JP2011239797A - Image diagnosing support apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve accuracy of detecting an abnormal shadow candidate in a medical image by setting a parameter in detection of the abnormal shadow candidate on the basis of patient information.SOLUTION: By an image diagnosing support apparatus, electronic medical record information of the medical image subjected to the detection of the abnormal shadow candidate is acquired from an image management server and the parameter to be used in abnormal shadow candidate processing is set on the basis of the acquired electronic medical record information.

Description

  The present invention relates to an image diagnosis support apparatus.

  In the medical field, doctors display medical images generated in modalities such as CR (Computed Radiography) devices, FPD (Flat Panel Detector) devices, CT (Computed Tomography) devices, MRI (Magnetic Resonance Imaging) devices, etc. Interpretation diagnosis is performed to observe the state of lesions and changes over time.

  In order to support interpretation diagnosis, various techniques for detecting abnormal shadow candidates from medical images have been proposed. In order to improve the detection accuracy of abnormal shadow candidates, for example, Patent Document 1 has a plurality of detection algorithms so that an abnormal shadow candidate detection algorithm can be selected according to the type and detection purpose of abnormal shadow candidates. A diagnostic imaging support apparatus is described. Patent Document 2 describes a technique for changing processing conditions in abnormal shadow candidate detection processing according to a difference in imaging conditions.

By the way, in the detection of an abnormal shadow candidate, various parameters such as a contrast threshold when deleting a false positive candidate are used. Conventionally, parameters used for detection of abnormal shadow candidates are constant regardless of the gender of the patient, the result of an inquiry, and the like.
JP 2002-112986 A JP 2002-191585 A

  However, for example, in a medical image obtained by photographing the chest, when the gender is female, the breast is reflected in the lower lung field and the contrast between the abnormal shadow candidate and the surrounding lung field is lowered. For this reason, there is a problem that abnormal shadow candidates cannot be detected with high accuracy if the same contrast threshold value used for false positive deletion is used as the surrounding lung field. In addition, when the inquiry result is pneumonia, the density of the entire lung field is low, and therefore there is a problem that abnormal shadow candidates cannot be detected with high accuracy when using the same contrast threshold as that in the normal state.

  The subject of this invention is improving the detection precision of the abnormal shadow candidate in a medical image by setting the parameter at the time of abnormal shadow candidate detection based on patient information.

In order to solve the above-described problem, an image diagnosis support apparatus according to the first aspect of the present invention provides:
An abnormal shadow candidate detecting means for detecting an abnormal shadow candidate from a medical image;
Patient information acquisition means for acquiring patient information of a patient of a medical image to be detected by the abnormal shadow candidate detection means from a storage device that stores patient information;
Parameter setting means for setting parameters used in the abnormal shadow candidate detection means based on the acquired patient information;
Is provided.

The invention according to claim 2 is the invention according to claim 1,
The parameter setting means sets parameters used in the abnormal shadow candidate detection means based on the gender included in the patient information of the patient.

The invention according to claim 3 is the invention according to claim 2,
The medical image is a chest image;
When the gender included in the patient information of the patient is female, the parameter setting means preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lower lung region in the abnormal shadow candidate detection means. It is set to be different from the predetermined threshold value.

The invention according to claim 4 is the invention according to claim 3,
When the gender included in the patient information of the patient is female, the parameter setting means has a predetermined contrast threshold used for detecting an abnormal shadow candidate existing in a lower lung field region in the abnormal shadow candidate detecting means. The standard deviation threshold is set higher than a predetermined threshold.

The invention according to claim 5 is the invention according to claim 1,
The parameter setting means sets parameters used in the abnormal shadow candidate detection means based on the age included in the patient information of the patient.

The invention according to claim 6 is the invention according to claim 5,
The medical image is a chest image;
When the age included in the patient information of the patient is equal to or greater than a predetermined age, the parameter setting unit sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lung field region by the abnormal shadow candidate detection unit. It sets so that it may differ from a predetermined threshold value.

The invention according to claim 7 is the invention according to claim 6,
When the age included in the patient information of the patient is greater than or equal to a predetermined age, the parameter setting means predetermines a contrast threshold value used in the abnormal shadow candidate detection means for detecting an abnormal shadow candidate existing in a lung field region. Set lower than the specified threshold.

The invention according to claim 8 is the invention according to claim 1,
The parameter setting means sets parameters used in the abnormal shadow candidate detection means based on the body temperature included in the patient information of the patient.

