JP4253055B2 - Computer-aided diagnosis device - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a computer-aided diagnosis apparatus for detecting an abnormal shadow candidate from a medical image and presenting it, thereby reducing the trouble of interpretation and encouraging the subsequent diagnosis work.
[0002]
[Prior art]
X-ray image, X-ray computed tomography image (hereinafter abbreviated as “CT image”) by X-ray computed tomography apparatus (X-ray CT), magnetic resonance image (hereinafter referred to as “MRI image”) by magnetic resonance imaging apparatus (MRI) A medical image such as an abbreviation) is interpreted by a specialized interpretation doctor or doctor after imaging, and a diagnosis is made by the doctor based on the interpretation result.
[0003]
Interpretation is therefore an important task where accuracy, that is, oversight or misdiagnosis of anomalies can lead to delays in early detection and subsequent inappropriate treatment of patients, careful observation and appropriate judgment, and This is one of the most difficult tasks requiring many years of experience.
[0004]
Such human interpretation has the risk that the interpretation results will vary depending on the ability of the interpreter, and it is possible to avoid this danger as much as possible and provide stable interpretation results. There is an active movement of semi-automated reading, that is, introduction of a computer-aided diagnosis apparatus (hereinafter abbreviated as “CAD apparatus”).
[0005]
The role of the CAD device is to find a candidate for abnormality from the medical image and present the candidate to the interpreting doctor. The interpreting doctor eventually refers to the candidate for abnormality presented from the CAD device, and finally The interpretation results will be compiled based on this judgment. In this way, by using a CAD device at the initial stage of interpretation, problems such as overlooking subtle contrast anomalies that are difficult to identify with the human eye can be reduced, and the judgment ability of humans can be demonstrated. It is now considered effective to leave a place to do. A typical function of such a CAD apparatus is that all abnormalities are detected from a medical image, abnormal candidates are picked up from these, and the extracted abnormal candidates are marked and displayed on the medical image. is there.
[0006]
What is important here is the judgment required to pick up abnormal candidates, and in many cases it is accumulated in the heads of many experienced interpreting physicians in terms of how this judgment is given to computers. It was based on medical knowledge. For example, when a lung cancer candidate is picked up from a chest CT image, the lung cancer candidate is present in the lung field region and has a CT value that is higher than the CT value of the aerobic alveolar region in the lung field region on average. The computer uses a judgment mechanism in the interpretation doctor's head that the shape is approximate to a circle, and a computer algorithm is set up along this mechanism.
[0007]
In other words, the lung field region is first identified by threshold processing or the like for the target image, a small region having a CT value higher than the surroundings is extracted from the identified region, and the circularity of the small region is determined. This is a procedure of calculating a parameter defined as representing and leaving a small region that is judged to have high circularity.
[0008]
Of course, various feature quantities other than circularity are calculated and used for determination. The amount that is the reference for inversion for each feature amount is determined in advance according to the purpose of the detection algorithm or the target abnormality, and for example, a lung cancer candidate in a size range corresponding to a circle with a diameter of 5 mm to 30 mm is detected. In the algorithm, the area of the small area is calculated, and if it is between the areas of the circles of the two sizes, it is left, and if it is out of the range, it is excluded.
[0009]
The small region finally left after some of these processes is stored as a lung cancer candidate in the storage area in the CAD device in association with the medical image. When the interpreting doctor interprets the CT examination image, the position information of all lung cancers detected by the CAD device is read from the storage area according to the doctor's instruction, and the corresponding image has a circle centered on the position. A display method has been implemented, such as displaying in a superimposed manner.
[0010]
In addition, in the case of a plurality of positionally continuous tomographic images such as CT examination and MRI examination, the same abnormal region is often seen in two or more images. The circles were displayed at almost the same position in the continuous two to three tomographic images.
[0011]
The accuracy of the algorithm for detecting and classifying abnormalities has been improved by recent research. However, in the case of detecting lung cancer candidates in a chest CT image, a CT image for one examination covering the entire lung is used. In this situation, several false positive candidates are detected by the CAD device.
