JP2011036684A - Computer supported image diagnosing system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To precisely and accurately match slice positions of images in past and present with each other even when photographing start points are different or intervals of slices are different in photographing of a specific organ in different photographing periods. <P>SOLUTION: In a computer-supported image diagnosing system 10A, a slice matching means 22A provided in a computer 12A includes: a means 41 for calculating an area change rate of the specific organ for each of slice images included in a group of the slice images of a first photographing period, and also calculating an area change rate of the specific organ for each of slice images included in a group of the slice images of a second photographing period; a means 42 for comparing the area change rate of the required slice image of the second photographing period with the area change rate for each of the slice images included in the group of the slice images of the first photographing period; and a means 43 for selecting from the group of the slice images of the first photographing period the slice image having the area change rate whose difference from the area change rate of the required slice image in the second photographing period is within a prescribed range. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a technique for automatically detecting a lung cancer candidate shadow at an early stage using CT images acquired by helical CT scans at different imaging timings. The present invention relates to a computer-aided diagnostic imaging system that matches a slice position with a slice obtained in the above.

  Looking at deaths by cause in Japan, cancer deaths are increasing year by year. In particular, lung cancer is one of the most well-known causes of death, and overcoming lung cancer is a major social problem. In order to overcome this, it is necessary to detect and treat appropriately at an extremely early stage. In particular, it is important to detect small lung cancer that is rapidly progressing. Therefore, in order to improve the rate of overcoming lung cancer, mass screening at the initial stage where cancer cells grow is essential.

  As a conventional mass screening method, a chest X-ray film has been used for lung cancer diagnosis. This chest X-ray film is a two-dimensional projection image, and the shadows of bones and organs overlap with each other, and as a result, early detection of microlung cancer is hindered. In recent years, a helical CT (Computerized Tomography) scanner that repeatedly collects projection data while drawing the spiral activation of the X-ray tube in the opposite direction of the examiner can take an image of the entire lung field in a short time. It became so.

  However, in a mass examination using a helical CT image collected by a helical CT scanner, the number of images read by the examination doctor is enormous. Therefore, group screening takes time for diagnosis and is difficult to apply to clinical practice.

  Thus, a filmless computer-aided diagnosis (CAD) system, which is an invention provided with a database of large-capacity CT images and provided with these image search and image matching functions, has been disclosed. In this system, slice positions are automatically matched (matched) with image data of the same examiner at different imaging times. In this slice matching, a slice of interest in the current CT image and a past slice including the same image information as this slice are obtained and a difference between these slices is calculated, and the slice between the current slice and the past slice is calculated. Match position. The past and current CT images are displayed side-by-side so that the examiner can easily compare and interpret (see, for example, Patent Document 1).

JP 2001-137230 A

  However, the extraction region of a specific organ in a very thin and large amount of slice images includes many approximate region forms. Therefore, it can be considered as one of the causes of reducing the accuracy for aligning the anatomical position in the body axis direction, and appropriate slice matching is not always realized.

  In addition, since one of the images acquired in the past and the present is an image taken in the past, it is considered that there are many images taken with an old CT apparatus or single slice CT. The slice interval (slice thickness) between different past slice images and the current slice image may be different.

  When the slice interval between the current CT image and the past CT image is different, if the slice matching process is performed according to Patent Document 1, the accuracy is lowered. If the accuracy of slice matching is low, the slice position of the current CT image and the slice position of the past CT image are different in comparing the current CT image and the past CT image, that is, in the opposite axis direction. The images at different positions are compared, and comparative interpretation loses its diagnostic significance.

  The present invention has been made in consideration of the above-described circumstances, and it is possible to accurately match the slice positions in the past and current images even when the imaging start point is different or the slice interval is different when imaging a specific organ at different imaging timings. An object of the present invention is to provide a computer-aided image diagnostic system that can be performed with high accuracy.

