JPWO2019185457A5 - - Google Patents

Description

The following is described generally with respect to imaging, and in particular with respect to generating spectroscopic inflammation images from spectroscopic imaging data, with particular application to computed tomography imaging.

Inflammation is thought to be involved in a wide variety of diseases, including neurodegenerative, cardiovascular, and cancer. Thus, non-invasive methods of inflammation detection support early disease detection when the damage caused by the disease is still reversible, improved patient stratification, and response to therapeutic assessment. Imaging of inflamed tissue through nuclear medicine techniques such as proton emission tomography (PET) imaging of radiolabeled white blood cells and 18F-fluorodeoxyglucose (18F-FDG) provides insight into the inflammatory response . information can be obtained. However, this technique suffers from high radiation dose requirements, limited availability, high cost, limited spatial and temporal resolution, and the ability to distinguish between high metabolic healthy organs such as the heart and nearby inflammation . There are limitations due to limitations in the ability to

Imaging inflamed tissue using computed tomography (CT) provides information about structural tissue changes. These changes include edema, fluid accumulation in the extracellular space, contrast enhancement, endothelial rupture, and organ damage. However, CT is limited in its ability to characterize inflammation due to low contrast between imaging artifacts including healthy tissue and/or fat, inflammation , and radiation sclerosis . For example, FIG. 1 depicts a cardiac CT image 100 with a first region of interest (ROI) 102 and a second ROI 104 . A first region of interest 102 was found to have high inflammation and little or no fat concentration, a second region of interest 104 had little or no inflammation , It has been found to have a high fat concentration. FIG. 2 presents boxplots 200 for ROIs 102 and 104 .

In FIG. 2, a first axis 202 represents Hounsfield Units (HU) values and a second axis 204 represents ROI. FIG. 2 shows minimum 206 , maximum 208 and median 210 values for first ROI 102 and minimum 212 , maximum 214 and median 216 values for second ROI 104 . Boxplot 200 shows that although there may be a difference in median HU values 210 and 216 between the two ROIs 102 and 104, this difference is not statistically significant. As such, inflammation is commonly assessed through delayed enhancement of contrast-enhanced CT scans, which primarily delineate vascular permeability associated with increased inflammation . Disadvantageously, these changes are subjectively determined and difficult to quantify due to various imaging artifacts including e.g. difficult.

Aspects described herein address the problems referenced above and others.

Described herein are techniques for assessing inflammation by generating spectroscopic maps that delineate and quantify inflammation within spectroscopic volumetric image data (eg, CT).

In one aspect, a system includes a memory configured to store an inflammation map generator module. The system receives at least one of spectroscopic projection data or spectroscopic volumetric image data and uses 2-base decomposition to decompose at least one of the spectroscopic projection data or spectroscopic volumetric image data into spectral projection data or spectroscopic volumetric image data. generating a set of vectors for each base represented in at least one of the metric image data; calculating the density of each base in the voxel from the set of vectors for each base; Further including a processor configured to determine a concentration of at least one of fat or inflammation . The system also includes a display configured to display the determined concentration of at least one of fat or inflammation .

In another aspect, a method comprises the steps of: receiving at least one of spectroscopic projection data or spectroscopic volumetric image data; decomposing at least one of the spectroscopic projection data or spectroscopic volumetric image data using 2-base decomposition; and generating a set of vectors for each base represented in at least one of the spectroscopic projection data or the spectroscopic volumetric image data. The method comprises the steps of: calculating a density of each base within the voxel from the set of vectors for each base; determining a density of at least one of fat or inflammation within the voxel from the density of each base; and displaying the determined concentration of at least one of

In another aspect, a computer-readable storage medium is encoded with computer-readable instructions that, when executed by a processor of a computing system, cause the processor to generate at least one of the spectroscopic projection data or the spectroscopic volumetric image data. and decomposing at least one of the spectroscopic projection data or the spectroscopic volumetric image data using a two-base decomposition for each base represented in the at least one of the spectroscopic projection data or the spectroscopic volumetric image data. generating a set of vectors; The computer readable instructions, when executed by the processor , further cause the processor to calculate a concentration of each base within the voxel from the set of vectors for each base; determining the concentration of at least one; and displaying the determined concentration of at least one of fat or inflammation .