[0012]
In the above-mentioned conventional CAD device, since the number of false positive candidates per test has not reached a sufficiently low level, the interpreting doctor actually detects each of the abnormal candidates detected and displayed by the CAD device many times. It was necessary to make a decision as to whether or not it was a candidate, which was a time-consuming task.
[0013]
There is not necessarily a uniform standard regarding the criteria for determining what kind of shadow the interpreting doctor recognizes as abnormal. For example, in the case of a lung cancer candidate in a chest CT image, if the diameter is about 5 mm or more, it is determined as a candidate. Some interpretation doctors also accept candidates as candidates if they are 10 mm or more. Further, since it is very unlikely that a large abnormality will be overlooked, there are some interpretation doctors who do not need to display a candidate having a size of about 20 mm or more even if the CAD apparatus detects it.
[0014]
However, since the conventional CAD apparatus displays all the detected lung cancer candidate regions, it cannot respond to various demands that differ depending on the individual interpreting doctor. Was often annoying.
[0015]
In addition, when interpreting a plurality of tomographic images constituting a multi-slice obtained by a CT apparatus, an MRI, or the like, when an abnormality is found in a certain image, it is performed to observe almost the same position of adjacent images. In a conventional CAD apparatus, a mark “◯” indicating that a lung cancer candidate has been detected is often displayed at almost the same position in adjacent CT images. Therefore, a detection abnormality displayed in one of the images is detected. ○ indicates that a candidate for lung cancer was detected at almost the same position in the adjacent image that was observed immediately after the adjacent image was also observed when it was judged whether it should be a lung cancer candidate by referring to the position of the candidate Is displayed together with the image, which is also annoying and takes time for interpretation.
[0016]
[Problems to be solved by the invention]
A first object of the present invention is to provide a computer-aided diagnosis apparatus that can reduce the burden on the determination work for selecting and collecting abnormal candidates.
A second object of the present invention is to provide a computer-aided diagnosis apparatus capable of improving the efficiency of interpretation work for a plurality of tomographic images constituting a multi-slice.
[0017]
[Means for Solving the Problems]
  The present invention includes means for detecting a candidate for an abnormal shadow from a medical image, means for selecting a specific candidate as a display target based on a morphological feature amount from the candidates for the abnormal shadow, and the display target. It comprises a means for displaying a selected specific candidate and an operation means for setting or changing the morphological feature amount.
[0018]
According to the present invention, even if the morphological features to be recognized as abnormality candidates are different depending on the doctors, only the abnormality candidates having the feature amounts that match the doctor's criteria can be displayed, so that the doctor recognizes that they are unnecessary. Therefore, it is possible to reduce the burden on the determination work of selecting and collecting abnormal candidates by the doctor.
[0019]
In addition, since the type of the morphological feature is set or changed, even when two or more doctors share one CAD device, the type of feature that matches the criteria of each doctor is selected. Therefore, the abnormal candidates that each doctor recognizes as unnecessary can be prevented from being displayed, and the burden of the collecting work by the doctors can be reduced.
[0020]
According to a fourth aspect of the present invention, an abnormal mass is determined based on means for detecting an abnormal shadow from each of a plurality of medical tomographic images constituting a multi-slice and continuity between the slices of the abnormal shadow. Means for selecting one specific slice for each of the determined abnormal masses, and means for displaying a medical tomographic image corresponding to the selected slice.
[0021]
According to the present invention, only a tomographic image corresponding to one specific slice is displayed for a single abnormal region recognized as a lump. Compared to the conventional case of displaying a tomographic image, the doctor does not have to perform confirmation work many times even though the abnormality is originally the same, and the burden of the confirmation work can be reduced.
[0022]
According to the present invention, as shown in claim 7, a means for detecting an abnormal shadow from each of a plurality of medical tomographic images constituting a multi-slice, a means for creating a three-dimensional image from the plurality of medical tomographic images, And a means for displaying the detected abnormal region so as to overlap the three-dimensional image.
