JP2007275318A - Image display device, image display method, and its program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To display a medical image processed so as to be suitable for the diagnosis of a medical doctor. <P>SOLUTION: A candidate area of an abnormal site is detected from the medical image 11 with a CAD program 20 and a disease type of the abnormal site estimated from the candidate site is determined with a disease type determining means 30. The disease type of the abnormal site of the candidate area, an anatomical structure where the candidate area is present, and the medical 11 by processing the same by using the processing program 50 according to the kind of a modality used for photographing the medical image is displayed in a process and display means 60. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to an image display device that displays a medical image so as to improve the diagnostic performance of the medical image, an image display method, and a program thereof.

  Conventionally, in the medical field, it is common practice to detect a diseased part by diagnosing the presence or absence of a disease by observing the radiographic image of a subject to find a lesioned part and observing the state of the affected part. Yes. However, interpretation of X-ray images depends on the experience of the interpreter and the level of image interpretation ability, and is not necessarily objective.

  Therefore, in response to this demand, various computer diagnosis support that automatically detects a candidate for an abnormal shadow in an image represented by the image data based on the image data of the subject acquired as a diagnostic image. A system (CAD: Computer Aided Diagnose) has been proposed.

  Computer diagnosis support system detects abnormal shadow candidates based on image data and displays markings on the detected parts to alert the radiographer or to make an objective judgment In order to be able to focus on the abnormal shadow candidates detected as material, the image of the local area including the part where the abnormal shadow candidates are detected is extracted from the whole image and displayed so that the image reader's observation consciousness can be concentrated. By doing so, it is possible to prevent oversight or misunderstanding by the interpreter as described above and improve diagnostic performance (for example, Patent Document 1).

  In recent years, three-dimensional image data (three-dimensional voxel data) configured as an aggregate of a plurality of slice image data using CT (Computed Tomography), MRI (Magnetic Resonance Imaging), or the like. Has come to be obtained. Such three-dimensional image data has a large amount of information, and a cross-sectional image generated based on the three-dimensional image data and a three-dimensional display are performed. Although the diagnosis can be performed, the amount of interpretation is enormous, the burden on the doctor who observes the image and makes the diagnosis is very large, and there are problems unique to the tomographic image.

  For example, when a doctor actually makes a diagnosis by looking at the detected abnormal shadow candidates, it is important to observe the position of the detected abnormal shadow candidates and the surrounding tissue state. In some cases, it is not possible to know which slice position is being observed. Therefore, when an abnormal shadow candidate is detected, a circular marker indicating the position of the detected candidate is marked on the tomographic image and displayed, and the relationship between the image of the lung field viewed from the front and the slice position is shown. There are some which display a marker indicating the position of a lesion candidate point on the chest front image to make it easier to understand a three-dimensional position (for example, Patent Document 2).

On the other hand, since 3D voxel data obtained by CT, MRI, etc. has 3D information, organs such as blood vessels and livers are extracted and displayed three-dimensionally, and images are displayed along the cross section of the blood vessels. In addition, the site of the disease can be observed in detail.
JP-A-8-294479 JP 2001-87228 A

  As described above, in recent years, CAD has been frequently used, and as a result, it has become possible to automatically detect a part that appears to be an abnormal shadow from various medical images obtained by photographing a subject.

  However, the automatically detected site is simply a region that is a candidate for an abnormal shadow, and whether or not it is a true lesion site is diagnosed by observing a medical image taken by an interpreting doctor or the like. There is a need. In addition, it may be difficult to make a diagnosis simply by displaying an image of a local area or making the three-dimensional position of a lesion easy to understand, as in the past. In many cases, it is more effective to perform diagnosis by processing an image using a processing method that seems appropriate. For example, in liver cancer, an image obtained by extracting a blood vessel or liver so that only a specific tissue such as a blood vessel or liver can be observed from 3D voxel data obtained by photographing the abdomen, or a soft part from an image obtained by photographing the chest It is effective to process and display an image suitable for diagnosis with a gradation that makes it easy to observe only the image.

  Therefore, in view of the above circumstances, an object of the present invention is to provide an image display apparatus, an image display method, and a program for automatically processing and displaying a medical image so as to obtain an image suitable for diagnosis. To do.