Those skilled in the art will appreciate further aspects of the present application upon reading and understanding the accompanying description.

The invention may take form in various components and arrangements of components, and in various steps and arrangements of steps. The drawings are for purposes of illustrating preferred embodiments only and are not to be construed as limiting the invention.

1 is an exemplary CT image of a heart with one region of interest containing primarily fat and another region of interest containing primarily inflammation ; FIG. 2 is an exemplary boxplot for each region of interest in the image shown in FIG. 1; FIG. 1 schematically illustrates a system including a computing system with an inflammation map generator and an imaging system; FIG. FIG. 2 schematically illustrates an example of an inflammation map generator; 1 is an exemplary CT image of a heart with one region of interest containing primarily fat and another region of interest containing primarily inflammation ; FIG. FIG. 11 schematically illustrates another example of an inflammation map generator; FIG. 10 is a vector separation plot; [0012] Figure 4 illustrates an exemplary method according to one embodiment described herein; FIG. 4 illustrates an exemplary method according to another embodiment described herein;

FIG. 3 schematically illustrates a system 300 including an imaging system 302, such as a CT scanner, configured for spectroscopic (multi-energy) imaging. Imaging system 302 includes a generally stationary gantry 304 and a rotating gantry 306 that is supported by stationary gantry 304 and rotates about an examination region 308 about a z-axis. A subject support 310 , such as a couch, supports an object or subject within the examination region 308 .

A radiation source 312 , such as an x-ray tube, is rotatably supported by and rotates with rotating gantry 306 to emit radiation across examination region 308 . In one example, radiation source 312 includes a single broad spectrum x-ray tube. In another example, radiation source 312 includes a single X-ray tube configured to switch between at least two different emission voltages (eg, 80 kVp and 140 kVp) during scanning. In yet another example, radiation source 312 includes two or more x-ray tubes configured to emit radiation having different average spectra. In yet another example, radiation source 312 includes combinations thereof.

Radiation sensitive detector array 314 includes a one- or two-dimensional array of rows of detector elements 316 , on the opposite side of radiation source 312 across examination region 308 , on an angular arc. A radiation sensitive detector array 314 detects radiation that traverses the examination region 308 and produces spectral projection data. If the radiation source 312 comprises a single broad-spectrum X-ray tube, the radiation sensitive detector array 314 may include energy resolving detectors (e.g., direct conversion photon counting detectors, at least two sets of photodiodes with different spectral sensitivities). (multilayer), etc.). With kVp switching and a multi-tube configuration, detector element 316 can alternatively be a non-energy resolving detector.

A reconstructor 318 receives the spectroscopic projection data and converts it into spectroscopic images, such as one or more of a photoelectric image, a Compton scatter image, an iodine image, a virtual non-contrast image, a bone image, a soft tissue image, and/or other base spectroscopic images. Reconstruct the volumetric image data. Reconstructor 318 may also generate non-spectral volumetric image data, for example, by combining spectral projection data to reconstruct combined spectral projection data or combining spectral volumetric image data.

Computing system 320 acts as an operator console. Console 320 includes human-readable output devices such as a display and input devices such as a keyboard and mouse. Software resident on console 320 allows an operator to interact with and/or operate scanner 302 via a graphical user interface (GUI) or otherwise. The console 320 further includes a processor 322 (e.g., microprocessor, controller, central processing unit (CPU), etc.) and computer readable storage media 324, including temporary media such as physical memory devices, but not non-transitory media. including.

Computer readable storage medium 324 includes instructions 326 for at least inflammation map generator 328 . In one variation, processor 322 and computer-readable storage medium 324 are part of another computing system that is separate and remote from (i.e., not part of) computing system 320. . In another variation, processor 322 additionally or alternatively executes one or more computer readable instructions carried by a carrier wave, signal, and/or other transitory medium.