  In order to solve the above-described problem, the computer-aided image diagnosis system according to the present invention is configured to acquire a slice image group and a second imaging time at the first imaging time acquired by acquiring images of the specific organs of the examiner at different imaging times. Storage unit for storing the slice image group of the image, a diagnosis process execution unit for reading out the slice image group at the first imaging time and the slice image group at the second imaging time from the storage unit, and executing the support diagnosis process, and the support Display means for displaying a result of the support diagnosis process performed by the diagnosis process execution means, and the support diagnosis process execution means includes a region of the specific organ for each slice image included in the slice image group at the first imaging time. An area change for calculating a change rate of the specific organ for each slice image included in the slice image group at the second imaging time is calculated. A rate change means, a rate of change of area in a slice image at a required second imaging time, and a rate of change of area in a slice image group included in the slice image group at the time of the first imaging time; Slice image selection means for selecting a slice image having a region change rate whose difference from a region change rate in the slice image at the required second photographing time is within a predetermined range from the slice image group at the first photographing time; It was equipped.

  According to the computer-aided image diagnosis system according to the present invention, matching of slice positions in the past and current images is performed accurately and accurately even when the imaging start point is different or the slice interval is different in imaging of a specific organ at different imaging timings. be able to.

1 is a block diagram showing a first embodiment of a computer-aided image diagnostic system according to the present invention. The figure which shows an example of the relationship between the slice position of each image, and the area of a lung field area | region as a graph. The figure which showed the relationship between the distance between slices from a reference | standard slice image to the other slice image contained in a slice image group, and the area of a lung field area | region as a graph. (A) (b) The figure which shows the example of a display of the CT image image | photographed by the past and the present examination. The figure which shows the example of a display of CT image image | photographed by the past and the present examination. The block diagram which shows 2nd Embodiment of the computer-aided image diagnostic system which concerns on this invention.

  Embodiments of a computer-aided image diagnostic system according to the present invention will be described with reference to the accompanying drawings.

  FIG. 1 is a block diagram showing a first embodiment of a computer-aided image diagnostic system according to the present invention.

  Fig. 1 shows the initial lung cancer candidate shadows automatically from the progress of the disease using slice images (tomographic images) acquired and collected by current and past helical CT (Computerized Tomography) scans at different imaging times. FIG. 1 shows a computer-aided diagnosis (Computer Assisted Diagnosis: CAD) system 10 to be detected. Among the algorithms that support comparative interpretation of slice images acquired at different imaging times, the most important is the slice position of the current CT image acquired by the CT scan and the slice position of the past CT image for the same examiner. Is a slice matching algorithm that aligns the anatomical position in the body axis direction of the examiner between the current and past images.

  The computer-aided image diagnosis system 10 includes a slice image group of the entire lung acquired by a helical CT scan of a part including a specific organ of the examiner (patient), for example, a chest, several times until the current examination. The database 11 as storage means for storing each imaging time, and the past and current slice image groups having different imaging times are read from the database 11 and, for example, support diagnosis processing for executing lung cancer support diagnosis processing according to the program code A computer 12 as an execution unit and a display unit 13 for displaying the result of support diagnosis processing performed by the computer 12 are provided.

  The computer 12 includes, as a subprogram code, a lung field region extraction means 21 that extracts a lung field region that is an index of slice matching processing from slice images included in each of the current slice image group and the past slice image group, and this Using the lung field region extracted by the lung field region extraction unit 21, the slice matching unit 22 that matches the slice position between the current slice image of interest and the past slice image, and the past and present And a feature analysis means 23 for measuring a feature amount of a region suspected of having lung cancer in the slice image. Although not shown, the feature analysis unit 23 includes a feature amount calculation unit that calculates feature amounts from past and current slice images, and a lung cancer candidate detection unit that detects lung cancer candidates from the feature amounts.

  The slice matching unit 22 of the computer 12 sets a past reference slice image from the past slice image group and a reference slice image setting unit 31 that sets the current reference slice image from the current slice image group. And calculating the inter-slice distance between the past reference slice image set by the reference slice image setting means 31 and another slice image having a lung field area included in the past slice image group, and the current reference slice The inter-slice distance calculating means 32 for calculating the inter-slice distance between the image and the other slice image included in the current slice image group and having the lung field region, and the inter-slice distance between the past reference slice image Same as the inter-slice distance between the reference slice image and the current slice image of interest in the current slice image group. A slice image that a slice image selection means 33 for selecting from the past of the slice image group is provided.