An image display device of the present invention includes an image storage unit that stores a medical image obtained by photographing a subject,
Candidate area detecting means for detecting a candidate area of an abnormal part from the medical image;
Disease type determination means for determining a disease type of an abnormal site estimated from the candidate region;
A display method storage means for storing a table that predetermines a display method corresponding to a disease type of an abnormal part, an anatomical structure in which the abnormal part exists, and a type of modality used for imaging;
Processing means for processing the medical image corresponding to the display method;
A display method according to the disease type of the abnormal region of the candidate region, the anatomical structure in which the candidate region exists, and the type of modality used to capture the medical image is acquired from the table, The image processing apparatus includes a processing display unit that processes and displays the medical image using a processing unit corresponding to a display method.

The image display method of the present invention includes a candidate region detection step of detecting a candidate region of an abnormal part from a medical image obtained by photographing a subject,
A disease type determination step of determining a disease type of an abnormal site estimated from the candidate region;
The disease type of the abnormal part of the candidate area, the anatomical structure in which the candidate area exists, and the display method according to the type of modality used for photographing the medical image, the disease type of the abnormal part, and And obtaining the display method corresponding to the anatomical structure in which the abnormal site exists and the type of modality used for imaging from a predetermined table, and using the processing means corresponding to the display method, the medical device And a processing display step for processing and displaying the image.

The program of the present invention is a computer,
Image storage means for storing a medical image obtained by photographing a subject;
Candidate area detecting means for detecting a candidate area of an abnormal part from the medical image;
Disease type determination means for determining a disease type of an abnormal site estimated from the candidate region;
A display method storage means for storing a table that predetermines a display method corresponding to a disease type of an abnormal part, an anatomical structure in which the abnormal part exists, and a type of modality used for imaging;
Processing means for processing the medical image corresponding to the display method;
A display method according to the disease type of the abnormal region of the candidate region, the anatomical structure in which the candidate region exists, and the type of modality used to capture the medical image is acquired from the table, The processing means corresponding to the display method is used as a processing display means for processing and displaying the medical image.

  The “abnormal part” refers to a part in which a shadow that does not appear in normal tissue appears due to a disease or the like on an image obtained by photographing a subject. Further, different characteristics appear in the “abnormal part” depending on the disease or the like.

  “Abnormal site candidate area” refers to an area that appears on the image of the subject with features that appear to be abnormal shadows, and is an area that has characteristics close to abnormal shadows even if it is a normal tissue image. including.

  The “region where the candidate region appears” includes not only the candidate region of the abnormal part but also the vicinity of the candidate region. The “anatomical structure existing in the area where the candidate area appears” refers to an organ or bone existing in the area where the candidate area appears.

  “A table of predetermined display methods according to the disease type of the abnormal part, the anatomical structure where the stomach part exists, and the type of modality used for imaging” is “the disease type of the abnormal part, stomach Depending on the anatomical structure where the site exists and the type of modality used for imaging, the medical image to be displayed is processed in order to display it with an appropriate display method from among various display functions. This is a table in which the processing to be performed and its parameters are set.

  According to the present invention, a plurality of CADs and a plurality of display functions are provided, and the disease type of the abnormal part detected using CAD, the anatomical structure in which the abnormal part exists, and the imaging are used. It is possible to improve the diagnostic performance of an observer such as a doctor by setting a display method using an appropriate display function corresponding to the type of modality that has been used and automatically processing and displaying the medical image. it can.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a diagram showing a schematic configuration of an image display apparatus according to the present embodiment.

  The image display device 1 includes an image storage unit 10 that stores a medical image 11 obtained by photographing a predetermined part of a subject to be inspected, a plurality of CAD programs 20 that detect candidate regions of abnormal parts from the medical image 11, and a CAD 20 The disease type determination means 30 for determining the disease type of the abnormal part estimated from the detected candidate region, the type of abnormal shadow, the anatomical structure existing in the region where the abnormal shadow appears, and the imaging Display method storage means 40 for storing a table 41 in which a display method corresponding to the type of modality used is set in advance, a plurality of processing means 50 for processing the medical image 11 corresponding to the display method, and candidate regions Abnormal site disease type, anatomical structure in which candidate region exists, and display method according to the type of modality used for imaging medical image 11 Obtains from the table 41 includes a processing display means 60 for displaying by processing the medical image 11, display means 70 such as a CRT using a processing means 40 corresponding to the display method.

  The image storage means 10 is a memory that stores a medical image 11 input via an interface (not shown).