Inflammation map generator 328 is configured to process at least the spectroscopic volumetric image data and/or spectroscopic projection data to generate an inflammation map. As will be described in more detail below, the inflammation map generator 328 generates at least spectroscopic projection data and/or spectroscopic data by quantifying contributions of tissue components (e.g., inflammation and/or fat) per pixel through at least the inflammation map. Quantify inflammation and/or fat in the volumetric image data. As such, the techniques described herein provide new spectroscopic maps (i.e., inflammation maps) that, in one example, improve non-invasive assessment of inflammation by CT, e.g., through improved delineation and quantification of inflammation . do.

FIG. 4 schematically shows an example inflammation map generator 328 .

Inflammation map generator 328 includes material decomposer 402 . A material decomposer 402 receives spectroscopic data (spectroscopic projection data and/or spectroscopic volumetric image data) and at least directly decomposes the received spectroscopic data into at least fat and inflammation . The material decomposer 402 outputs a set of vectors representing fat and a set of vectors representing inflammation , one fat vector and one inflammation vector per voxel.

ここで、Ｌｏｗ ｋｖは、低い方のエネルギー獲得についての分光データであり、Ｈｉｇｈ ｋｖは、高い方のエネルギー獲得についての分光データであり、

は、原点からのバイアスベクトルであり（例えば、バイアスがない場合は

）、

は、物質分解器４０２によって出力される脂肪を表すベクトルであり、ｃ_ｆａｔは、未知の脂肪濃度であり、

は、物質分解器４０２によって出力される炎症を表すベクトルであり、ｃ_ｉｎｆは、未知の炎症濃度である。一般に、式１は、ｃ_ｆａｔ及びｃ_ｉｎｆを求めるために同時に解くことができる線形方程式の系を表す。 Inflammation map generator 328 further includes a fat and inflammation density calculator 404 that determines fat and inflammation density within voxels based on the vectors. In one example, fat and inflammation concentration calculator 404 determines the concentration of fat and the concentration of inflammation as shown in Equation 1.

where Low kv is the spectroscopic data for the lower energy gain and High kv is the spectroscopic data for the higher energy gain;

is the bias vector from the origin (e.g., if there is no bias,

),

is a vector representing the fat output by the material decomposer 402, c _fat is the unknown fat concentration,

is a vector representing the inflammation output by the material decomposer 402 and c _inf is the unknown inflammation concentration. In general, Equation 1 represents a system of linear equations that can be solved simultaneously to find c _fat and c _inf .

Inflammation map generator 328 further includes a spectral inflammation map visualization module 406 . Spectral inflammation map visualization module 406 displays the quantified fat concentration and inflammation by visually enhancing (eg, coloring) each pixel based on its fat and/or inflammation concentration c _fat and c _inf . display the concentration. For example, in one example, pixels containing only fat are colored with a first or darker shading, pixels containing only inflammation are colored with a second or lighter shading, and fat and inflammation are colored. Pixels containing both and are colored with a shading of light and dark between them, the shading being closer to the first shade if the pixel contains more inflammation than fat and the pixel contains more inflammation than . If it contains a lot of fat, it approaches the second shade.

This is illustrated in FIG. 5 , which depicts a CT image 500 with a first ROI 502 , a second ROI 504 and a third ROI 506 . The first ROI 502 has a higher inflammatory concentration and as a result appears brighter than the second ROI 504 which has a higher fat concentration. The third ROI 506 contains both fat and inflammation , and as a result appears as a shade of gray between the gray values corresponding to the first ROI 502 and the second ROI 504 . In this example, first ROI 502 and second ROI 504 are located at substantially the same locations as first ROI 102 and second ROI 104 depicted in FIG. As shown in FIG. 1, the voxels within the first ROI 102 and the second ROI 104 are substantially the same shade of gray. As a result, whereas it is difficult to distinguish fat from inflammation in the image of FIG. I have.

FIG. 6 schematically shows another example of inflammation map generator 328 .

In this example, material decomposer 602 decomposes the input spectroscopic data into two base materials (eg, iodine and water), or components (eg, photoelectric effect and Compton scattering). The material decomposer 602 outputs a set of vectors representing one base and a set of vectors representing another base, one vector per voxel.