The invention according to claim 9 is the invention according to claim 8,
The medical image is a chest image;
When the body temperature included in the patient information of the patient is equal to or higher than a predetermined value, the parameter setting unit preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a tracheal region by the abnormal shadow candidate detection unit. It is set to be different from the predetermined threshold value.

The invention according to claim 10 is the invention according to claim 9,
When the body temperature included in the patient information of the patient is equal to or higher than a predetermined value, the parameter setting means has a predetermined contrast threshold used for detecting an abnormal shadow candidate existing in the tracheal region in the abnormal shadow candidate detecting means. Set lower than the threshold value.

The invention according to claim 11 is the invention according to claim 1,
The parameter setting means sets parameters used in the abnormal shadow candidate detection means based on an inquiry result included in the patient information of the patient.

The invention according to claim 12 is the invention according to claim 11,
The medical image is a chest image;
When the inquiry result included in the patient information of the patient is pneumonia, the parameter setting unit preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lung field region by the abnormal shadow candidate detection unit. It is set to be different from the predetermined threshold value.

The invention according to claim 13 is the invention according to claim 12,
The parameter setting means is configured to determine in advance a contrast threshold value used for detection of an abnormal shadow candidate existing in a lung field region in the abnormal shadow candidate detection means when the inquiry result included in the patient information of the patient is pneumonia. Set lower than the threshold value.

  According to the present invention, parameters for detecting abnormal shadow candidates are set based on patient information, so that the detection accuracy of abnormal shadow candidates in medical images can be improved.

[Configuration of hospital system 100]
FIG. 1 shows a system configuration of the hospital system 100.
As shown in FIG. 1, the in-hospital system 100 includes a modality 10, an image management server 20, an image diagnosis support device 30, an electronic medical record input device 40, and the like. Each device has a local area network (LAN), Data is transmitted and received via a communication network N formed of a communication line such as a WAN (Wide Area Network). Each device constituting the in-hospital system 100 conforms to the DICOM (Digital Image and Communications in Medicine) standard, and communication between the devices is performed according to DICOM. The number of each device is not particularly limited.

  The modality 10 captures a part of the patient to be examined, digitally converts the captured image, generates a medical image, and outputs the medical image to the image management server 20. The modality 10 includes, for example, a CR device, an FPD device, and the like.

  The image management server 20 is a computer device that accumulates, stores and manages image data of medical images generated by the modality 10 and incidental information related to medical images. Specifically, the image management server 20 includes a storage unit 25 configured by a hard disk or the like. The storage unit 25 stores a medical image DB (Data Base) 251.

  Medical images stored in the medical image DB 251 are stored in a DICOM file format that conforms to the DICOM standard. The DICOM file is composed of an image part and a header part. Actual data of the medical image is written in the image portion, and supplementary information related to the medical image is written in the header portion. The medical image DB 251 has a medical image management table that stores supplementary information of medical images, and stores medical images so as to be searchable.

  The incidental information includes, for example, patient information, examination information, and detailed image information.

The patient information includes patient identification information for identifying the patient (for example, patient ID), various information related to the patient of the medical image such as the patient's name, sex, date of birth, and the like.
The examination information includes examination identification information (for example, examination ID) for identifying the examination, examination date, modality type, examination site, and various types of information related to the examination such as a doctor in charge.

  The detailed image information includes various information related to the medical image such as an image ID, an image generation time, and a file path name indicating the storage location of the medical image.

  When a medical image is received from the modality 10, the image management server 20 stores the received medical image in the medical image DB 251 and registers the incidental information in the medical image management table.

The storage unit 25 of the image management server 20 includes an electronic medical record information DB 252 that stores electronic medical record information input from the electronic medical record input device 40.
FIG. 2 shows an example of data storage in the electronic medical record information DB 252. As shown in FIG. 2, the electronic medical record information DB 252 stores patient information for each item such as patient ID, name, sex, age, weight, height, body temperature, medical examination result, and entry date.

The image diagnosis support apparatus 30 is an apparatus that detects abnormal shadow candidates from medical images acquired from the image management server 20. The detailed configuration of the image diagnosis support apparatus 30 will be described later.
The electronic medical record input device 40 is a device for inputting electronic medical record information such as patient information and an inquiry result.

[Configuration of diagnostic imaging support apparatus 30]
FIG. 3 shows a functional configuration of the image diagnosis support apparatus 30.
As shown in FIG. 3, the diagnostic imaging support apparatus 30 includes a CPU 31, an operation unit 32, a display unit 33, a communication unit 34, a RAM 35, a storage unit 36, and the like, and each unit is connected by a bus 37.