  The display unit 13 includes an image display mode setting unit 35 that displays a current slice image and a past slice image matched with the current slice image in a pageable manner, and a display set by the image display mode setting unit 35. A CRT (Cathode Ray Tube) 36 for displaying an image on the screen is provided.

  Next, the operation of the computer-aided image diagnostic system 10 will be described.

  The helical CT scanner can image the entire lung field in a short time. The imaging conditions for mass screening of lung cancer are, for example, a beam width of 10 mm, a top plate speed of 20 mm / s, a tube voltage of 120 kV, a tube current of 50 mA, and a reconfiguration interval of 10 mm, during one breath hold (about 15 sec.). A helical CT image of the entire lung field is acquired. The helical CT image is a three-dimensional image having a 180-degree linear interpolation, an image size of 512 × 512 pixels reconstructed at an interval of 10 mm, and a gradation of 12 bits per pixel. About 35 slices are taken for every diagnosis from all examiners and stored in the database 11.

  Next, when interpretation is performed using the helical CT images of the same examiner collected in the past group examination imaging, the computer 11 of the computer-aided image diagnosis system 10 accesses the database 11. Past slice image data consisting of multiple slice images acquired by imaging the entire lungs of the examiner in the past and current slice image consisting of multiple slice images acquired by imaging the entire lungs of the same examiner The group data is retrieved from the database 11, and the past and current slice image groups are read out to the computer 12, respectively. In the computer 12, using the past and current slice image groups relating to the examiner read from the database 11, support diagnosis processing, for example, lung cancer support diagnosis processing, is executed according to the program code.

  At the time of slice matching executed by the computer 12, the lung field region extracting means 21 determines the lung field regions of the slice images included in the past and current slice image groups based on the threshold processing and the defect portion correction processing, respectively. Extracted. The lung field region contains a large amount of air, and the CT value in the lung field region is lower than the CT value of the skeleton or soft tissue, so that the extraction ability is relatively high.

  FIG. 2 is a graph showing an example of the relationship between the slice position of each image and the area of the lung field region in the slice images included in the past and current slice image groups.

  The graph shown in the upper part of FIG. 2 shows the relationship between the anatomical position in the body axis direction and the area of the lung field area, and the area of the lung field area becomes 0 at the anatomical position without the lung field area. A lung field region exists at an anatomical position on the right side of the figure from the end of the field region. The graph shown in the middle row shows the relationship between the slice position of each image and the area of the lung field region, and the slice image is sliced with the anatomical position a1 in the graph shown in the upper row as the imaging start point (imaging end point). (Slice thickness) This is a group of past slice images collected at 10 mm. Further, the graph shown in the lower part shows the relationship between the slice position of each image and the area of the lung field region, and the slice images were collected at the slice interval of 5 mm using the anatomical position b1 of the graph shown in the upper part as the imaging start point. The current slice image group.

  In the graph shown in the middle stage, the lung field region is not extracted between the image of the slice position A1 obtained by photographing the anatomical position a1 and the image of the slice position A2 having a slice interval of 10 mm from the slice position A1 (the lung field). On the other hand, there is a slice interval of 10 mm from the slice position A2, and a lung field area is extracted from the image of the slice position A3 obtained by photographing the anatomical position a3 (area of the lung field area> 0). ).

  In the lower graph, an image of the slice position B1 obtained by photographing the anatomical position b1, an image of the slice position B2 having a slice interval of 5 mm from the slice position B1, and a slice interval of 5 mm from the slice position B2 are shown. A lung field region is not extracted from an image at a certain slice position B3, while there is a slice interval of 5 mm from the slice position B3, and a lung field region is extracted from an image at slice position B4 obtained by photographing the anatomical position b4. Yes.

  In a general matching method, for example, the slice position B10 is a slice position having a lung field area approximate to the area of the lung field area extracted from the image of the slice position B10, for example, slice positions A4, A5, A6, A7. That is, it is matched with a plurality of slice positions, and the accuracy of aligning the anatomical positions in the body axis direction with current and past data is not sufficient.

