JP5710166B2 - Image analysis apparatus and method - Google Patents

Description

  The present invention relates to an analysis technique for RI (Radio Isotope) images, and more particularly to a technique for comparing RI images obtained by photographing the same examination site of the same subject at different times and evaluating the amount of RI that has flowed out from the examination site.

  When an RI myocardial blood flow preparation (for example, 99mTc-SESTAMIBI or 99mTc-Tetrofosmin) is administered to a subject as an RI drug, the RI drug flows out from the myocardium after acute myocardial infarction treatment or severe angina An increase in washout is observed. Therefore, there is an inspection method that is useful for diagnosing a heart disease or the like by using a plurality of images obtained by administering a RI medicine to a subject and then performing SPECT imaging of the subject's heart in different time zones thereafter. For example, washing from the heart can be obtained from the count value of the early image taken in the first time zone after the RI drug administration and the count value of the late image taken in the second time zone after the first time zone. There is known a washout rate (Washout rare, WR) indicating the ratio of the amount of RI drug that has been released (for example, Non-Patent Document 1).

A.M.Peters.A Unified Approach to Quantification by Kinetic Analysis in Nuclear Medicine.J Nucl Med. 1993; 34: 706-13.

  However, although WR is a quantitative index, the calculated index value often deviates greatly from the qualitative evaluation performed when visually comparing the images of the early image and the late image. For example, even if the value of WR is the same, there are cases where the visual evaluation determines that the subject is normal or abnormal. This is because the imaging time interval is different and there is an individual difference in the amount of RI drug that is washed out for each subject. Therefore, in the clinical field, there has been a demand for a quantitative evaluation index that is in good agreement with a qualitative evaluation in which images of early and late images are visually compared.

  Accordingly, an object of the present invention is to provide a quantitative evaluation index that is consistent with qualitative evaluation in which an early image and a later image are visually compared.

  The image analysis apparatus according to one embodiment of the present invention provides first captured image data relating to an RI image obtained by capturing an examination site of a subject in a first time zone after administering a RI (Radio Isotope) drug to the subject. And storage means for storing second captured image data relating to an RI image obtained by imaging the examination site of the subject in a second time zone after a predetermined time from the first time zone, Normalization means for normalizing the captured image data to obtain first normalized data, normalizing the second captured image data to obtain second normalized data, the first normalized data and the second normalized data Calculation means for calculating an index for evaluating the washout amount of the RI drug from the examination target site based on the digitized data, and output means for outputting the index calculated by the calculation means, Obtain.

  In a preferred embodiment, the RI agent may be an RI myocardial blood flow product.

  In a preferred embodiment, the RI agent may be 99mTc-SESTAMIBI.

  In a preferred embodiment, the examination site may be the heart.

  In a preferred embodiment, the normalizing means divides the examination target part into a plurality of segments for each of the first photographed image data and the second photographed image data, and uses a representative value of each segment. And normalizing each segment to obtain the first and second normalized data. And the said calculation means may be made to calculate the said parameter | index about each segment based on the said 1st and 2nd normalization data normalized for every said segment.

  In a preferred embodiment, the index may be an index indicating a change rate from the first normalized data to the second normalized data.

In a preferred embodiment, the index may be WOINDx (n) calculated by the following formula.
E_SEG (n): representative value of each segment of early image L_SEG (n): representative value of each segment of late image n: segment identifier E_SEG_MAX: maximum value of representative value of each segment of early image L_SEG_MAX: each value of late image Maximum segment representative value

  In a preferred embodiment, the first photographed image data is array data obtained by converting a SPECT (Single Photon Emission Tomography) image photographed in the first time zone, and the second photographed image data is Array data obtained by converting a SPECT image captured in the second time zone, and the image analysis device further includes display means for displaying a polar map based on the first and second captured image data. Also good. The first and second normalized data are data normalized for each segment of the polar map, and the calculation unit may calculate the index for each segment of the polar map. Good.

It is a lineblock diagram of image analysis device 1 concerning one embodiment of the present invention. It is a flowchart which shows the production | generation procedure of array data. It is a figure which shows the tomographic image of the heart typically. It is a flowchart which shows the whole process sequence of WOINDx calculation. It is a flowchart which shows the specific procedure of a normalization process. It is explanatory drawing of conversion from array data to normalization segment data. It is a display example of a polar map and WOINDx.