  The CPU 31 reads out the system program stored in the storage unit 36 and various processing programs for executing the diagnosis support processing shown in FIG. 4, develops it in a work area formed in the RAM 35, and controls each unit according to the program. To do.

  The operation unit 32 includes a keyboard having cursor keys, numeric input keys, various function keys, and the like, and a pointing device such as a mouse, and sends an instruction signal input to the CPU 31 by key operation or mouse operation on the keyboard. Output.

  The display unit 33 is configured by a monitor such as an LCD (Liquid Crystal Display) or a CRT (Cathode Ray Tube), and displays various screens, medical images, and the like according to instructions of display signals input from the CPU 31.

  The communication unit 34 includes a LAN adapter, a router, a TA (Terminal Adapter), and the like, and transmits and receives data to and from each device connected to the communication network N.

  The RAM 35 temporarily stores various programs, input or output data, parameters, and the like read from the storage unit 36 in various processes executed and controlled by the CPU 31.

The storage unit 36 includes an HDD (Hard Disc Drive), a semiconductor nonvolatile memory, or the like.
The storage unit 36 is a system program executed by the CPU 31, various programs for performing various processes including a diagnosis support process and an abnormal shadow candidate detection process, and data necessary for executing these programs (for example, FIG. 5). The parameter setting table 361) shown in FIG. Various programs are stored in the storage unit 36 in the form of computer-readable program code, and the CPU 31 sequentially executes operations according to the program code.

[Operation of diagnostic imaging support apparatus 30]
Next, the operation of the image diagnosis support apparatus 30 will be described.
FIG. 4 shows a flow of diagnosis support processing in the image diagnosis support apparatus 30. The process shown in FIG. 4 is realized by the cooperation of the CPU 31 and the diagnosis support processing program stored in the storage unit 36.

First, in accordance with the operation of the operation unit 32, a medical image (hereinafter referred to as a medical image Sa) that is a detection target of an abnormal shadow candidate is acquired from the image management server 20 (step S1).
Specifically, when a medical image Sa that is a detection target of an abnormal shadow candidate is selected from the medical image list screen displayed on the display unit 33 by the operation of the operation unit 32, a request for acquiring the medical image Sa is transmitted to the communication unit. 34 to the image management server 20. In the image management server 20, when the acquisition request for the medical image Sa is received, the requested medical image Sa is read from the medical image DB 251 and transmitted to the image diagnosis support apparatus 30. In the diagnostic imaging support apparatus 30, the medical image Sa transmitted from the image management server 20 is received and acquired by the communication unit 34. The acquired medical image Sa is stored in the RAM 35.

Next, electronic medical record information of a patient ID that matches the patient ID attached to the medical image Sa is acquired from the image management server 20 (step S2).
Specifically, a patient ID is transmitted to the image management server 20, and an acquisition request for electronic medical record information in which the patient ID matches the patient ID is transmitted. In the image management server 20, electronic medical record information in which the received patient ID matches the patient ID is read from the electronic medical record DB 252 and transmitted to the image diagnosis support apparatus 30. When there are a plurality of electronic medical record information with matching patient IDs, the electronic medical record information with the latest “entry date” is read and transmitted. In the image diagnosis support apparatus 30, the electronic medical record information transmitted from the image management server 20 is received and acquired by the communication unit 34. The acquired electronic medical chart information is stored in the RAM 35.

  Next, parameters used in the subsequent abnormal shadow candidate detection process are set based on the acquired electronic medical record information (step S3).

Here, various parameters used in the abnormal shadow candidate detection process are determined in advance and stored in the storage unit 36. However, depending on the patient's condition, detection may not be performed with high accuracy using predetermined parameters.
For example, if the patient is a woman, in the medical image taken of the chest, the breast is reflected in the lower lung area, so the density of the lower lung area is low, and the contrast between the abnormal shadow and the lung field is normal (here, Lower than male). Therefore, when the patient is a woman, if the contrast threshold used in the abnormal shadow candidate detection process is set to a predetermined threshold, the abnormal shadow candidate may not be detected with high accuracy.
For example, when the patient has pneumonia, in the medical image obtained by photographing the chest, the density of the entire lung field is lower than normal, and the contrast between the abnormal shadow and the lung field is lower than normal (normal). Therefore, when the patient has been diagnosed as suspected of having pneumonia through an interview, if the contrast threshold used in the abnormal shadow candidate detection process is set to a predetermined threshold, the abnormal shadow candidate may not be detected accurately. is there.