  Next, in the reference slice image setting means 31 included in the slice matching means 22 of the computer 12 shown in FIG. 1, the past and current slice image groups are adjacent to the non-extracted slice images of the lung field region, and The slice image from which the lung field region is extracted by the lung field region extracting means 21 is set as the reference slice image. Then, each reference slice image is regarded as an image of the imaging start point of the lung field region. For example, from the graph shown in the middle part of FIG. 2, the image at the slice position A3 is set as the past reference slice image in the past slice image group, while from the graph shown in the lower part, the image at the slice position B4 is The current reference slice image in the slice image group is set, and the images at the slice positions A3 and B4 are regarded as images of the respective imaging start points in the lung field region.

  In the inter-slice distance calculating means 32 included in the slice matching means 22, the past reference slice image set by the reference slice image setting means 31 and other slice images having a lung field region included in the past slice image group. The distance between slices is calculated, and the distance between slices between the current reference slice image and another slice image included in the current slice image group is calculated. For example, since the image at the slice position A3 is regarded as the past reference slice image in the graph shown in the middle of FIG. 2, the inter-slice distance calculation unit 32 determines the slice position A3 and the slice positions A4, A5, A6,. In the graph shown in the lower part, the image at the slice position B4 is regarded as the current reference slice image. Therefore, the slice distance calculation unit 32 determines the slice position B4 and the slice position. The distance between slices with B5, B6, B7,... Is calculated.

  FIG. 3 is a graph showing the relationship between the distance between slices from the reference slice image to other slice images included in the slice image group and the area of the lung field region.

  From the graph shown in the middle of FIG. 2, the image at the slice position A3 is set as a past reference slice image, and the slice image A4, which is a slice image having a lung field region included in the past slice image group from the slice position A3. The relationship between the distance between slices to A5, A6,... And the area of the lung field is shown as a graph in the upper part of FIG. On the other hand, from the graph shown in the lower part of FIG. 2, the image at the slice position B4 is set as the current reference slice image, and the slice image that is a slice image having a lung field region included in the current slice image group from the slice position B4. The relationship between the distance between slices to B5, B6, B7,... And the area of the lung field region is shown as a graph in the lower part of FIG.

  Next, when selecting an image at a slice position that matches the slice position of the target slice image in the current slice image group from the past slice image group, first, the slice matching means 22 of the computer 12 shown in FIG. In the slice image selection means 33 included in the above, the inter-slice distance between the slice image of interest and the current reference slice image is read. Then, a slice image in which the inter-slice distance with the past reference slice image is equal to the read inter-slice distance is selected from the past slice image group.

  Therefore, even if the imaging start point is different between the past and current chest imaging, or the slice interval is different between the past and current chest imaging, the lung field is changed from the past and current slice image groups by the transition of the area of the lung field. A reference slice image serving as an end of the region is set, and slice position matching can be performed based on the inter-slice distance from the past reference slice image and the inter-slice distance from the current reference slice image. As a result, it is possible to accurately acquire a past slice image at a slice position that matches the slice position of the slice image of interest in the current slice image group.

  For example, in the graph shown in the lower part of FIG. 3, in the graph shown in the upper part, the past image of the slice position to be matched with the slice position B6 that is 10 mm from the slice position B4 of the current reference slice image. The image at the slice position A4 that is 10 mm from the slice position A3 of the past reference slice image is selected. If there is no image at a slice position with a distance of 10 mm between slices from the slice position A3 of the past reference slice image, the image is selected by matching a slice position where the distance between slices is approximately 10 mm from the slice position A3. Alternatively, the image may not exist at the corresponding slice position.

  Next, the image display form setting unit 35 of the display unit 13 shown in FIG. 1 applies slice images included in the past and current slice image groups to a predetermined frame and displays them side by side on the CRT 36. For example, the image display form setting means 35 displays three current slice images and three anatomically corresponding past slice images on the same screen of the CRT 36, respectively, in a predetermined frame.

  FIG. 4 is a diagram showing a display example of CT images taken in the past and current examinations.