ここで、ＰｈｏｔｏＥは、光電効果分光データであり、ＣＳｃａｔｔｅｒは、コンプトン散乱分光データであり、

は、物質濃度計算器６０４によって生成される光電効果を表すベクトルであり、ｃ_{ＰｈｏｔｏＥ}は、光電効果についての未知の濃度であり、

は、物質濃度計算器６０５によって生成されるコンプトン散乱を表すベクトルであり、ｃ_{ＣＳｃａｔｔｅｒ}は、コンプトン散乱についての未知の濃度である。 Inflammation map generator 328 further includes substance concentration calculator 604 . For a given 2-base decomposition (eg, photoelectric effect and Compton scattering as shown below), the material concentration calculator 604 finds the concentration of each base for each voxel, eg, as shown in Equation 2.

Here, PhotoE is photoelectric effect spectroscopy data, CSscatter is Compton scattering spectroscopy data,

is a vector representing the photoelectric effect produced by the material concentration calculator 604, _cPhotoE is the unknown concentration for the photoelectric effect,

is a vector representing the Compton scatter produced by the substance concentration calculator 605, and _cCScatter is the unknown concentration for the Compton scatter.

Inflammation map generator 328 further includes fat and inflammation vector identifier 606 . A fat and inflammation vector identifier 606 identifies fat and inflammation vectors using a 2-base diagram. Fat and inflammation vectors can be learned from training data or from subjects using prior knowledge (eg, location of inflammation and healthy fat).

The inflammation map generator 328 also includes a fat and inflammation density calculator 608 . Fat and inflammation concentration calculator 608 conceptually uses vector separation techniques as shown in FIG. 7 to determine the concentration of fat and inflammation within a given voxel. In general, vector separation techniques are based on aligning the graphical location of fat and/or inflammation corresponding to a given voxel with the corresponding fat vector and/or inflammation vector to determine the fat and/or inflammation within the voxel. Determine the concentration of In FIG. 7, a first axis 702 represents a first base and a second axis 704 represents a second base (eg, photoelectric and Compton scattering).

FIG. 7 further shows a first vector 706 corresponding to inflammation and a second vector 708 corresponding to fat. The two vectors meet at a point 710 representing the bias of the two vectors from the origin. In this case the bias from the origin is zero (0). FIG. 7 also shows a first extrapolation 712 of the first vector 706 and a second extrapolation 714 of the second vector 708 . Based on the extrapolations 712 and 714, the concentration of fat and/or inflammation can be determined for a given voxel 716 with known values corresponding to the first and second bases 702 and 704. FIG.

ここで、

は、物質分解によって定義され、ピクセル値を与えられる２Ｄ座標系中でのボクセルの表現であり、軸として、

及び

を持つ２Ｄ座標で、

及び

によって定義される座標系中で

に等しく、

は、原点からのバイアスを表し、

は、脂肪及び炎症ベクトル識別器６０６によって識別された脂肪を表すベクトルであり、ｃ_ｆａｔは、脂肪の未知の濃度であり、

は、脂肪及び炎症ベクトル識別器６０６によって識別された炎症を表すベクトルであり、ｃ_ｉｎｆは、炎症の未知の濃度である。 In one example, the fat and inflammation concentration calculator 608 determines the fat concentration and the inflammation concentration using a vector separation technique, as shown in Equation 3.

here,

is a representation of a voxel in a 2D coordinate system defined by material decomposition and given pixel values, with

as well as

in 2D coordinates with

as well as

in a coordinate system defined by

equal to

represents the bias from the origin, and

is a vector representing the fat identified by the fat and inflammation vector identifier 606, c _fat is the unknown concentration of fat,

is the vector representing the inflammation identified by the fat and inflammation vector identifier 606, and _{c_inf} is the unknown concentration of inflammation .

Inflammation map generator 328 further includes a spectral inflammation map visualization module 610 . Spectral inflammation map visualization module 610 displays the quantified fat by coloring each pixel in the spectral projection data and/or spectral volumetric image data based on its fat concentration and/or inflammation concentration. Concentration and inflammation concentration are displayed.

FIG. 8 illustrates an exemplary method according to embodiments described herein.

At 802, spectroscopic data (ie, spectroscopic projection data or spectroscopic volumetric image data) is received as described herein and/or otherwise.

At 804, the spectroscopic data is materially decomposed to generate a set of vectors for two different bases, as described herein and/or otherwise.