Therefore, in the present embodiment, parameters used in the abnormal shadow candidate detection process are set based on the electronic medical record information. The parameter setting is performed with reference to a parameter setting table 361 stored in the storage unit 36.
FIG. 5 shows an example of the parameter setting table 361 used when the examination site of the medical image Sa is the chest. In the parameter setting table 361, the contents of the electronic medical record information, which is a parameter change condition, are stored in association with the area where the parameter should be changed, the type of parameter to be changed, and the amount of change when the contents correspond to the contents. Has been. Specifically, as shown in FIG. 5, the “item” indicating the item of the electronic medical record information related to the parameter change, the “condition” indicating the content corresponding to the parameter change condition in the item, and the parameter are changed. A “change area” indicating the image area to be changed, a “change parameter” indicating the type of the parameter to be changed, and a “change amount” indicating the parameter change amount are stored in association with each other. The contrast threshold value and the standard deviation threshold value, which are parameters shown in FIG.

  The change amount of each parameter in the parameter setting table 361 is a value obtained experimentally and empirically. For example, when the content of the “sex” item is “female”, the amount of change in the contrast threshold is calculated as follows. First, true positive and false positive shadow images existing in the lower lung region of a woman are accumulated, and the contrast of each of the true positive and false positive shadows is calculated. The calculation of contrast will be described later. Next, a contrast is calculated such that the ratio of the Mahalanobis distance from the true positive group to the Mahalanobis distance from the false positive group when the contrast is a variable is 1. This contrast value is determined as a contrast threshold used for the lower lung region in the abnormal shadow candidate detection process. Then, the amount of change is calculated based on the difference between the determined threshold and a predetermined threshold. The change amounts of other change parameters are values calculated in the same manner.

Hereinafter, the process of step S3 will be described.
First, the parameter setting table 361 is referred to, and the item (“age” in FIG. 5) specified first in the “item” of the parameter setting table 361 is searched from the electronic medical record information.
Next, it is determined whether or not the content of the retrieved item of the electronic medical record information corresponds to the “condition” associated with the first “item” of the parameter setting table 361. When it is determined that the content of the retrieved item of the electronic medical record information corresponds to “condition”, the “change area”, “change parameter”, and “change amount” associated with this “condition” are parameter setting tables. The data is read from 361 and stored in the RAM 35.
Similarly, each item defined in the parameter setting table 361 is sequentially searched from the electronic medical record information, and it is determined whether or not the content of the item corresponds to “condition”. If it is determined that the “condition” is met, the “change area”, “change parameter”, and “change amount” associated with the “condition” are read from the parameter setting table 361 and stored in the RAM 35. Is done.

When the search for all items is completed, the parameters are set based on the predetermined parameters stored in the storage unit 36, the change area, the type of change parameter, and the change amount stored in the RAM 35.
For example, when the examination part of the medical image Sa is the chest and the gender of the electronic medical record information is female, the above processing causes the RAM 35 to display “change area: lower lung field, change parameter: contrast threshold, change amount. : -9 "and" change area: lower lung field area, change parameter: standard deviation threshold, change amount: +8 "are stored. Therefore, when the predetermined contrast threshold is 25 and the standard deviation threshold is 20, the contrast threshold of the lower lung field region is set to 16 and the standard deviation threshold is set to 28.
For example, when the examination site of the medical image Sa is the chest and the age of the electronic medical record information is 60 years or older, the above processing causes the RAM 35 to display “change area: lung field area, change parameter: contrast threshold, change Amount: −3 ”is stored. Therefore, when the predetermined contrast threshold is 25, the contrast threshold of the lung field region is set to 22.
For example, when the examination part of the medical image Sa is the chest and the body temperature of the electronic medical record information is 38 ° C. or higher, the above processing causes the RAM 35 to display “change area: trachea area, change parameter: contrast threshold, change amount. : -3 "is stored. Therefore, when the predetermined contrast threshold is 25, the contrast threshold of the tracheal region is set to 22.
For example, when the examination part of the medical image Sa is the chest and “pneumonia” is included in the inquiry result of the electronic medical record information, the RAM 35 stores “change area: lung field area, change parameter: contrast” by the above processing. “Threshold, change amount: −4” is stored. Accordingly, when the predetermined contrast threshold is 25, the contrast threshold of the lung field region is set to 21.
When the change area and the change parameter overlap, the parameter with the larger amount of change is set with priority. For example, in the case of a woman over 65 years old, “contrast threshold: 16” is set in the lower lung region. In the other lung field regions, “contrast threshold value: 22” is set. Predetermined parameters are set for areas other than the change area and parameters other than the change parameter even in the change area.
The set parameters are stored in the RAM 35.