  Hereinafter, an image analysis apparatus according to an embodiment of the present invention will be described with reference to the drawings.

  FIG. 1 is a configuration diagram of an image analysis apparatus 1 according to the present embodiment. The image analysis apparatus 1 according to the present embodiment analyzes a plurality of RI images of the examination site of the subject taken at different time zones after administering the RI drug to the subject, and the amount of RI that has flowed out of the examination target site An index for evaluating is calculated. In this embodiment, in particular, analysis is performed using a SPECT (Single Photon Emission Tomography) image of the heart imaged by a SPECT imaging apparatus (not shown) using 99mTc-SESTAMIBI as the RI agent.

  The image analysis apparatus 1 is configured by, for example, a general-purpose computer system, and individual components or functions of the image analysis apparatus 1 described below are realized by executing a computer program, for example. This computer program can be stored in a computer-readable recording medium.

  As shown in the figure, the image analysis apparatus 1 includes an image data storage unit 11, an array data generation unit 13, an array data storage unit 15, a normalization processing unit 16, an index calculation unit 17, and an index storage unit. 18 and a display processing unit 19.

  The image data storage unit 11 stores SPECT image data of the heart taken by a SPECT imaging apparatus (not shown). Here, the SPECT image data includes the early image data 111 obtained by imaging the subject's heart in the first time zone after administration of the RI drug, and the second time zone after a predetermined time from the first time zone. The late image data 113 obtained by photographing the heart of the same subject is stored. The early image is, for example, a SPECT image taken 1 hour after RI drug administration, and the late image is, for example, a SPECT image taken 6 hours after RI drug administration.

  Both the early image data 111 and the late image data 113 are three-dimensional voxel data. The value of each voxel is a count value counted by the SPECT imaging device.

  The array data generation unit 13 generates the polar image (bulls eye) based on the early image data 111 and the late image data 113 stored in the image data storage unit 11, and the early image array data 151 and the late image array data. 153 is generated.

  Here, FIG. 2 is a flowchart showing a sequence of generating array data performed by the array data generating unit 13. The processing procedure of the array data generation unit 13 will be described in detail with reference to FIG. The array data generation unit 13 generates the early image array data 151 and the late image array data 153 by applying the following processing to the early image data 111 and the late image data 113, respectively.

  First, the array data generation unit 13 reads image data of an image to be processed (early image data 111 or late image data 113) (S11).

  The array data generation unit 13 selects the range of the heart region in the target image and sets the range as the processing range (S12). The selection of the heart region may be performed in accordance with an instruction from the operator, or the array data generation unit 13 may automatically detect the heart region.

  FIG. 3 schematically shows a tomographic image of the heart. FIG. 3A is a schematic diagram of a vertical long-axis image, and FIG. 1B is a schematic diagram of a short-axis image. For example, the operator selects each heart region by using the input device 21 when the display processing unit 19 displays images as shown in FIGS. For example, the operator designates the base portion be and the apex portion ae using the vertical long axis image of FIG. Also, the center ce and radius re (outer circumference) of the heart are set using the short axis image of FIG.

  Returning to FIG. 2, the array data generation unit 13 generates a predetermined number of short-axis images (here, 16 images) based on the image data of the short-axis image, according to the length between the base portion be and the apex portion ae. ) To normalize (S13).

  As shown in FIG. 3B, the array data generation unit 13 uses the radius re in the 3 o'clock direction from the center ce of the heart as the starting point at 7.5 degrees, and for each radius re in the 48 directions, The largest pixel value on re is extracted (S14). The array data generation unit 13 performs this process for each of the 16 short-axis images obtained by normalization (S15).

  Thereby, 48 × 16 array data is extracted. Based on this array data, a polar map, which will be described later, is displayed.

  The early image array data 151 (first captured image data) and the late image array data 153 (second captured image data) generated by the array data generation unit 13 are stored in the array data storage unit 15, respectively.