At 806, the vector representing each base is used to obtain the concentration of fat and/or inflammation within the voxel, as described herein and/or otherwise.

At 808, fat and/or inflammation within the determined voxels is displayed as described herein and/or otherwise.

FIG. 9 illustrates an exemplary method according to embodiments described herein.

At 902, spectroscopic data (ie, spectroscopic projection data or spectroscopic volumetric image data) is received as described herein and/or otherwise.

At 904, the spectroscopic data is materially decomposed to generate a set of vectors for two different bases, as described herein and/or otherwise.

At 906, the vector representing each base is used to obtain the density of each base within the voxel, as described herein and/or otherwise.

At 908, the locations in the two base maps with known fat or inflammation are used to obtain vectors representing fat and inflammation , as described herein and/or otherwise.

At 910, the concentration of fat and/or inflammation within the voxel is determined as described herein and/or otherwise.

At 912, fat and/or inflammation within the determined voxels are displayed as described herein and/or otherwise.

It should be understood that the order of the above operations is not limiting. As such, other orderings are contemplated herein. Also, one or more acts may be omitted and/or one or more additional acts may be included.

The foregoing is embodied by computer readable instructions encoded on or embodied in a computer readable storage medium which, when executed by a computer processor, cause the processor to perform the operations described above. Additionally or alternatively, at least one of the computer-readable instructions may be carried by a signal, carrier wave, or other transitory medium that is not a computer-readable storage medium.

The invention has been described with reference to preferred embodiments. Modifications and alterations will occur to others upon reading and understanding the preceding detailed description. It is intended that the invention be construed as including all such modifications and alterations insofar as they come within the scope of the appended claims or the equivalents thereof.

Claims (20)

a memory storing an inflammation map generator module;
receiving at least one of spectroscopic projection data or spectroscopic volumetric image data and decomposing said at least one of spectroscopic projection data or spectroscopic volumetric image data using 2-base decomposition to produce said spectroscopic projection data or said spectroscopic volumetric image data; generating a set of vectors for each base represented in said at least one of said image data; calculating from said set of vectors for each base a density of each base within a voxel; from said density of each base; a processor that determines a concentration of at least one of fat or inflammation within the voxel;
a display displaying the determined concentration of the at least one of fat or inflammation ;
A system comprising: a radiation source that emits X-ray radiation;
a detector array for detecting x-ray radiation and generating said spectroscopic projection data;
a reconstructor for reconstructing a signal to produce the spectroscopic volumetric image data;
3. The system of claim 1, further comprising an imaging system comprising: 3. The system of claim 1 or 2, wherein one substance of the 2-base breakdown is fat and the other is inflammation . The processor further

determining the concentration of the at least one of fat or inflammation within the voxel by solving
where Low kv is the spectroscopic volumetric image data from the low energy image, High kv is the spectroscopic volumetric image data from the high energy image,

is the bias from the origin, and

is a vector representing fat, c _fat is the fat concentration,

4. The system of claim 3, wherein is a vector representing inflammation and c _inf is an inflammation concentration. 3. The system of claim 1 or 2, wherein prior to performing the 2-base decomposition, values for photoelectric effect and Compton scattering are determined from spectroscopic projection data or spectroscopic volumetric image data. The processor further

determining values for at least one of the photoelectric effect and Compton scattering by solving for
where PhotoE is the photoelectric effect, CSscatter is the Compton scattering,

is a vector representing the photoelectric effect as one axis of 2D coordinates , c _PhotoE is a value for the photoelectric effect,

6. The system of claim 5, wherein is a vector representing the Compton scatter as the other axis of 2D coordinates , and c _CSscatter is a value for the Compton scatter. The processor further

determining the concentration of the at least one of fat or inflammation within the voxel by solving
here,

is the representation of the voxel in the 2D coordinate system defined by the 2-base decomposition and given the pixel values, with

