The light-emitting window plane of adjustment integrating sphere uniform source of light system 31 is 54mm to the distance of the camera lens front end face of high-sensitive CCD camera 11, with the complete diameter in integrating sphere uniform source of light system 31 in the zone that guarantees high-sensitive CCD camera 11 captured images is in the light-emitting window disc scope of 10cm, and keeps the light-emitting window center to overlap with the primary optical axis of high-sensitive CCD camera 11; Demarcate the one-to-one relationship of the pixel grey scale average of absolute irradiance value that integrating sphere uniform source of light system 31 records and institute's images acquired.

Measure the maximum noise of integrating sphere uniform source of light system and the sensitivity of high-sensitive CCD camera 11 and determine the calibration system linear zone, in organism illuminating source irradiance scope, integrating sphere is set 20 different absolute irradiance values, gather the image of at least three time of exposure under each absolute irradiance value, three time of exposure preferentially select to cover the time of exposure scope in the biotic experiment, and each time of exposure is gathered 5 width of cloth images down; And gather the background noise image of the CCD camera 11 of corresponding time of exposure.Utilize these images to demarcate the one-to-one relationship of absolute irradiance value and gradation of image average.

Set up the quantitative visual field function of toy body surface light source, the light that enters high-sensitive CCD camera 11 is carried out quantitatively

Shown in Figure 4 is integrating sphere uniform source of light system visual field quantitative relationship sketch map, and integrating sphere uniform source of light system light-emitting window centre of surface is R to the distance at high-sensitive CCD camera lens front end face edge₀, integrating sphere uniform source of light system light-emitting window centre of surface is d to the distance at high-sensitive CCD camera lens front end face center₀, d here₀Be 54mm; The visual field function is the bottom surface by toy surface source of light center of distribution to the formed circumferential plane of high-sensitive CCD camera lens end face edge, with the surface source of light center of distribution is the centre of sphere, is that the spherical crown that radius is intercepted on camera lens is the viewing area with this centre of sphere to the length of high-sensitive CCD camera lens edge line segment; Foundation is by the energy size that enters the high-sensitive CCD camera lens that the visual field determined:

$E=\frac{R_0(R-d)}{R(R_0-d_0)}$

In the formula, R₀, when R represents with the system calibration of integrating sphere uniform source of light respectively and when measurement the light-emitting window centre of surface or the distribution of light sources center of toy body surface to the distance at high-sensitive CCD camera lens front end face edge, d₀, when d represents with the system calibration of integrating sphere uniform source of light respectively and when measurement the light-emitting window centre of surface or the distribution of light sources center of toy body surface to the distance at high-sensitive CCD camera lens front end face center, E represents to enter the unit energy of high-sensitive CCD camera lens, and unit energy is generally used nW/mm in the bioluminescence computed tomography (SPECT) system²Represent.

Step 304: bimodal data registration

Utilize bioluminescence fault imaging data acquisition platform to gather the bioluminescence signal of a plurality of visual angles toy body surface and the outline drawing picture of petty action object respectively as previously mentioned, obtain the spherical labels dot information on the toy support when gathering the outline drawing picture of toy, and can read the coordinate of spherical labels point; Gather the image of MicroCT system, and carry out the three-dimensional reconstruction of MicroCT system view data, can on the volume data after the reconstruction, read the three-dimensional coordinate of spherical labels point simultaneously; (wherein u represents object distance to utilize the axial simplification imaging formula of high-sensitive CCD camera 11 key lights 1/u+1/v=1/f in the bioluminescence fault imaging data acquisition platform 1, v represents that image distance, f represent focal length) can obtain the third dimension coordinate of optics outline drawing picture, according to imaging formula with optical three-dimensional coordinate transformation to the toy body surface, and and the spherical labels point data of MicroCT system coupling, the view data on the optical CCD is mapped on the body surface of toy; The bioluminescence light source of toy body surface is R to the distance at high-sensitive CCD camera lens front end face edge during measurement, and the bioluminescence light source of toy body surface is d to the distance at high-sensitive CCD camera lens front end face center; Visual angle energy with when calibration is a benchmark, and the Energy Conversion the when visual field of visual field that can be when measuring during divided by calibration be multiply by calibration again arrives the energy value of high-sensitive CCD camera 11 when going out to measure.

Step 305 and 306: data pretreatment and quantitatively reconstruction

1, petty action object inner light source feasible region determines and the initial mesh subdivision

To the toy volume data of MicroCT system reconstructing, determine the light source feasible region and toy volume data body is carried out the initial mesh subdivision according to the bioluminescence data that multi-angle is measured.