  A program (hereinafter simply referred to as CAD) 20 that performs CAD processing includes candidate area detection means 21 that detects a candidate area of an abnormal region from the medical image 11. Each CAD 20 has a function of detecting a candidate region of an abnormal site related to each disease by the candidate region detection means 21 according to one disease, and further calculating the disease likelihood and malignancy of the candidate region. For example, as CAD 20, CAD for detecting lung cancer or emphysema from medical image 11 obtained by imaging the chest, CAD for searching for aneurysm or stenosis of blood vessels as abnormal blood vessels, CAD for extracting calcification of blood vessels, and abdomen There is a CAD for detecting liver cancer or the like from the medical image 11. A plurality of CADs 20 are provided in the image display device 1, and the CAD 20 suitable for analysis is selected according to each part of the subject imaged in the medical image 11. Specifically, CAD 20 of the following method can be used.

  For example, the CAD 20 that detects lung cancer detects lung cancer based on the appearance of a white area and white area of lung cancer on a simple X-ray image obtained by imaging the chest. The candidate area detection unit 21 performs enhancement processing on a region where density gradient vectors of pixel values are concentrated, extracts a white and high-brightness region from the enhanced image, the shape, size, pixel value in the region, and the like. Using the statistic of, the region considered to be lung cancer is taken as a candidate region, and the malignancy of the candidate region is calculated to determine whether the cancer is lung cancer according to whether the malignancy is high (details) Is disclosed in, for example, JP-A-2005-198890).

  Further, the CAD 20 for detecting breast cancer detects breast cancer based on the fact that a breast or liver mass appears at a concentration different from that of surrounding normal tissue. Candidate area detection means 21 detects a candidate area that is likely to be cancer using the density gradient of the pixel value, calculates the malignancy of the candidate area, etc., and determines whether the cancer is breast cancer depending on whether the malignancy is high. (For details, see, for example, JP-A-8-294479).

  In the above-mentioned JP-A-8-294479 and JP-A-2005-198890, what has been imaged by a simple X-ray imaging apparatus such as a CR apparatus is described, but on a CT image (tomographic image) imaged by CT. In addition, since an abnormal part such as cancer appears at a concentration different from the surrounding concentration, the candidate region detecting means 21 detects the candidate region of the abnormal region from the change in the concentration gradient and the concentration profile, and the malignant region is detected. Whether it is cancer or not can be determined from the degree.

  Alternatively, the CAD 20 that detects emphysema can detect pulmonary emphysema on the basis of the fact that it appears as a low-concentration area on a CT image in which pulmonary emphysema is imaged as compared to a normal lung field. The candidate area detection means 21 detects a low concentration absorption value area (hereinafter referred to as LAA) appearing on the CT image, and detects LAA as a candidate area of an abnormal part. LAA appears as a low-concentration area compared to normal lung fields, but the bias component and the intrabronchial air area are erroneously extracted in the processing using the threshold, so the bias component is removed and the concentration enhancement processing is performed. After detection, the area detected using the threshold is detected as a candidate area for the abnormal part (For details, refer to IEICE Technical Report vol.104, no.580, pp145-149,2005 (Tokushima University) ("Pulmonary emphysema analysis algorithm using multi-slice").

  The CAD 20 for detecting aneurysms and stenosis of blood vessels uses 3D volume data (MRI images, CT images, etc.) including vascular structures taken by MRI or CT, for automatic detection of aneurysms and stenosis. A necessary feature amount is determined, and vascular stenosis is detected using the feature amount. First, the candidate region detection means 21 obtains a blood vessel axis passing through the center of the blood vessel structure in the 3D volume data of the blood vessel structure, and detects the entire structure of the blood vessel. In addition, since blood vessels have characteristic density values in the data, blood vessels are extracted from the density values and, if necessary, are detected to fill in missing parts without detecting blood vessels. After determining the shape of the surface of the blood vessel structure by interpolating the obtained blood vessels, the path of the blood vessel serving as the skeleton is obtained through determination of the center of gravity of the blood vessel in the plane perpendicular to the blood vessel. Further, along the blood vessel path, a feature amount for identifying an abnormality such as an aneurysm or stenosis is obtained. Specifically, as feature quantities, the minimum and maximum blood vessel diameters orthogonal to the blood vessel path, the local curvature of the skeletal path, the blood vessel wall perimeter, the blood vessel wall perimeter curvature, and the path length and blood vessel diameter Alternatively, a discriminator learned using these characteristic quantities is generated using a ratio with the circumference of the blood vessel wall, and a blood vessel abnormality (aneurysm, stenosis, etc.) in the 3D volume data is generated using this discriminator. This is detected as a candidate region for an abnormal part. Further, it is determined whether or not the aneurysm or stenosis is based on the shape of such an abnormal site (for details, see, for example, JP-A-2004-329929).