When the parameter setting is completed, an abnormal shadow candidate detection process is executed (step S4).
Hereinafter, detection of an abnormal shadow candidate when the examination site is the chest will be described as an example of the process of step S4 of FIG. Here, an example of detecting a nodule-like shadow candidate from the lung field region will be described.

  FIG. 6 shows a flow of abnormal shadow candidate detection processing executed in step S4. The abnormal shadow candidate detection process is realized by the cooperation of the CPU 31 and the abnormal shadow candidate detection process program stored in the storage unit 36.

First, a lung field region is detected from the medical image Sa (step S101).
For example, a technique described in Japanese Patent Laid-Open No. 2003-6661 can be used to extract a lung field region. In this extraction method, first, an extension direction-specific contour detection mask and an extension direction-specific edge detection mask in six different directions (six directions of 0 °, 30 °, 60 °, 90 °, 120 °, and 150 °) are used. Thus, six images in which the contour extending in the direction corresponding to each contour detection mask is emphasized and three images in which the edge extending in the direction corresponding to each edge detection mask is emphasized are created from the medical image Sa. The Next, an outline image is created by selecting the pixel having the lowest density by associating the pixels of these nine images. The extension direction-specific contour detection mask is a mask for detecting a straight line extending in a specific angle direction. The extension direction edge detection mask is a mask that detects edges extending in a specific angle direction and having a specific density gradient directionality.
Next, smoothing processing is performed on the medical image Sa to create a smoothed image in which ribs and clavicles are not conspicuous. By multiplying the approximate contour image and the pixel value of the smoothed image for each corresponding pixel, the thoracic internal smoothing is performed. An image is created.
Next, the internal smoothed image of the chest is subjected to polar coordinate conversion around a reference point (a point that is approximately equidistant from the contour portion at the top of the lung field), and a template for detecting the outline of the thorax in the obtained polar coordinate converted image (clinical) The outline of the rib cage is detected by performing template matching using an average of a large number of cardio-thoracic rib contours obtained in a typical manner. Then, the lung field region is extracted by returning the detection result to the medical image Sa, performing spline interpolation and connecting with a closed curve.

The extraction of the lung field region is not limited to the above. For example, the horizontal direction and the vertical direction of the medical image Sa are sequentially scanned to create a signal value profile in each direction, and based on the inflection points in the profile. Thus, a technique for extracting a lung field region can be used.
The extracted position information of the lung field region is stored in the RAM 35.

  Next, initial detection of abnormal shadow candidates in the lung field region extracted from the medical image Sa is performed (step S102).

  Hereinafter, as an example of the processing of step S101, an example of detecting a nodular shadow candidate as an abnormal shadow candidate will be described (Shigehiko Katsagawa: “Computer-Aided Diagnosis” radiology) See Physics).

  First, an image in which the contrast of the nodular shadow is enhanced by the matched filter and an image in which the contrast is attenuated by the smoothing filter are created from the medical image Sa, and a difference image is created by taking the difference between the corresponding pixel values of both. Is done. The matched filter is designed so that, for example, the contrast-to-noise ratio of a nodular shadow having a diameter of 9 mm is maximized. In the difference image, the contrast of a normal chest structure such as a rib is lowered, and the contrast of a nodular shadow is enhanced.

  Next, threshold processing is performed a plurality of times on the difference image. The threshold processing is processing for converting a pixel value equal to or greater than a predetermined threshold to 1 and converting the others to 0 to binarize. The threshold value is a value expressed by the area ratio of the histogram of the pixel values of the difference image. For example, the threshold processing is performed a plurality of times by changing the threshold in the upper 1% to 40% range of the histogram of the difference image. By this threshold processing, shadows such as abnormal shadows, rib cages, and blood vessel shadows are extracted. Of the extracted shadows, shadows other than the lung field region extracted in step S101 are deleted.

  Next, feature quantities such as the effective diameter, circularity, irregularity, etc., of the shadows extracted by multiple threshold processings are calculated. FIG. 7 shows formulas for calculating the effective diameter (indicated by D in FIG. 7), circularity, and irregularity. The effective diameter is the diameter of a circle having the same area as the shadow. The circularity is a ratio of the area (number of pixels) of the shadow included in the circle when the circle is superimposed on the center of gravity of the shadow. The degree of irregularity represents the degree of fraud at the edge of the shadow, and is a value obtained by subtracting from 1 the ratio of the circumference of the circle to the length of the periphery of the shadow.