Each organic region of cutting apart in the MicroCT system data in the anatomical structure data that obtain is determined optical parametric, carry out the initial mesh subdivision to finding the solution the zone, and obtain the energy density of surperficial node, distribute, delimit priori light source feasible region according to surface energy.

2, set up the linear relationship between the inner unknown light source and surperficial known measurements on the k level grid based on the hp mixed finite element method;

Set up the linear relationship AS between the inner unknown light source and surperficial known measurements on the k level grid with the hp mixed finite element method_k=Φ_k, S wherein_kBe the unknown light source in inside on the k level grid, Φ_kIt is the surperficial light distribution on the k level grid;

3, to the calculating formula of the optimization objective function on the k level grid and find the solution

In the formula

It is the zone boundary

The measured value of last pharosage, λ_kThe expression regularization parameter, η_kThe expression penalty; Utilize Level Set Method to find the solution this object function, obtain the distribution of light sources on the k level grid;

4, judge whether to satisfy the precision conditions of finding the solution

K＞L_MaxIf satisfy then stop,

To distribution of light sources S_k, otherwise forward next step (determining k+1 level light source feasible region and grid segmentation Strategy) to,

Be value of calculation, L_MaxIt is maximum grid segmentation number of times;

5, determine k+1 level light source feasible region and grid segmentation Strategy, according to the distribution of light sources S that reconstructs on the k level grid_kDetermine light source feasible region new on the k+1 level grid and grid segmentation Strategy.Change the step of front, with the hp mixed finite element method at the linear relationship AS that sets up on the k+1 level grid between inner unknown light source and the surperficial known measurements_K+1=Φ_K+1Rebuild at the enterprising line light source of k+1 level grid, the energy density of light source on the k level grid is carried out integration, obtain the gross energy of light source reconstruction, unit represents with nW.

Step 307: cell quantity is demarcated

Toy liver model to different cell quantities is rebuild, and obtains the gross energy of corresponding reconstructed.The gross energy w that the bioluminescence light source reconstruction that different cell quantity constituted is obtained_nWith pairing cell quantity c_nCarry out linear fit and obtain c_n=aw_n+ b, wherein a and b are for rebuilding the gross energy w that obtains_nThe linear fit coefficient; Thereby extrapolate the intravital cell quantity of toy by the gross energy of light source reconstruction.

Shown in Fig. 5 A (b) is the gross energy of PC3 prostate gland cancer cell three-dimensional reconstruction and the relation of cell quantity, the fitting result y=44402x+19491 of the gross energy of reconstruction and the data of cell, and the coefficient of determination of linear relationship match reaches 0.993; Also have good linear fit effect y=13170x+23571 at the gross energy of the A549 lung carcinoma cell three-dimensional reconstruction shown in Fig. 5 B (b) and the relation of cell quantity equally, the match coefficient of determination of linear relationship is 0.986.The 3 D anatomy structure of two kinds of cells three-dimensional reconstructions under varying number and gross energy and the sketch map of bioluminescence light source in Fig. 5 A (a) and Fig. 5 B (a), have been provided simultaneously; The cell that is adopted carries luciferase reporter gene, and luciferase can be luminous under the catalysis situation of substrate fluorescein.

Step 308: judge whether precision meets the demands

Cell mass to the varying number that carries optical probe carries out the bioluminescence light source reconstruction, draw the gross energy of reconstruction, the selection of cell quantity is carried the different of optical probe luminous intensity according to the difference of the degree of depth in the cell mass place petty action object with cell and difference to some extent in the cell mass, and the data of degree of depth cell mass in 1cm the time are usually 1 * 10⁴To 1 * 10⁷Quantity, the light source model quantity that the cell mass of each quantity makes up is no less than 5, the bioluminescence light source gross energy that same quantity cell is rebuild carries out statistical analysis, all the other results averaged, and analysis deviation; 5 cell masses to known same cell quantity carry out the bioluminescence light source reconstruction, obtain gross energy and calculate the total quantity of cell, relatively result of calculation and known results e according to the calculating formula of rebuilding bioluminescence light source gross energy and simulate₀Deviation, as value of calculation at e₀Then satisfy accuracy requirement in (1 ± 5%) scope and forward step 309 to and finish, carry out data acquisition again, registration, data pretreatment again, quantitatively rebuild and the demarcation of cell quantity otherwise forward step 303 to.

Step 309: finish

In the detection of biological cell or molecular method quantification, the cell quantity in the living small animal body (only being used to obtain average information) is that biological study presses for the important information of knowing accurately.