[0023]
According to this invention, the doctor can observe all of the abnormal areas detected from the image for one examination at a time, and can reduce the burden of the confirmation work.
Further, as shown in claim 8, since it is easy to specify from which tomographic image the abnormal region displayed superimposed on the three-dimensional image is extracted, the doctor can reduce the burden of confirmation work. it can.
[0024]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, a computer-aided diagnosis apparatus according to the present invention will be described with reference to the drawings according to a preferred embodiment.
(First embodiment)
FIG. 1 is a block diagram showing the configuration of a computer-aided diagnosis apparatus (CAD apparatus) according to the first embodiment. As shown in FIG. 1, the CAD apparatus according to the present embodiment includes an image input unit 1, an image storage unit 2, an abnormality candidate detection unit 4, a detection result storage unit 5, an instruction input unit 6, a display determination information storage unit 7, The detection result display determination unit 8, the display unit 3, and a control unit that controls the movement of the entire device (not shown) are configured. In FIG. 1, arrows between blocks indicate the main data flow.
[0025]
The image input unit 1 is connected to a network (not shown), and directly or from a medical image interpolation communication from an image acquisition apparatus such as an X-ray computed tomography apparatus (X-ray CT) or a magnetic resonance imaging apparatus (MRI) via the network. It has a function as an interface for inputting medical images transferred from an image storage device such as a system (PACS). The image storage unit 2 is equipped with a storage medium such as a large-capacity magnetic disk or magneto-optical disk in order to store medical image data input from the image input unit 1.
[0026]
The abnormality candidate detection unit 4 processes the medical image stored in the image storage unit 2 with any conventional general algorithm to detect abnormal shadow candidates included in the medical image. This is provided for calculating the area of the morphological feature amount of each detected abnormality candidate. As shown in FIG. 2, the detection result storage unit 5 indicates the position of the abnormal candidate detected by the abnormal candidate detection unit 4 on the image and the calculated morphological feature amount, as shown in FIG. It is provided for associating with numbers and storing them as abnormality candidate information tables.
[0027]
The detection result display determination unit 8 displays from among all abnormality candidates detected by the abnormality candidate detection unit 4 in accordance with display determination reference information stored in a storage medium such as a magnetic disk of the display determination information storage unit 7. It is provided to determine and select an abnormality candidate (a specific abnormality candidate) to be provided to. The criteria for display determination can be changed or set freely by the interpretation doctor by operating various input devices such as a mouse, a keyboard, and a touch panel provided in the instruction input unit 6.
[0028]
The display unit 3 is equipped with a display device such as a cathode ray tube (CRT) or a liquid crystal display device, and displays a medical image and is determined to be displayed on the medical image by the detection result display determination unit 8. The position of the selected specific abnormality candidate is presented by overlaying a mark on the coordinates of the abnormality candidate.
[0029]
Next, the operation of the CAD apparatus according to the first embodiment configured as described above will be described using an example in which a lung cancer candidate is detected from a chest CT image and the detection result is displayed. First, an operation from input of image data to detection of a detection result through a lung cancer candidate detection process will be described.
[0030]
First, image data transferred over the network from the X-ray computed tomography apparatus arrives at the image input unit 1. The image data is sent to the image storage unit 2 and stored therein. The image data includes image supplementary information such as a patient name, examination identification number, and image number, and a distribution of CT values (pixel values).
[0031]
The image data stored in the image storage unit 2 is read by the inspection unit and sent to the abnormality candidate detection unit 4 one by one. When receiving the image data, the abnormality candidate detection unit 4 executes a lung cancer candidate detection process. There is no restriction on the algorithm of this processing, for example,
"Lung cancer screening support system using helical CT images for mass screening" by Toshioka et al. (JAMIT Frontier'97, Proceedings of Medical Imaging Engineering, pages 111-116, January 1997)
The method described in is adopted.