  Further, the CAD 20 that detects arteriosclerosis of the coronary artery particularly in vascular diseases detects calcification (coronary artery calcification) that appears on the coronary artery based on the fact that the arteriosclerosis appears as a calcification image on the CT image. Therefore, the candidate region detection means 21 extracts the isolated region by removing the left and right lung fields from the thorax region by using the CT image having characteristic density values in the lung field, heart, and blood vessels. Further, a region obtained by removing a cardiovascular region including blood vessels such as the heart and the aorta from the follower region is extracted as a cardiac fat region, and this cardiac fat region is set as a detection target region of the coronary artery, and is increased white using threshold processing. The place where the bright spot appears is detected as a candidate region of the abnormal part, and it is determined whether or not it is arteriosclerosis (for details, refer to IEICE Technical Report vol.104, no.579, pp127-131, 2004 (Tokushima University) "Construction of coronary artery calcification detection algorithm using multi-slice CT images").

  The CAD described above is an example, and it is desirable to provide various CAD in the image display device 1. Even if the same lung cancer is detected, CAD of another algorithm may be provided.

  The disease type determination means 30 determines the disease type from the candidate region of the abnormal part detected using the above-described various CAD 20 and the determination result. For example, when diagnosing pulmonary diseases, CAD 20 for a plurality of pulmonary diseases is applied to the medical image 11, and based on the disease likelihood and malignancy of each disease obtained from the CAD 20 for each pulmonary disease, Identify disease names in the candidate area.

  The processing means (processing program) 50 is a gradation processing program for changing the gradation of a photographed image, a density processing program, a frequency band enhancement program, or three-dimensional voxel data obtained by photographing a photographed image using a CT apparatus or the like. In this case, there are a MIP processing program, a CPR display processing program, a program for extracting a specific organ based on the concentration value, and the like.

  The display method storage means 40 is a table in which the disease type of the abnormal shadow, the anatomical structure existing in the region where the abnormal shadow appears, and the display method corresponding to the type of modality used for imaging are set in advance. 41 is stored. Specifically, as shown in FIG. 2, the disease type of the abnormal shadow, the anatomical structure existing in the region where the abnormal shadow appears, the type of modality used for imaging, and the display method corresponding to these As described above, a processing program used for processing an image, parameters used for the processing program 50, and the like are set in the table 41 in advance. For example, if the disease type is lung cancer and the anatomical structure in the region where the abnormal shadow appears is a soft part, it is preferable to change the gradation so that the soft part is easy to observe. In addition, since the density of the anatomical structure that appears on the medical image 11 differs depending on the captured modality, it is necessary to have a gradation corresponding to each modality. Therefore, a gradation conversion program is set in the processing program of the table 41, and gradation parameters that allow easy observation of the soft part are set in advance according to the captured modality.

  In addition, when one medical image 11 is processed into an image suitable for diagnosis, when a plurality of types of processing are performed to make an image suitable for diagnosis, one disease A plurality of processing programs and parameters thereof may be set in the table 41 for the eyes.

  The processing display means 60 is a display method according to the disease type of the abnormal part of the candidate area detected by the CAD 20, the anatomical structure in which the candidate area exists, and the type of modality used for photographing the medical image 11. Is obtained from the table 41, and the medical image 11 is processed using the processing means 40 corresponding to the display method and displayed on the screen of the display means 70.

  Next, the operation of the image display apparatus 1 of the present invention will be described with reference to the flowchart of FIG.

  First, the case where lung disease is determined will be specifically described. A CT image (medical image 11) obtained by imaging the chest through the interface is input and stored in the image storage means 10 (S100), and an abnormal region is detected from the CT image using CAD for lung cancer or CAD for emphysema. A candidate area is detected (S101). The disease type determination means 30 determines the disease type based on the malignancy of the candidate area obtained from each CAD (S102). If it is determined from the determination result that the candidate area existing in the lung field area is lung cancer, the processing display means 60 indicates that the disease type is lung cancer, the anatomical hand structure is the lung soft area, and the modality is The processing program 50 and its parameters for CT are extracted from the table 41 (S103), and the gradation processing program (processing program 50) and parameters (A1, A2, A3,...) Set in the table 41 are obtained. (See FIG. 2), the gradation of the CT image is changed (S104) and displayed on the screen of the display means 70 (S105).