Next, initial candidates for abnormal shadows are narrowed down from the extracted shadows based on the calculated feature values.
FIG. 8A shows a relationship between a change in threshold value in threshold processing and a change in effective radius of nodular shadow and blood vessel shadow extracted by the threshold processing. FIG. 8B shows the relationship between the threshold value change in the threshold value process, the nodular shadow extracted by the threshold value process, and the change in the circularity of the pulmonary blood vessel. FIGS. 8A and 8B are graphs showing the results obtained based on experimental experience. As shown in FIGS. 8 (a) and 8 (b), the effective radius of the nodular shadow changes gently as the threshold value increases, and the circularity of the blood vessel shadow has a threshold value of 16 while maintaining a relatively high circularity. % Shows a sharp increase in effective radius and a decrease in circularity.
Therefore, initial candidates for abnormal shadows can be narrowed down based on changes in effective radius and circularity in the vicinity of the threshold value of 16% in each shadow.

When the initial detection of abnormal shadow candidates is completed, false positive candidates are deleted from the detected initial candidates (step S103).
Here, in the false positive candidate deletion process, first, the parameters of each area set in the RAM 35 are referred to. When there is an area where a parameter different from the predetermined parameter is set, processing for extracting the area is performed.

For example, when the parameters of the lower lung field are changed, the lower lung field is extracted. When the parameters of the lower lung field are changed, the breasts overlap in the lower lung field, so the density is lower than that of the surrounding lung fields. Therefore, the lower lung region can be extracted by the following method.
First, for each of the left and right lung field regions, an n pixel × n pixel ROI (region of interest) is sequentially set while shifting from the bottom in the horizontal direction and the vertical direction, and the average value of the pixel values of each ROI is calculated. Next, for each ROI, the difference in the average value of the pixel values from the ROI adjacent in the vertical direction is calculated. When the calculated difference exceeds a predetermined threshold value, a pixel located at the center in the vertical direction of the ROI is extracted as a boundary of the lower lung region.

  Further, for example, when the parameters of the trachea region are changed, the trachea region is extracted. The tracheal region can be extracted using a known method. For example, as described in Japanese Patent Application Laid-Open No. 2004-8419, a technique for detecting an anatomical feature position using a plurality of artificial images representing the structure of a subject (here, a tracheal region) can be used. In this technique, a plurality of artificial images representing the trachea, each of which is accompanied by the shape information of the trachea represented by the artificial image, is stored in the storage unit 36 in advance. Then, an artificial image having a structure that substantially matches the medical image Sa is selected from the stored artificial images, and the tracheal region in the medical image Sa is extracted based on the shape information attached to the selected artificial image. .

  Next, image feature amounts (contrast and standard deviation) are calculated for each of the initial candidate regions detected in step S102.

The contrast is calculated as follows.
First, as shown in FIG. 9, a circle at a distance d3 from the center of gravity of the initial candidate region to one point on the edge is designated as the candidate edge. Next, an outer area and an inner area are set based on the distance d3. For example, as shown in FIG. 9, a circular area at a distance of 0.8d3 from the center of gravity is set as an inner area, and an area at a distance of 1.2d3 or more and a distance of 1.7d3 or less from the center of gravity is set as an outer area.
Next, an average value of pixel values is calculated for each of the outer region and the inner region. Then, the difference between the average values is calculated. The difference between the average values is the contrast.

  When the contrast and standard deviation are calculated for each initial candidate, whether each initial candidate is a true positive candidate or false based on the parameters (contrast threshold and standard deviation threshold) set in the area where each initial candidate is located. It is determined whether or not it is a positive candidate. As a result of the determination, the initial candidate determined to be a true positive candidate is finally detected as an abnormal shadow candidate, and the position information is stored in the storage unit 36 in association with the image ID. Then, the abnormal shadow candidate detection process ends.

Returning to FIG. 4, when the abnormal shadow candidate detection process is completed, the detection result is displayed on the display unit 33 (step S5). For example, the medical image Sa is displayed on the display unit 33, and an annotation or the like is displayed at the position of the detected abnormal shadow candidate.
When an instruction to end the display of the detection result is given, the diagnosis support process ends and the RAM 35 is released.