Claims (4)

1. quantitative optical molecular tomographic device, it is characterized in that: this device is by bioluminescence fault imaging data acquisition platform, the MicroCT system, quantitatively calibration system and quantitatively reconstructing computer formation, described bioluminescence fault imaging data acquisition platform is by the high-sensitive CCD camera, toy support and rotation platform constitute, the high-sensitive CCD camera is used to catch the distribution situation that the intravital bioluminescence light source of toy overflows body surface, toy support fixedly petty action object bodily form in rotary course does not change, and rotation platform is used to rotate measurand to obtain the bioluminescence distribution of light sources information of the toy body surface of toy body surface different visual angles by the high-sensitive CCD camera; Described MicroCT system is made up of X-ray tube and flat panel detector, realize the cross sectional reconstruction of various visual angles data for projection by the cone-beam backprojection algorithm, obtain the 3 D anatomy structure information of petty action object, this 3 D anatomy structure information is identical with the actual size of toy; Described quantitative calibration system is an integrating sphere uniform source of light system, realizes the quantitative calibration to bioluminescence fault imaging data acquisition platform; Quantitatively reconstructing computer is realized the quantitative reconstruction of the intravital bioluminescence light source of data acquisition, control and toy of optics and MicroCT system.

2. quantitative optical molecular tomographic device according to claim 1, it is characterized in that: MicroCT system and high-sensitive CCD camera merge the formation bimodal system of installing, wherein, the primary optical axis of high-sensitive CCD camera is vertical mutually with the height of the cone-beam of the X ray of the emitter projection of MicroCT system, the plane coordinates of reading on the toy outline drawing picture of high-sensitive CCD camera photographs adds the third dimension coordinate that primary optical axis direction imaging formula is determined, minification according to optical data transforms to actual size, with the registration results that obtains after institute of the MicroCT system image data coupling, guarantee the registration accuracy of subsequent optical data and institute of MicroCT system image data.

3. quantitative optical molecular tomographic device according to claim 1, it is characterized in that: the spherical labels point that on the toy support 4 diameters to be set be 0.5-1.0mm so that on the three-dimensional data of the MicroCT system that rebuilds the direct centre coordinate of mark-sense point.

4. use the quantitative method for reconstructing of quantitative optical molecular tomographic device realization optical molecular imaging for one kind, it is characterized in that: this quantitative method for reconstructing is set up the visual field function by quantitative calibration system the light that enters the high-sensitive CCD camera is carried out quantitatively; In quantitative reconstructing computer, utilize hp finite element grid segmentation Strategy that surface data is rebuild and obtain the intravital bioluminescence light source of toy gross energy; By bioluminescence fault imaging data acquisition platform and quantitatively reconstructing computer the equivalent bioluminescence light source that cell mass produced that varying number in the petty action object is carried optical probe rebuild and obtain its gross energy, with the relation of the different gross energies of demarcation reconstruction, and verify with different cell quantities:

[1] utilize hp finite element grid segmentation Strategy that surface data is rebuild the method that obtains the intravital bioluminescence light source of toy gross energy and comprise following steps:

Step 1: utilize integrating sphere uniform source of light system to set up the one-to-one relationship of the gradation of image average of absolute irradiance value and high-sensitive CCD camera institute images acquired

A) measure the maximum noise of integrating sphere uniform source of light system and the sensitivity of high-sensitive CCD camera and determine the calibration system linear zone, image data and carry out the calibration of camera in linear zone is with precision and the reliability that guarantees to calibrate;

B) regulating integrating sphere uniform source of light system in the calibration system linear zone makes and equidistantly changes n time between the minimum and maximum irradiance of absolute irradiance value at linear zone on light-emitting window surface, the value of n is not less than 10, and each absolute irradiance value of n the variation that integrating sphere uniform source of light system reads and the gradation of image average that the high-sensitive CCD camera is caught all have unique corresponding relation;

C) when calibration high-sensitive CCD camera captured images the visual field in the scope of integrating sphere uniform source of light system light-emitting window disc, and keep light-emitting window to be centered close on the primary optical axis of high-sensitive CCD camera;

Step 2: bimodal data registration

A) utilize bioluminescence fault imaging data acquisition platform to gather the outline drawing picture of petty action object under the bioluminescence signal of a plurality of visual angles toy body surface under the fully airtight dark situation and the light conditions respectively, obtain the spherical labels dot information on the toy support when gathering the outline drawing picture of toy, read the coordinate of spherical labels point;

B) gather the image of MicroCT system, and carry out the three-dimensional reconstruction of MicroCT system view data, on the data after the reconstruction, read the three-dimensional coordinate of spherical labels point;