  The normalization processing unit 16 normalizes the early image array data 151 and the late image array data 153, respectively. That is, the normalization processing unit 16 normalizes the early image array data 151 (first captured image data) to obtain normalized segment data (first normalized data) of the early image, and the late image array data 153 (first data). 2), normalized segment data (second normalized data) of the late image is obtained. The normalization processing unit 16 divides the heart into a plurality of segments for each of the early image array data 151 (first captured image data) and the late image array data 153 (second captured image data). Normalization is performed for each segment using the representative value to obtain normalized segment data (first and second normalized data) of the early and late images. The normalized segment data of the early image and the late image is data normalized for each segment obtained by dividing the heart into a plurality of segments.

  Based on the normalized segment data of the early image (first normalized data) and the normalized segment data of the later image (second normalized data), the index calculating unit 17 extracts the RI drug from the heart of the subject. An index for evaluating is calculated. For each segment, the index calculation unit 17 calculates an index for evaluating the RI drug washout amount for each segment. The index calculated here is referred to as WashoutINDEX (WOINDx).

  FIG. 4 is a flowchart showing an overall processing procedure for calculating WOINDx. The overall processing procedure for WOINDx calculation will be described with reference to FIG.

  First, the normalization processing unit 16 applies normalization processing, which will be described later, to the early image array data 151 to generate normalized segment data of the early image (S21).

  Similarly, the normalization processing unit 16 applies normalization processing, which will be described later, to the late image array data 153 to normalize, and generates normalized image data of the late image (S23). Note that normalization of the early image and normalization of the late image may be performed first.

  Then, the index calculation unit 17 calculates WOINDx using the normalized segment data of the early image normalized in step S21 and the normalized segment data of the later image normalized in step S23 (S25). . Hereinafter, each process of step S21-S25 is demonstrated in detail.

  First, details of the normalization process in steps S21 and S23 will be described. Steps S21 and S23 have the same processing contents except for the target data. FIG. 5 is a flowchart showing a specific procedure of the normalization process. FIG. 6 is an explanatory diagram of conversion from array data to normalized segment data. Hereinafter, the normalization process will be described with reference to FIGS. 5 and 6.

  The normalization processing unit 16 normalizes each value in the array data so that the maximum value becomes 100 in the target array data ar [i, j] (early image array data 151 or late image array data 153). (% Uptake) (S31).

  Next, each value of the target array data n_ar [i, j] normalized in step S31 is divided into 17 segments when creating a polar map, and a representative value SEG (n) (n is a segment) Segment data (early image segment data or late image segment data) is obtained (S33). Here, it is determined in advance to which segment each value of the target array data belongs. Note that the number of segments is not necessarily 17 and may be any number. The representative value SEG [n] of each segment represents each segment such as an average value, maximum value, minimum value, intermediate value, or mode value of normalized array data n_ar [i, j]. Value is acceptable.

The normalization processing unit 16 further normalizes the segment data SEG (n) determined in step S33 to generate normalized segment data SV [n] (S35). For example, the normalization processing unit 16 calculates the representative value SEG (n) of 17 segments so that the maximum value (SEG_MAX) of the representative values SEG (n) of each segment becomes 100 according to the following equation (1). To generate a normalized representative value SV (n) for each segment (% Uptake for each segment).

  In this embodiment, what was normalized to% Uptake in step S31 is further normalized in step S35. Thereby, since the base of the data used as a comparison object is prepared, there exists an effect which suppresses the dispersion | variation by the individual difference for every test subject. It is not essential to repeat normalization in this way. For example, step S31 may be omitted from the two normalization processes, and only step S35 may be performed.

Here, the segment data of the early image is represented by E_SEG (n), and the segment data of the later image is represented by L_SEG (n). Further, the maximum value in the segment data E_SEG (n) of the early image is E_SEG_MAX, and the maximum value in the segment data L_SEG (n) of the late image is L_SEG_MAX. At this time, the normalized segment data E_SV (n) of the early image and the normalized segment data L_SV (n) of the late image are defined by the following equations (2) and (3), respectively.

The index calculation unit 17 uses the above-mentioned early-image normalization segment data E_SV (n) and the late-image normalization segment data L_SV (n) to identify the segment-specific index WOINDx (n ) Is calculated.