[0032]
When the lung cancer candidate detection process for one image is completed, the data associated with the image, the coordinates of the detected lung cancer candidate, and the data regarding the calculated morphological feature amount are sent to and stored in the detection result storage unit 5. As this morphological feature amount, for example, the effective diameter (the diameter of a circle having the same area as the abnormal shadow region of the lung cancer candidate), the area of the abnormal shadow region of the lung cancer candidate, the diameter of the circumscribed circle of the abnormal shadow region of the lung cancer candidate, The circumscribed rectangle area of the abnormal shadow region of the lung cancer candidate, the circularity of the abnormal shadow region of the lung cancer candidate, the degree of unevenness of the edge of the abnormal shadow region of the lung cancer candidate, and the like can be considered. These are "mm", "mm² "Calculated in units."
[0033]
The detection result storage unit 7 is supplied with image supplementary information, lung cancer candidate coordinates, and feature amount data from the abnormality candidate detection unit 4, and extracts the patient name, examination identification number, and image number from the image supplementary information. The candidate coordinates and feature quantity data are registered and stored in the abnormality candidate information table. An example of the contents of the stored abnormality candidate information table is shown in FIG. The first six lines indicate that one lung cancer candidate region is detected from each of the six images of the examination with the examination identification number “123456” of the patient “Nippon Taro”.
[0034]
Next, an operation in which an interpreting doctor interprets a CT image using the CAD device will be described. Here, it is assumed that image interpretation itself is performed by observing a film on which an image is printed, as is conventionally done.
[0035]
First, when an interpreting physician inputs a display request for a lung cancer candidate detection result via the instruction input unit 6, the abnormality candidate information table stored in the detection result storage unit 5 is loaded into the control unit under the control of the control unit, The patient name, examination identification number, and imaging date are displayed on the display unit 3 as a list.
[0036]
Furthermore, a screen for changing the initial value of the display determination criterion information stored in the display determination information storage unit 7 and the value of each display determination criterion information is displayed on the touch panel. As shown in FIG. 3, this screen is realized by a known graphic user interface, and the display determination reference value can be arbitrarily set or changed by dragging the slider with the mouse. .
[0037]
Further, display determination reference information is read from the display determination information storage unit 7 to the detection result display determination unit 8 in accordance with an instruction from the control unit. The display determination reference information is sent to the detection result display determination unit 8 every time a change is made.
[0038]
Subsequently, the image interpretation doctor proceeds with film interpretation according to the following procedure. Here, it is assumed that the CT examination film having the examination identification number “123456” of the patient “Nippon Taro” is being read.
[0039]
(Step 1) When the interpretation doctor completes the interpretation of the film, the examination of the examination identification number “123456” of the patient “Nippon Taro” is displayed from the list of patient names, examination identification numbers, and photographing dates displayed. Select. When the inspection selection is input, the control unit controls to display an image corresponding to the inspection with the inspection identification number “123456”. Here, the display unit 3 displays the images one by one. To do.
[0040]
(Step 2) Subsequently, according to an instruction from the control unit, the detection result storage unit 5 reads out the abnormality candidate information of the image with the image number “6” of the selected inspection identification number “123456” from the abnormality candidate information table, The result is sent to the detection result display determination unit 8.
[0041]
(Step 3) The detection result display determination unit 8 is sent from the abnormality candidate information of the image number 6 of the examination with the examination identification number “123456” sent from the detection result storage unit 5 and the display judgment information storage unit 7. When the display determination criterion information is received, the display determination criterion is compared with the feature amount included in the abnormality candidate information for each display determination criterion. As the morphological features used for the display determination criteria, one type may be used, or two or more types may be arbitrarily combined. For example, when the effective diameter of the abnormality candidate is described as an example, whether or not the effective diameter of the abnormality candidate included in the abnormality candidate information is included within the range of the display determination reference value (at this time, a minimum of 5 mm and a maximum of 30 mm). Determine.