  When detecting a vascular disease, a chest CT image (medical image 11) is input and stored in the image storage means 10 (S100), and a vascular disease such as CAD 20 for detecting a vascular aneurysm or stenosis is detected. A candidate region of an abnormal part is detected from a CT image of the chest using a plurality of CAD 20 (S101). The disease type determination means 30 determines the disease type (S102), and if it is determined as an aneurysm (or stenosis), the CPR processing program (processing program 50) set in the table 41 is used. (S103) The processing display means 60 uses a CPR process (for example, a technique disclosed in JP-A-2002-92621) along an axis passing through the center of the blood vessel for a blood vessel existing in the candidate region of the abnormal part. (S104), and CPR is displayed on the screen of the display means 70 (S105).

  Alternatively, even when a vascular disease is detected, a candidate region for arteriosclerosis is detected from the chest CT image by the CAD 20 that detects arteriosclerosis (S101), and the disease type determination means 30 determines that the disease is arteriosclerosis. (S102), using the MIP processing program (processing program 50) set in the table 41, the processing display means 60 applies MIP processing (for example, Japanese Patent Laid-Open No. 7-303647) to blood vessels existing in the candidate region of the abnormal region. (S104) and MIP display on the screen of the display means 70 (S105).

  In liver cancer, it is important to determine what kind of blood vessels exist not only in the liver where the cancer exists but also in the surroundings where the cancer exists. Accordingly, an abdominal CT image (medical image 11) is input and stored in the image storage means 10 (S100), and a candidate region for liver cancer is detected from the CT image using the CAD 20 for detecting liver cancer (S101). When the disease type determination means 30 determines that the cancer is liver cancer based on its malignancy (S102), the volume rendering processing program (processing processing program 50) set in the table 41 is used (S103) to process Volume obtained by performing volume rendering processing (a method disclosed in Japanese Patent Application Laid-Open No. 2004-33757) on the CT image by the display unit 60 and extracting the liver and blood vessels near the liver where the candidate area exists. A rendering image is generated (S104) and displayed on the screen of the display means 70 (S105).

  As described above in detail, it is possible to facilitate diagnosis by changing the display method according to the type of disease detected using CAD and the organ in which the disease exists.

Schematic configuration of display device An example of a table in which a processing program corresponding to a disease type, an anatomical structure in which an abnormal site exists, and a type of modality used for imaging is set Flow chart representing operation of image display device

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Image display apparatus 10 Image memory | storage means 11 Medical image 20 CAD program 21 Candidate area | region detection means 30 Disease type determination means 40 Display method memory | storage means 41 Table 50 Processing processing means 60 Processing display means 70 Display means

Claims (3)

Image storage means for storing a medical image obtained by photographing a subject;
Candidate area detecting means for detecting a candidate area of an abnormal part from the medical image;
Disease type determination means for determining a disease type of an abnormal site estimated from the candidate region;
A display method storage means for storing a table that predetermines a display method corresponding to a disease type of an abnormal part, an anatomical structure in which the abnormal part exists, and a type of modality used for imaging;
Processing means for processing the medical image corresponding to the display method;
A display method according to the disease type of the abnormal region of the candidate region, the anatomical structure in which the candidate region exists, and the type of modality used to capture the medical image is acquired from the table, An image display apparatus comprising: a processing display unit that processes and displays the medical image using a processing unit corresponding to a display method. A candidate region detection step of detecting a candidate region of an abnormal region from a medical image obtained by photographing a subject;
A disease type determination step of determining a disease type of an abnormal site estimated from the candidate region;
The disease type of the abnormal part of the candidate area, the anatomical structure in which the candidate area exists, and the display method according to the type of modality used for photographing the medical image, the disease type of the abnormal part, and And obtaining the display method corresponding to the anatomical structure in which the abnormal site exists and the type of modality used for imaging from a predetermined table, and using the processing means corresponding to the display method, the medical device An image display method comprising: a processing display step for processing and displaying an image. Computer
Image storage means for storing a medical image obtained by photographing a subject;
Candidate area detecting means for detecting a candidate area of an abnormal part from the medical image;
Disease type determination means for determining a disease type of an abnormal site estimated from the candidate region;
A display method storage means for storing a table that predetermines a display method corresponding to a disease type of an abnormal part, an anatomical structure in which the abnormal part exists, and a type of modality used for imaging;
Processing means for processing the medical image corresponding to the display method;
A display method according to the disease type of the abnormal region of the candidate region, the anatomical structure in which the candidate region exists, and the type of modality used to capture the medical image is acquired from the table, A program that functions as a processing display unit that processes and displays the medical image using a processing unit corresponding to a display method.

Images