  In the display example shown in FIG. 4A, a past slice image group is displayed in the upper part of the screen, and a current slice image group is displayed in the lower part of the screen. For example, the display area in the lower part of the screen is composed of a frame that displays three slice images included in the current slice image group, and in that frame, the slice position B4 that is the current reference slice image shown in the lower part of FIG. , The image at the slice position B5, and the image at the slice position B6 are sequentially applied and displayed. On the other hand, the display area in the upper part of the screen in FIG. 4A is composed of frames for displaying three slice images included in the past slice image group, and the reference slice image shown in the upper part of FIG. An image at a certain slice position A3 and an image at a slice position A4 matching the slice position B6 are sequentially applied and displayed. When there is no image at a slice position that matches the slice position B5 included in the current slice image group, the position (dashed line in the figure) where the image at the slice position that matches the slice position B5 is to be displayed is not displayed. To do.

  On the other hand, in the display example shown in FIG. 4B, an image of the slice position A3 that is the reference slice image and an image of the slice position A3 or A4 that matches the slice position B5 are displayed in the frame of the display area at the top of the screen. The slice position B6 and the image of the matching slice position A4 are sequentially applied and displayed.

  Therefore, on the screens shown in FIGS. 4A and 4B, the slice image to be noticed in the current slice image group and the slice image and the past slice image corresponding to the anatomy are displayed together. be able to. In addition, when the examiner operates, comparative interpretation is performed while paging the current slice image on the screen and the past slice image matched with these on the left and right in the figure.

  Further, the image display form setting means 35 of the display means 13 shown in FIG. 1 includes a slice image group collected at the first examination as a past CT image, and a slice image collected at the previous examination as a past CT image. The slice images included in the group and the slice image group collected at the current examination can be applied to a predetermined frame and displayed side by side on the same screen of the CRT 36.

  FIG. 5 is a diagram showing a display example of CT images taken in the past and current examinations.

  In the display example shown in FIG. 5, past slice image groups are displayed in the upper and middle stages of the screen, and the current slice image group is displayed in the lower part of the screen. For example, the image at the slice position B4, the image at the slice position B5, and the image at the slice position B6, which are the current reference slice images shown in the lower part of FIG. 3, are sequentially applied to the frame in the display area at the lower part of the screen. Is displayed. On the other hand, the reference slice image for the first visit, the slice image for the first visit that matches the slice position B5, and the slice image for the first visit that matches the slice position B6 are sequentially displayed in the display area frame in the upper part of the screen of FIG. Is done. In addition, the previous reference slice image, the previous slice image that matches the slice position B5, and the previous slice image that matches the slice position B6 are sequentially applied to the frame in the display area in the middle of the screen.

  Further, the feature analysis means 23 of the computer 12 shown in FIG. 1 measures the feature amount of the region suspected of having lung cancer in the past and present CT images, and displays the measurement result on the CRT 36.

  According to the computer-aided imaging diagnostic system 10 shown in FIG. 1, the past and current images are acquired even when the imaging start points are different between the past and current chest imaging, and when the slice intervals are different between the past and current chest imaging. Matching of slice positions in can be performed accurately and accurately.

  FIG. 6 is a block diagram showing a second embodiment of the computer-aided image diagnostic system according to the present invention.

  FIG. 6 shows a computer-aided image diagnostic system 10A. The slice matching means 22A provided in the computer 12A of the computer-aided image diagnostic system 10A shows the lung field region change rate for each slice image included in the past slice image group. And a region change rate calculating means 41 for calculating a lung field region change rate for each slice image included in the current slice image group, and a lung field region change rate in the slice image of interest in the current slice image group, The difference between the lung field region change rate comparison means 42 for comparing the lung field region change rate for each slice image included in the past slice image group and the lung field region change rate in the slice image of interest is within a predetermined range. Slice image selection means 43 for selecting a slice image having a region change rate from a past slice image group. In the computer-aided image diagnostic system 10A shown in FIG. 6, the same parts as those in the computer-aided image diagnostic system 10 shown in FIG.

  In the lung field region change rate calculating means 41, for example, the area of the lung field region in the slice image of interest in the current slice image group and the area of the lung field region in the slice image whose position is adjacent to the slice image of interest. The lung field region change rates in two directions, the positive direction and the negative direction, are calculated. In the lung field region change rate calculating means 41, the lung field region change rate in the right direction (positive direction) in the image at the slice position A4 (the lung field region in the image at the slice position A4) is obtained from the graph shown in the middle part of FIG. Area / area of the lung field in the image at slice position A3) is “2.8”, and the lung field area change rate (A4 / A5) in the left direction (negative direction) in the figure is “0.68”, the image at slice position A5 The positive lung field region change rate (A5 / A4) is calculated as “1.5” and the negative lung field region change rate (A5 / A6) is calculated as “0.85”.