C) two dimension coordinates that obtain on the outline drawing that has spherical labels point that utilizes the high-sensitive CCD camera of bioluminescence fault imaging data acquisition platform to take add at the third dimension coordinate of primary optical axis direction by imaging formula calculating, three-dimensional coordinate is reverted to the actual size of object according to the image minification, with the three-dimensional coordinate coupling of the spherical labels point of MicroCT system, the data of MicroCT system are cut apart the anatomical information that obtains each organ;

Step 3: the quantitative visual field function of setting up toy body surface light source, the light that enters the high-sensitive CCD camera is carried out quantitatively, center by the bioluminescence distribution of light sources of a visual angle body surface of toy is the bottom surface to the formed circumferential plane of high-sensitive CCD camera lens end face edge, with surface source of light distribution center is the centre of sphere, is radius with this centre of sphere to high-sensitive CCD camera lens edge line segment length, and the intercepting spherical crown is the viewing area on camera lens; Foundation is by the visual field function that enters high-sensitive CCD camera lens energy size that the visual field determined:

$E=\frac{R_0(R-d)}{R(R_0-d_0)}---\left(1\right)$

(1) in the formula, R₀, when R represents with the system calibration of integrating sphere uniform source of light respectively the light-emitting window centre of surface and when measuring the bioluminescence distribution of light sources center of toy body surface to the distance at high-sensitive CCD camera lens front end face edge, d₀, light-emitting window centre of surface and the bioluminescence distribution of light sources center of toy body surface is to the distance at high-sensitive CCD camera lens front end face center when measuring when d represents with the system calibration of integrating sphere uniform source of light respectively, E represents to enter the unit energy of high-sensitive CCD camera lens;

Step 4: bioluminescence light source feasible region determines and the initial mesh subdivision in the petty action object

The optical parametric of each organic region in the anatomical structure data that determining step 2 obtains carries out the initial mesh subdivision to finding the solution the zone, and obtains the energy density of surperficial node, distributes according to surface energy, delimits priori light source feasible region;

Step 5: set up the linear relationship between the inner unknown light source and surperficial known measurements on the k level grid based on hp mixed finite element grid segmentation Strategy;

Step 6: the optimum target on the k level grid is found the solution;

Step 7: determine to satisfy the precision conditions of finding the solution,, obtain distribution of light sources, otherwise forward step 8 to if satisfy then stop;

Step 8: determine k+1 level light source feasible region and grid segmentation Strategy

Determine light source feasible region new on the k+1 level grid and grid segmentation Strategy according to the distribution of light sources that reconstructs on the k level grid;

Step 9: change step 5, dwindle the feasible region, make k increase progressively 1, rebuild, obtain the gross energy of bioluminescence light source reconstruction at the enterprising line light source of k+1 level grid.

[2] the equivalent bioluminescence light source that cell mass produced of the varying number of toy vivo carrying optical probe is rebuild the gross energy that obtains light source,, comprises following steps to demarcate the relation of different reconstruction gross energy with different cell quantities:

Step [1]: the cell mass to the varying number that carries optical probe carries out the bioluminescence light source reconstruction, draws the gross energy of reconstruction;

Step [2]: the light energy that the quantity of pair cell and cell are produced is rebuild the gross energy that obtains and is carried out linear fit, demarcates the gross energy of reconstruction bioluminescence light source and the relation of cell quantity; The bioluminescence light source gross energy that same quantity cell is rebuild carries out statistical analysis, to results averaged, and analysis deviation; The gross energy w that the bioluminescence light source reconstruction that different cell quantity constituted is obtained_nWith pairing cell quantity c_nCarry out linear fit:

c_n＝a·w_n+b (2)

In the formula (2), a and b are the cell quantity C of the cell mass of formation bioluminescence light source_nThe gross energy w that obtains with its reconstruction_nThe linear fit coefficient;

Step [3]: the result to match does the data accuracy checking

The bioluminescence light source that cell mass produced to known same cell quantity carries out light source reconstruction, obtains gross energy and according to rebuilding bioluminescence light source gross energy w_nMeansigma methods utilize the resulting The Fitting Calculation formula of step [2] (2) to calculate the total quantity of cell, relatively result of calculation and known results e₀Deviation, as value of calculation at e₀Then satisfy accuracy requirement in (1 ± 5%) scope, otherwise forward step [2] to again to many groups quantity, a plurality of cell mass gross energy average and the cell quantity of rebuilding the bioluminescence light source carries out linear fit on the same group.

Images