[0042]
In this determination, an abnormality candidate whose effective diameter is between 5 mm and 30 mm as a reference is selected as a display object, and an abnormality candidate whose effective diameter is less than 5 mm or exceeding 30 mm is excluded from the display object. When the determination regarding all the abnormal candidates of the image is completed, the detection result display determination unit 8 notifies the control unit of the image number of the CT image including the lung cancer candidate selected as the display target. In addition, the detection result display determination unit 8 sends the coordinate values of the lung cancer candidates selected as display targets to the display unit 3.
[0043]
(Step 4) Upon receiving the notification from the detection result display determination unit 8, the control unit instructs the image storage unit 2 to read out the CT image having the image number 6 including the lung cancer candidate to be displayed. Then, the image data of the image number 6 of the examination identification number “123456” is sent from the image storage unit 2 to the display unit 3. If there is no lung cancer candidate to be displayed, the process returns to (Step 2), and processing for display relating to the next image (image number 10) is performed.
[0044]
(Step 5) Upon receiving the abnormality candidate coordinate values sent from the detection result display determination unit 8 and the corresponding image sent from the image storage unit 2, the display unit 3 has a predetermined radius centered on the coordinate values. A circular mark is displayed over the image.
[0045]
(Step 6) The image interpretation doctor performs an operation of confirming whether or not there is an abnormality recognized as a lung cancer candidate represented in the image displayed on the display unit 3 by looking at the film.
(Step 7) The interpretation doctor clicks the next image request button displayed on the display unit 3 to display the next image. Then, the process returns to (Step 2).
[0046]
As described above, the display criterion information can be changed at any time, and the corresponding value is changed by dragging the slider on the touch panel of the instruction input unit 6 and moving it left and right. Then, the changed display determination reference information is sent to the detection result display determination unit 8, and (step 2) and subsequent steps are executed for the currently displayed image.
[0047]
In this way, the interpreting physician leaves the computer the task of detecting abnormal shadows and picking up candidates that are considered to have a high probability of being abnormal, and only sets the criteria for judging abnormalities. Since it is possible to prevent the display of abnormal candidates having morphological features that are recognized as unnecessary, it is possible to effectively coexist with a computer and to reduce the burden of determining the selection of abnormal candidates.
(Second Embodiment)
Next, a CAD apparatus according to the present invention will be described according to a second embodiment. The CAD apparatus according to the second embodiment is effective when a plurality of spatially continuous tomographic images constituting a multi-slice obtained by a CT apparatus, an MRI, or the like are to be interpreted.
This configuration is shown in FIG. In FIG. 4, components having the same functions as those in FIG. 1 are denoted by the same reference numerals, and fundamentally different components are denoted by different numerals. Here, the difference from FIG. 1 will be mainly described.
[0048]
The identity determination unit 11 conforms to the identity of abnormality candidates detected from a plurality of tomographic images constituting a multi-slice according to the same abnormality candidate determination criteria stored in the same abnormality candidate determination information storage unit 10, that is, adjacent images. This is provided to determine whether or not the plurality of lung cancer candidates detected in step 1 belong to a single lump of lung cancer.
[0049]
This determination is performed based on the distance between the coordinates of the abnormality candidate. That is, when the coordinates of the abnormality candidate detected from one of the adjacent images and the coordinates of the abnormality candidate detected from the other are close within the distance indicated in the same abnormality candidate determination criterion, these abnormality candidates are Determined to be for a single mass of lung cancer. The distance used as a reference for this determination can be freely set or changed by the interpretation doctor via the operation input unit 9, and this value is stored in the storage unit 10.