  WOINDx (n) is the rate of change from the normalized segment data E_SV (n) (first normalized data) of the early image by segment to the normalized segment data L_SV (n) (second normalized data) of the later image It is an index showing. That is, WOINDx (n) represents the rate of change of the count value from the early image to the late image with respect to the early image for each segment. At that time, since the count values of the early image and the late image are compared using values normalized in the respective images, attenuation of the radioactivity released by the RI and normal washing of the normal region (physiologically) Can be removed.

  The WOINDx calculated by the index calculation unit 17 is stored in the index storage unit 18.

  The display processing unit 19 generates a polar map based on the early image array data 151 and the late image array data 153 (the first and second captured image data), and causes the display device 22 to display the polar map. Further, the display processing unit 19 causes the display device 22 to display WOINDx stored in the index storage unit 18. These display examples are shown in FIG.

  Fig. A is a polar map of an early image based on the early image array data 151, Fig. B is a polar map of an late image based on the late image array data 153, and Fig. C is a table in which WOINDx is displayed by segment. It is an example. WOINDx indicates that the greater the number, the greater the amount of RI washed out of the heart.

  As a result, according to the evaluation of doctors who looked at FIGS. A and B, a segment with a large amount of washing out is a region indicated by A in the figure. Then, according to FIG. 3C, it can be seen that the WOINDx of the segment in the area A has a larger value than the other segments. That is, WOINDx is in good agreement with the qualitative evaluation in which the early image and the later image are visually compared.

  In general, the time interval between the early image and the late image is often different when the facility where the examination is performed is different. Therefore, since the degree of attenuation of the radioactivity released by RI and the amount of normal RI that is washed out from the normal region are different, it is difficult to compare the test results at different facilities. However, WOINDx according to the present embodiment removes the effects of attenuation of radioactivity released by RI over time and normal washing out of normal areas, so comparisons can be made even with test results at different facilities. It is. That is, the test results at different facilities can be evaluated using the same threshold value.

  Further, by using% Uptake in normalization, it is possible to objectively and quantitatively evaluate the difference between the normal region and the abnormal region even if there is a difference in RI nuclide. Therefore, the clinical use of WOINDx calculated using data normalized by% Uptake is considered wide.

  Furthermore, in this embodiment, when calculating WINDx, the normalized segment is normalized so that the maximum value E_SEG_MAX in the segment data of the early image and the maximum value L_SEG_MAX of the segment data of the later image are both 100. Data E_SV (n) and L_SV (n) are used. For this reason, since the base for comparing the early image and the later image is prepared, it is possible to quantitatively evaluate the segment in which the washing is promoted when compared with the early image in the later image.

  Further, since the value normalized by E_SEG_MAX or L_SEG_MAX is used for both the denominator and numerator in Expression (4) (see Expressions (2) and (3)), WOINDx is 0 in the normal region where the amount of washing is relatively small. A value close to. For this reason, if WOINDx is used, it is possible to clearly distinguish the normal region from the region where washing out is promoted. This is in contrast to WR, where the difference in values between normal and abnormal areas is not clear.

  Further, if the late image is collected under the same conditions as the early image, the count value is small, and overestimation of the abnormal part becomes a problem. Therefore, in order to maintain the image quality of the late image, it is necessary to lengthen the collection time of the late image. In the WOINDx of the present embodiment, the count value is normalized, so that it is not affected by the length of the collection time.

  The above-described embodiments of the present invention are examples for explaining the present invention, and are not intended to limit the scope of the present invention only to those embodiments. Those skilled in the art can implement the present invention in various other modes without departing from the gist of the present invention.

  For example, the examination target site may be an organ other than the heart. Furthermore, although the image analysis based on the SPECT image has been described in the present embodiment, the present invention can also be applied to a PET (Positron Emission Tomography) image.