  As described above, according to the image diagnosis support apparatus 30, the electronic medical record information of the medical image that is the detection target of the abnormal shadow candidate is acquired from the image management server 20, and the abnormal shadow candidate process is performed based on the acquired electronic medical record information. Set the parameters used in.

  Therefore, parameters for detecting abnormal shadow candidates are set based on electronic medical record information such as the patient's sex and medical examination results, so that the detection accuracy of abnormal shadow candidates in medical images can be improved.

For example, because the parameters for abnormal shadow candidates are set based on the gender included in the patient's electronic medical record information, abnormal shadow candidates can be detected using parameters according to the patient's gender, improving detection accuracy. Can be made.
For example, if the medical image is a chest image and the patient's gender is female, the contrast between the abnormal shadow in the lower lung field and the surrounding area is lower than that of the male due to the reflection of the breast, and the standard deviation of the abnormal shadow increases. . In consideration of this feature, the diagnostic imaging support apparatus 30 sets a contrast threshold used for detection of an abnormal shadow candidate existing in the lower lung region lower than a predetermined threshold, and sets a standard deviation threshold in advance. Since it is set higher than the threshold value, the detection accuracy can be improved.

In addition, for example, since the parameters for the abnormal shadow candidate are set based on the age included in the electronic medical record information of the patient, the abnormal shadow candidate can be detected using the parameter according to the age of the patient, and the detection accuracy Can be improved.
For example, when the medical image is a chest image and the patient's age is greater than or equal to a predetermined age, the density of the entire lung field decreases and the contrast between the abnormal shadow and the surroundings decreases. In consideration of this feature, the diagnostic imaging support apparatus 30 sets the contrast threshold used for detecting abnormal shadow candidates to be lower than a predetermined threshold, so that the detection accuracy can be improved.

In addition, for example, parameters for abnormal shadow candidates are set based on the body temperature included in the electronic medical record information of the patient, so that abnormal shadow candidates can be detected using parameters according to the patient's body temperature, and the detection accuracy Can be improved.
For example, when the medical image is a chest image and the patient's body temperature is equal to or higher than a predetermined value, the density of the trachea region decreases and the contrast between the abnormal shadow and the surroundings decreases. In consideration of this feature, the diagnostic imaging support apparatus 30 sets the contrast threshold used for detecting abnormal shadow candidates to be lower than a predetermined threshold, so that the detection accuracy can be improved.

In addition, for example, because the parameters for abnormal shadow candidates are set based on the interview results included in the patient's electronic medical record information, the abnormal shadow candidates can be detected using parameters according to the patient's interview results. Detection accuracy can be improved.
For example, when the medical image is a chest image and the patient has pneumonia, the density of the entire lung field decreases, and the contrast between the abnormal shadow and the surroundings decreases. In consideration of this feature, the diagnostic imaging support apparatus 30 sets the contrast threshold used for detecting an abnormal shadow candidate to be lower than a predetermined threshold, so that the detection accuracy of the abnormal shadow candidate can be improved.

In addition, the description content in the said embodiment is a suitable example of this invention, and is not limited to this.
For example, in the above embodiment, the case where parameters used for abnormal shadow candidate detection are set based on the patient information of the electronic medical record information has been described as an example, but the patient information is not limited to that of the electronic medical record information. Patient information incidental to medical images may be used.

  In addition, the medical image in which the examination target site is the chest has been described as an example, but the type of medical image is not limited to this. Also, the abnormal shadow candidate detection method and parameters are not limited to those described above.

  For example, in the above description, an example in which an HDD or a semiconductor nonvolatile memory is used as a computer-readable medium of the program according to the present invention is disclosed, but the present invention is not limited to this example. As another computer-readable medium, a portable recording medium such as a CD-ROM can be applied. Further, a carrier wave is also applied as a medium for providing program data according to the present invention via a communication line.

  In addition, the detailed configuration and detailed operation of each apparatus constituting the hospital system 100 and the hospital system 100 can be changed as appropriate without departing from the spirit of the invention.

It is a figure which shows the function structural example of the system in a hospital in embodiment of this invention. It is a figure which shows the example of data storage of electronic medical chart information DB of FIG. It is a block diagram which shows the functional structure of the image diagnosis assistance apparatus of FIG. It is a flowchart which shows the diagnostic assistance process performed by CPU of FIG. It is a figure which shows the data storage example of the parameter setting table memorize | stored in the memory | storage part of FIG. It is a flowchart which shows the abnormal shadow candidate detection process performed by CPU of FIG. It is a figure for demonstrating the effective diameter, circularity, and irregularity used in the abnormal shadow candidate detection process of FIG. (A) is a figure which shows the relationship between the threshold value of the threshold process performed in the abnormal shadow candidate detection process of FIG. 6, and an effective radius, (b) is the threshold value threshold value process performed in the abnormal shadow candidate detection process of FIG. It is a figure which shows the relationship of circularity. It is a figure for demonstrating the calculation method of contrast.