[0050]
More specifically, first, for each lung cancer candidate detected from all images for one examination, lung cancer detected from an image in which a slice is adjacent to the candidate (image having a larger image number by 1 and smaller by 1). The distance between the coordinates with the candidate is calculated, and if the distance is equal to or less than the criterion value, it is regarded as a candidate belonging to the same lung cancer. Then, among the lung cancer candidates regarded as belonging to the same lung cancer, a lung cancer candidate having a maximum morphological feature, for example, the effective diameter, is selected as a lung cancer candidate to be displayed for the lump of lung cancer. Then, the image selection unit 12 is notified of the image number of the image including the selected lung cancer candidate. The image selection unit 12 selectively reads out the image with the transmitted image number from the storage unit 2 and sends it to the display unit 3 together with the coordinates of the lung cancer candidate having the maximum effective diameter. The display unit 3 displays this image, and on the image, a mark is superimposed on the position of the lung cancer candidate with the maximum effective diameter determined to be displayed by the coordinate determination unit 11 on the coordinates of the abnormal candidate. To present.
[0051]
Next, the operation of the CAD apparatus according to the second embodiment configured as described above will be described by taking an example of detecting a lung cancer candidate from a chest CT image and displaying the detection result. First, since the operation from the input of image data to the detection result of the lung cancer candidate through the detection result is the same as the operation in the first embodiment, the description is omitted and the operation at the time of interpretation is described.
[0052]
First, when an interpreting doctor inputs a display request for a lung cancer candidate detection result to the instruction input unit 9, an abnormality candidate information table stored in the detection result storage unit 5 is read to the control unit according to an instruction from the control unit, and the patient The name, examination identification number, and photographing date are displayed on the display unit as a list.
[0053]
Furthermore, a screen for changing the reference value (distance) regarding the same abnormality candidate determination stored in the same abnormality candidate determination information storage unit 10 is displayed on the touch panel of the operation input unit 9. An example of this screen is shown in FIG. In this screen configuration, it is possible to freely change the reference value (distance) for the same abnormality candidate determination as needed by moving the slider left and right while dragging with the mouse. Further, this reference value is sent from the same abnormality candidate determination information storage unit 10 to the identity determination unit 11. Thereafter, the reference value for determining the same abnormality candidate is sent to the determination unit 11 every time it is changed.
[0054]
The doctor interprets the film according to the following procedure. Here, it is assumed that the CT examination film having the examination identification number “123456” of the patient “Nippon Taro” is being read.
[0055]
(Step 1) When the interpretation doctor completes the interpretation of the film, the examination of the examination identification number “123456” of the patient “Nippon Taro” is displayed from the list of patient names, examination identification numbers, and photographing dates displayed. Select. When the inspection selection is input, the control unit controls to display an image corresponding to the inspection with the inspection identification number “123456”. Here, the display unit 3 displays the images one by one. To do.
[0056]
(Step 2) Subsequently, according to an instruction from the control unit, the detection result storage unit 5 reads out the abnormality candidate information of all the images of the examination with the selected examination identification number “123456” from the abnormality candidate information table, and the identity determination unit 11 to send. The instruction from the control unit is given only when the interpretation doctor selects an examination.
[0057]
(Step 3) The identity determination unit 11 sends the abnormality candidate information of all the images of the examination with the examination identification number “123456” sent from the detection result storage unit 5 and the same abnormality candidate judgment information storage unit 10. When the determination criterion value is received, the inter-coordinate distance is calculated for each abnormality candidate between adjacent slice images, and if the distance between the coordinates is close within the determination reference value, the lung cancer candidates are the same lump. Memorize as a candidate belonging to lung cancer. When this process is performed until the end of the multi-slice image is reached, the effective diameters of multiple lung cancer candidates that can be regarded as the same lung cancer candidate are compared, and the lung cancer candidate that has the maximum is selected as the lung cancer candidate to be displayed. To do.
[0058]
When the determination regarding all the abnormal candidates of the image is completed, the identity determination unit 11 notifies the image selection unit 12 of the image number of the image from which the lung cancer candidate to be displayed is detected.
(Step 4) Upon receiving the notification of the image signal from the identity determination unit 11, the image selection unit 12 reads the image data from the storage unit 2 for the notified image numbers in ascending order of the image numbers. Thereafter, this instruction is regarded as an instruction to read one image every time there is an image display request from the interpretation doctor.