1 Image Analysis Device 11 Image Data Storage Unit 13 Array Data Generation Unit 15 Array Data Storage Unit 16 Normalization Processing Unit 17 Index Calculation Unit 18 Index Storage Unit 19 Display Processing Unit 111 Early Image Data 113 Late Image Data 151 Early Image Array Data 153 Late image array data

Claims (11)

First radiographed image data relating to an RI image obtained by imaging an examination site of the subject in a first time zone after administering a RI (Radio Isotope) drug to the subject, and a predetermined time after the first time zone Storage means for storing second photographed image data relating to an RI image obtained by photographing the examination site of the subject in a second time period;
A plurality of first representative values are determined from the first photographed image data , and each first representative value is normalized with a maximum value among the plurality of first representative values to obtain first normalized data. A plurality of second representative values are determined from the second photographed image data , and each second representative value is normalized with the maximum value among the plurality of second representative values to obtain second normalized data. Normalization means,
A calculation means for calculating an index for evaluating the amount of the RI drug washed out from the examination site based on the first normalized data and the second normalized data;
An image analysis apparatus comprising: output means for outputting an index calculated by the calculation means.   The image analysis apparatus according to claim 1, wherein the RI drug is an RI myocardial blood flow preparation.   The image analysis apparatus according to claim 1, wherein the RI drug is 99mTc-SESTAMIBI.   The image analysis apparatus according to claim 1, wherein the examination site is a heart. The normalization unit divides the examination target part into a plurality of segments for each of the first photographed image data and the second photographed image data, and sets representative values of each segment to the plurality of first and 2 to obtain the first and second normalized data by normalizing each segment ,
5. The calculation unit according to claim 1, wherein the calculation unit calculates the index for each segment based on the first and second normalized data normalized for each segment. 6. Image analysis device. The index is, the image analysis apparatus according to any one of claims 1 to 5, characterized in that from said first normalized data is an index showing the rate of change to the second normalized data. The image analysis apparatus according to claim 5, wherein the index is WOINDx (n) calculated by the following formula.
E_SEG (n): representative value of each segment of early image L_SEG (n): representative value of each segment of late image n: segment identifier E_SEG_MAX: maximum value of representative value of each segment of early image L_SEG_MAX: each value of late image Maximum segment representative value The first captured image data is array data obtained by converting a SPECT (Single Photon Emission Tomography) image captured in the first time zone,
The second captured image data is array data obtained by converting a SPECT image captured in the second time zone,
The image analysis apparatus further includes display means for displaying a polar map based on the first and second captured image data,
The first and second normalized data are data normalized for each segment of the polar map,
The image analysis apparatus according to claim 1, wherein the calculation unit calculates the index for each segment of the polar map. A method performed by an image analysis device,
First radiographed image data relating to an RI image obtained by imaging an examination site of the subject in a first time zone after administering a RI (Radio Isotope) drug to the subject, and a predetermined time after the first time zone Storing second photographed image data relating to an RI image obtained by photographing the examination site of the subject in a second time zone;
A plurality of first representative values are determined from the first photographed image data , and each first representative value is normalized with a maximum value among the plurality of first representative values to obtain first normalized data. A plurality of second representative values are determined from the second photographed image data , and each second representative value is normalized with the maximum value among the plurality of second representative values to obtain second normalized data. Steps,
Calculating an index for evaluating the washout amount of the RI medicine from the examination site based on the first normalization data and the second normalization data;
Outputting the calculated index; and an image analysis method. A computer program for an image analysis device,
In the image analysis device,
First radiographed image data relating to an RI image obtained by imaging an examination site of the subject in a first time zone after administering a RI (Radio Isotope) drug to the subject, and a predetermined time after the first time zone Storing second photographed image data relating to an RI image obtained by photographing the examination site of the subject in a second time zone;
A plurality of first representative values are determined from the first photographed image data , and each first representative value is normalized with a maximum value among the plurality of first representative values to obtain first normalized data. A plurality of second representative values are determined from the second photographed image data , and each second representative value is normalized with the maximum value among the plurality of second representative values to obtain second normalized data. A normalization step;
An index calculating step for calculating an index for evaluating the amount of the RI drug washed out from the examination site based on the first normalized data and the second normalized data;
And a step of outputting the calculated index.   The normalizing step divides the examination target part into a plurality of segments for each of the first photographed image data and the second photographed image data, and sets a representative value of each segment to the plurality of first and second parts. 2 to obtain the first and second normalized data by normalizing each segment,
  11. The computer program according to claim 10, wherein the index calculation step calculates the index for each segment based on the first and second normalized data normalized for each segment.