Explanation of symbols

100 Hospital System 10 Modality 20 Image Management Server 25 Storage Unit 251 Medical Image DB
252 Electronic medical record information DB
30 Image Diagnosis Support Device 31 CPU
32 Operation unit 33 Display unit 34 Communication unit 35 RAM
36 Storage Unit 361 Parameter Setting Table 37 Bus

Claims (13)

An abnormal shadow candidate detecting means for detecting an abnormal shadow candidate from a medical image;
Patient information acquisition means for acquiring patient information of a patient of a medical image to be detected by the abnormal shadow candidate detection means from a storage device that stores patient information;
Parameter setting means for setting parameters used in the abnormal shadow candidate detection means based on the acquired patient information;
An image diagnosis support apparatus comprising:   The image diagnosis support apparatus according to claim 1, wherein the parameter setting unit sets a parameter used in the abnormal shadow candidate detection unit based on gender included in patient information of the patient. The medical image is a chest image;
When the gender included in the patient information of the patient is female, the parameter setting means preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lower lung region in the abnormal shadow candidate detection means. The image diagnosis support apparatus according to claim 2, wherein the image diagnosis support apparatus is set so as to be different from a predetermined threshold value.   When the gender included in the patient information of the patient is female, the parameter setting means has a predetermined contrast threshold used for detecting an abnormal shadow candidate existing in a lower lung field region in the abnormal shadow candidate detecting means. The diagnostic imaging support apparatus according to claim 3, wherein the threshold value is set lower than the threshold value, and the standard deviation threshold value is set higher than a predetermined threshold value.   The image diagnosis support apparatus according to claim 1, wherein the parameter setting unit sets a parameter used in the abnormal shadow candidate detection unit based on an age included in patient information of the patient. The medical image is a chest image;
When the age included in the patient information of the patient is equal to or greater than a predetermined age, the parameter setting unit sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lung field region by the abnormal shadow candidate detection unit. The image diagnosis support apparatus according to claim 5, wherein the image diagnosis support apparatus is set to be different from a predetermined threshold value.   When the age included in the patient information of the patient is greater than or equal to a predetermined age, the parameter setting means predetermines a contrast threshold value used in the abnormal shadow candidate detection means for detecting an abnormal shadow candidate existing in a lung field region. The image diagnosis support apparatus according to claim 6, wherein the image diagnosis support apparatus is set to be lower than a set threshold value.   The image diagnosis support apparatus according to claim 1, wherein the parameter setting unit sets a parameter used in the abnormal shadow candidate detection unit based on a body temperature included in patient information of the patient. The medical image is a chest image;
When the body temperature included in the patient information of the patient is equal to or higher than a predetermined value, the parameter setting unit preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a tracheal region by the abnormal shadow candidate detection unit. The diagnostic imaging support apparatus according to claim 8, wherein the diagnostic imaging support apparatus is set so as to be different from a predetermined threshold.   When the body temperature included in the patient information of the patient is equal to or higher than a predetermined value, the parameter setting means has a predetermined contrast threshold used for detecting an abnormal shadow candidate existing in the tracheal region in the abnormal shadow candidate detecting means. The image diagnosis support apparatus according to claim 9, wherein the image diagnosis support apparatus is set to be lower than a threshold value.   The image diagnosis support apparatus according to claim 1, wherein the parameter setting unit sets a parameter used in the abnormal shadow candidate detection unit based on an inquiry result included in patient information of the patient. The medical image is a chest image;
When the inquiry result included in the patient information of the patient is pneumonia, the parameter setting unit preliminarily sets a threshold value of an image feature amount used for detection of an abnormal shadow candidate existing in a lung field region by the abnormal shadow candidate detection unit. The image diagnosis support apparatus according to claim 11, wherein the image diagnosis support apparatus is set so as to be different from a predetermined threshold value.   The parameter setting means is configured to determine in advance a contrast threshold value used for detection of an abnormal shadow candidate existing in a lung field region in the abnormal shadow candidate detection means when the inquiry result included in the patient information of the patient is pneumonia. The image diagnosis support apparatus according to claim 12, wherein the image diagnosis support apparatus is set lower than a threshold value.

Images