[0059]
(Step 5) The image selection unit 12 sends the coordinates of the lung cancer candidate with the largest effective diameter and its image data to the display unit 3.
(Step 6) Upon receiving the coordinate value of the abnormality candidate sent from the detection result display determination unit 8 and the corresponding image sent from the image storage unit 2, the display unit 3 has a predetermined radius centered on the coordinate value. A circular mark is displayed over the image.
[0060]
(Step 7) The image interpretation doctor checks whether the lung cancer candidate expressed in the image displayed on the display unit 3 is an abnormality that can be recognized as a lung cancer candidate, in light of the film.
[0061]
(Step 8) The interpretation doctor clicks the next image request button displayed on the display unit 3 to display the next image. Then, it returns to step 4.
In this way, only a tomographic image corresponding to one specific slice is displayed for a single abnormal region recognized as a lump, so that all faults that capture even a single abnormal region are displayed. Compared to the conventional case of displaying an image, the doctor does not need to perform confirmation work many times even though the abnormality is originally the same, and the burden of the confirmation work can be reduced.
(Third embodiment)
Next, a CAD apparatus according to the present invention will be described with reference to a third embodiment. Similar to the second embodiment, the CAD apparatus according to the third embodiment is effective when a plurality of spatially continuous tomographic images constituting a multi-slice obtained by a CT apparatus, an MRI, or the like are to be interpreted. Demonstrate. This configuration is shown in FIG. In FIG. 6, components having the same functions as those in FIG. 1 are denoted by the same reference symbols, and components that are fundamentally different are denoted by different symbols. Here, the difference from FIG. 1 will be mainly described.
[0062]
The detection result display image creating unit 14 creates a three-dimensional image of the target organ, for example, a wire frame image, from the tomographic image for one examination sent from the image storage unit 2, and the detection result storage unit The mark indicating the position of the abnormality candidate sent from 5 is synthesized. A three-dimensional image obtained by combining the marks is sent out and displayed on the display unit 3, and an image number of an image including an abnormality candidate corresponding to the mark is sent out and displayed on the display unit 3.
[0063]
The operation of the CAD apparatus according to the third embodiment configured as described above will be described by taking an example of detecting a lung cancer candidate from a chest CT image and displaying the detection result. Since the operation from the input of image data to the detection of the lung cancer candidate through the detection result of the lung cancer is the same as that in the first embodiment, a description thereof will be omitted, and the operation at the time of interpretation will be described.
[0064]
First, when an interpreting physician inputs a display request for a lung cancer candidate detection result to the instruction input unit 13, an abnormality candidate information table stored in the detection result storage unit 5 is read by the control unit according to an instruction from the control unit, and the patient The name, examination identification number, and photographing date are displayed on the display unit as a list.
[0065]
The interpreting physician proceeds with the interpretation of the film according to the following procedure. Here, it is assumed that the CT examination film having the examination identification number “123456” of the patient “Nippon Taro” is being read.
[0066]
(Step 1) When the interpretation doctor completes the interpretation of the film, the examination of the examination identification number “123456” of the patient “Nippon Taro” is displayed from the list of patient names, examination identification numbers, and photographing dates displayed. Select. When the inspection selection is input, the control unit instructs the image storage unit 2 to send all the images corresponding to the inspection with the inspection identification number “123456” to the detection result display image creating unit 14. The image storage unit 2 sends all images corresponding to the examination with the examination identification number “123456” to the detection result display image creation unit 14.
[0067]
(Step 2) Subsequently, according to an instruction from the control unit, the detection result storage unit 5 reads out the abnormality candidate information of all the images of the inspection with the selected inspection identification number “123456” from the abnormality candidate information table, and displays the detection result display image. The data is sent to the creation unit 14.