as well as

A 2D coordinate with

as well as

in a coordinate system defined by

equal to

represents the bias from the origin, and

is a vector representing fat, c _fat is the concentration of fat,

7. The system of claim 6, wherein is a vector representing inflammation and c _inf is a concentration of inflammation . The concentration of at least one of fat or inflammation is displayed by coloring pixels based on the concentration of at least one of fat or inflammation in corresponding voxels, voxels representing inflammation being displayed in a first color. voxels representing fat are displayed in a different second color and voxels representing a combination of inflammation and fat are displayed in a different third color in a color range between said first color and said second color. 2. The system of claim 1, displayed in a color of . receiving at least one of spectroscopic projection data or spectroscopic volumetric image data;
decomposing said at least one of spectroscopic projection data or spectroscopic volumetric image data using 2-base decomposition to produce a vector for each base represented in said at least one of spectroscopic projection data or spectroscopic volumetric image data; generating a set;
calculating the density of each base within a voxel from the set of vectors for each base;
determining a concentration of at least one of fat or inflammation within said voxel from said concentration of each base;
displaying the determined concentration of the at least one of fat or inflammation ;
A method. 10. The method of claim 9, wherein one substance of the 2-base breakdown is fat and the other is inflammation .

determining the concentration of the at least one of fat or inflammation within the voxel by solving
where Low kv is the spectroscopic volumetric image data from the low energy image, High kv is the spectroscopic volumetric image data from the high energy image,

is the bias from the origin, and

is a vector representing fat, c _fat is the fat concentration,

11. The method of claim 10, wherein is a vector representing inflammation and c _inf is an inflammation concentration. 10. The method of claim 9, wherein one base of the 2-base decomposition is the photoelectric effect and the other is Compton scattering.

determining values for at least one of the photoelectric effect and Compton scattering by solving for
where PhotoE is the photoelectric effect, CSscatter is the Compton scattering,

is a vector representing the photoelectric effect as one axis of the 2D coordinates , c _PhotoE is the value for the photoelectric effect,

13. The method of claim 12, wherein c is a vector representing Compton scattering as the other axis of the 2D coordinates , and _cCScatter is a value for Compton scattering.

determining the concentration of the at least one of fat or inflammation within the voxel by solving
here,

is a representation of said voxel in a 2D coordinate system defined by material decomposition and given pixel values, and as axes:

as well as

A 2D coordinate with

as well as

in a coordinate system defined by

equal to

represents the bias from the origin, and

is a vector representing fat, c _fat is the concentration of fat,

14. The method of claim 13, wherein is a vector representing inflammation and c _inf is the concentration of inflammation . The concentration of the at least one of fat and inflammation is displayed by coloring pixels based on the concentration of the at least one of fat and inflammation in corresponding voxels, and voxels representing inflammation are displayed in a first color. , voxels representing fat are displayed in a different second color and voxels representing a combination of inflammation and fat are displayed in a different third color in the color range between said first color and said second color. 10. The method of claim 9, displayed in color. A computer readable storage medium encoded with computer readable instructions which, when executed by a processor of a computing system, causes the processor to:
receiving at least one of spectroscopic projection data or spectroscopic volumetric image data;
decomposing said at least one of spectroscopic projection data or spectroscopic volumetric image data using 2-base decomposition to produce a vector for each base represented in said at least one of spectroscopic projection data or spectroscopic volumetric image data; generating a set;
calculating the density of each base within a voxel from the set of vectors for each base;
determining the concentration of at least one of fat or inflammation within the voxel from each base concentration;
displaying the determined concentration of the at least one of fat or inflammation ;
A computer-readable storage medium that causes 17. The computer-readable storage medium of claim 16, wherein one substance of the 2-base breakdown is fat and the other is inflammation . 17. The computer-readable storage medium of claim 16, wherein one base of the two-base decomposition is the photoelectric effect and the other is Compton scattering. Execution of the computer readable instructions further causes the processor to learn at least one of fat or inflammation vector identification from at least one of training data or known locations of fat and/or inflammation within the subject. Item 17. The computer-readable storage medium of Item 16. The concentration of the at least one of fat and inflammation is displayed by coloring pixels based on the concentration of the at least one of fat and inflammation in corresponding voxels, and voxels representing inflammation are displayed in a first color. , voxels representing fat are displayed in a different second color and voxels representing a combination of inflammation and fat are displayed in a different third color in the color range between said first color and said second color. 17. The computer-readable storage medium of claim 16, displayed in color.

Images