  On the other hand, in the lung field region change rate calculation means 41, the positive lung field region change rate (B5 / B4) in the image at the slice position B5 is “2.3” and the lung in the negative direction is shown in the graph shown in the lower part of FIG. The field region change rate (B5 / B6) is “0.61”, the positive lung field region change rate (B6 / B5) in the image at the slice position B6 is “1.6”, and the negative lung field region change rate (B6 / B7) is “0.74”, the positive lung field region change rate (B7 / B6) in the image at slice position B7 is “1.3”, and the negative lung field region change rate (B7 / B8) is “0.84”. The positive lung field region change rate (B8 / B7) in the slice position B8 image is calculated as “1.2”, and the negative lung field region change rate (B8 / B9) is calculated as “0.88”.

  In the lung field region change rate comparison means 42, the lung field region change rate in the slice image of interest in the current slice image group is compared with the lung field region change rate in the past slice image, and then the slice image selection means. In 43, a past slice image having a lung field region change rate in which the difference from the lung field region change rate in the slice image of interest is within a predetermined range is selected. For example, if the slice image of interest is an image at slice position B8 in the graph shown in the lower part of FIG. 2, the lung field region change rate comparison means 42 uses the lung field region change rate in the positive direction in the image at slice position B8. "1.2" and the negative lung field region change rate "0.88" are compared with the lung field region change rate in each slice image included in the past slice image group. In the slice image selection means 43, the difference between the positive lung field area change rate “1.2” and the negative lung field area change rate “0.88” in the image at the slice position B8 is within a predetermined range. A past slice image A5 having a field region change rate “1.5” and a negative lung field region change rate “0.85” is selected.

  According to the computer-aided diagnostic imaging system 10A shown in FIG. 6, the past and current images can be obtained even when the imaging start points are different between the past and current chest imaging, and when the slice intervals are different between the past and current chest imaging. Matching of slice positions in can be performed accurately and accurately.

10, 10A Computer-aided image diagnostic system 11 Database 12, 12A Computer 13 Display means 21 Lung field region extraction means 22, 22A Slice matching means 23 Feature analysis means 31 Reference slice image setting means 32 Interslice distance calculation means 33 Slice image selection means 35 Image display form setting means 36 CRT
41 Lung field region change rate calculation means 42 Lung field region change rate comparison means 43 Slice image selection means

Claims (3)

Storage means for storing a slice image group at a first imaging time and a slice image group at a second imaging time acquired by taking images at different imaging times for a specific organ of the examiner; A support diagnosis process execution unit that reads out the slice image group and the slice image group at the second imaging time and executes the support diagnosis process, and a display unit that displays a result of the support diagnosis process performed by the support diagnosis process execution unit. ,
In the support diagnosis processing execution means,
The area change rate of the specific organ is calculated for each slice image included in the slice image group at the first imaging time, and the area change of the specific organ is calculated for each slice image included in the slice image group at the second imaging time. An area change rate calculating means for calculating a rate;
A region change rate comparison means for comparing a region change rate in a slice image at a required second shooting time with a region change rate for each slice image included in the slice image group at the first shooting time;
Slice image selection means for selecting a slice image having a region change rate whose difference from a region change rate in the slice image at the required second photographing time is within a predetermined range from the slice image group at the first photographing time;
A computer-aided diagnosis system comprising: The computer-aided diagnosis system according to claim 1, wherein the region change rate calculation unit calculates a region change rate in two directions based on a difference from an area of a lung field region in adjacent slice images. The display means repeats selection of a slice image from the slice image group at the second imaging time, thereby selecting a slice image selected from the slice image group at the second imaging time, and the first imaging 2. The computer-aided diagnosis system according to claim 1, wherein a slice image selected from a group of slice images at a time is displayed so as to be pageable in conjunction with the slice image.

Images