[0068]
(Step 3) The detection result display creation unit 14 receives the abnormality candidate information of all the images of the examination with the examination identification number “123456” sent from the detection result storage unit 5 and the same sent from the image storage unit 2. When all the images of the inspection are received, a three-dimensional image such as a wire frame is created from these multi-slice tomographic images. Further, the coordinates of the abnormality candidate information are converted into coordinates in the three-dimensional image. When the processing is completed, the detection result display image creation unit 14 sends the created three-dimensional image, the converted coordinate values, and the image number of the image where the abnormality is detected to the display unit.
[0069]
(Step 4) When the display unit 3 receives the three-dimensional image sent from the detection result display image creation unit 14, the converted coordinate value, and the image number of the image where the abnormality is detected, the display unit 3 preliminarily sets the coordinate value as a center. A spherical image of a predetermined radius is combined as a mark with a three-dimensional image, and the image is configured and displayed so that the image number of the image where the abnormal candidate is detected is displayed near each mark. The three-dimensional image is a wire frame image so that the mark is not hidden, but may be a translucent surface image or the like if the mark is not hidden.
[0070]
(Step 5) The image interpretation doctor checks whether the lung cancer candidate expressed in the image displayed on the display unit 3 is an abnormality that is recognized as a lung cancer candidate, while checking it against the film.
[0071]
(Step 6) The interpretation doctor clicks the next image request button displayed on the display unit 3 to display the next image. Then, it returns to step 1.
In this way, a large number of lung cancer candidates detected from one examination image are collectively displayed on the three-dimensional image. Therefore, the doctor can observe all of the abnormal areas detected from the image for one examination at a time, and can reduce the burden of the confirmation work.
Although the present invention has been described with reference to the preferred embodiments, various modifications can be made without departing from the spirit of the present invention.
[0072]
【The invention's effect】
According to the computer-aided diagnosis apparatus of the present invention, it is possible to reduce the burden on the determination work for selecting and collecting abnormal candidates.
According to the computer-aided diagnosis apparatus of the present invention, it is possible to improve the efficiency of interpretation work for a plurality of tomographic images constituting a multi-slice.
[Brief description of the drawings]
FIG. 1 is a block diagram showing the configuration of a computer-aided diagnosis apparatus according to a first embodiment of the invention.
2 is a diagram showing a structure of an abnormality candidate information table stored in a detection result storage unit in FIG. 1;
FIG. 3 is a diagram showing an example of an operation panel constituting the instruction input unit of FIG. 1;
FIG. 4 is a block diagram showing a configuration of a computer-aided diagnosis apparatus according to a second embodiment of the present invention.
5 is a diagram showing an example of an operation panel that constitutes the instruction input unit of FIG. 4;
FIG. 6 is a block diagram showing a configuration of a computer-aided diagnosis apparatus according to a third embodiment of the present invention.
[Explanation of symbols]
1 ... Image input unit,
2 ... Image storage unit,
3 ... display part,
4 ... abnormality candidate detection unit,
5 ... detection result storage unit,
6 ... Instruction input part,
7: Display determination information storage unit,
8: Detection result display determination unit.

Claims (3)

Means for detecting a candidate for an abnormal shadow from a medical image;
Means for selecting a specific candidate as a display target based on the morphological feature amount from the abnormal shadow candidates;
Means for displaying the specific candidate selected as the display target ;
A computer-aided diagnosis apparatus comprising: an operation unit for setting or changing the morphological feature amount.   The computer-aided diagnosis apparatus according to claim 1, wherein the operation unit includes a unit for setting or changing the type of the morphological feature.   The morphological features include the diameter of a circle having the same area as the abnormal shadow, the area of the abnormal shadow, the diameter of the circumscribed circle of the abnormal shadow, the area of the circumscribed rectangle of the abnormal shadow, the circularity of the abnormal shadow, and the unevenness of the edge of the abnormal shadow The computer-aided diagnosis apparatus according to claim 1, wherein the computer-aided diagnosis apparatus is at least one of the following.

Images