JP4541988B2 - Alignment apparatus, alignment method, and program thereof - Google Patents

Description

  The present invention relates to an alignment apparatus, an alignment method, and a program for aligning medical images.

  In recent years, based on 3D image data configured as an aggregate of a plurality of slice image data acquired by CT (Computed Tomography), PET (Positron Emittion Tomography), etc., a cross-sectional image is obtained. Generation or three-dimensional display is performed to find a lesioned part, and the state of the lesioned part is observed to diagnose the presence or absence of disease and the progress. Since this three-dimensional image data has more information about the subject compared to conventional two-dimensional radiation image data, highly accurate diagnosis is possible. Furthermore, it has become possible to make a more accurate treatment plan by observing images taken between a plurality of different modalities.

  The PET image is an image taken by injecting a drug labeled with an isotope. As a drug (FDG) used in PET, glucose labeled with an isotope is used. Although cells use glucose as an energy source, cancer cells are more active than normal cells, so they have a property of taking in a lot of glucose, which is a nutrient, and FDG is also taken up by cancer tissues in large amounts. For this reason, radiation stronger than normal tissue is emitted from cancer cells. Since the amount of radiation is proportional to the amount of tumor cells taking up glucose, that is, the activity, PET can acquire an image reflecting the activity of the cancer cell portion. Thereby, the active cancer lesion can be observed from the PET image. However, since the PET image has a low spatial resolution, it becomes a blurred image. From this point, it is difficult to grasp the exact position of the lesion using only the PET image. Therefore, CT images and PET images that can accurately measure the position and morphology of organs are combined and used for diagnosis of treatment plans.

  However, although the position and form of the organ appear correctly in the CT image, the active cancer lesion appears in the PET image, but the form is blurred, so it is difficult to align the CT image and the PET image. is there. Also, between CT images and MRI images, bones are imaged in CT images, but bones are not imaged in MRI images, so alignment between CT images and MRI images is difficult. Since tissues appearing on images taken with different modalities (medical examination devices) differ in this way, it is difficult to perform alignment accurately.

  Therefore, a method has been proposed in which registration is performed using characteristic shapes captured in both images, such as lung field regions, when performing registration between images captured with different modalities (for example, non-modality). Patent Document 1).

Alternatively, a technique has been proposed in which a patient is photographed by wearing a mask having a convex protrusion that can accommodate a marker that appears on an image photographed by each modality. Since the markers appearing on the captured image differ depending on the modality, a plurality of types of markers for CT / MRI, SPECT / CT, MRI / CT, etc. are prepared. Therefore, when photographing with various modalities, the patient is wearing a mask containing a marker suitable for the modality in the projection, and the data between the different modalities is aligned based on the marker. A technique has also been proposed (for example, Patent Document 1).
22nd JCMI Nov., 2002 Medical Informatics 22 (suppl), 2002 pp706-707 Japanese Patent Laid-Open No. 2002-248083

  However, the technique of Non-Patent Document 1 needs to know a tissue effective for alignment for each combination of modalities, and a tissue effective for alignment is not always photographed. In addition, since most of the tissues used for alignment are soft tissues, they are deformed due to a change in posture and the alignment accuracy is not good.

  On the other hand, the technique of Patent Document 1 is not only burdensome to put a marker on a patient, but also when positioning with a plurality of modalities at the same time, alignment by the marker is effective. When there is an interval for imaging the PET, it is difficult to attach the patient so that the marker is at the same position.

  The present invention has been made in view of the above circumstances, and provides an alignment apparatus, an alignment method, and a program thereof that are capable of performing highly accurate alignment on images captured by different modalities. It is intended.

The alignment apparatus of the present invention is a standard that appears on an image when a plurality of anatomical structures obtained by photographing a subject using the first photographing apparatus are photographed by the first photographing apparatus. First model image storage means for storing a first standard density distribution model image expressed using density values;
A plurality of anatomical structures obtained by photographing a subject using the second photographing device are represented by using standard density values appearing on the image when the second photographing device is photographed. Second model image storage means for storing two standard density distribution model images;
First photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing device;
Second photographed image storage means for storing a second photographed image including an anatomical structure obtained by photographing the predetermined subject using the second photographing device;
First alignment means for obtaining a first corresponding position obtained by aligning corresponding anatomical structures between the first standard density distribution model image and the first captured image;
Second alignment means for acquiring a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. Corresponding position of the corresponding anatomical structure, the first corresponding position, and the corresponding position obtaining means for obtaining from the second corresponding position are provided.

The alignment method of the present invention is a standard that appears on an image when a plurality of anatomical structures obtained by photographing a subject using the first photographing device are photographed by the first photographing device. A plurality of anatomical images obtained by photographing a subject using a first model image storage means for storing a first standard density distribution model image and a second photographing device. Second model image storage means for storing a second standard density distribution model image expressed using a standard density value appearing on the image when the structure is photographed by the second photographing device; A first photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing apparatus, and the second photographing apparatus are used. The second photographed image including the anatomical structure obtained by photographing the predetermined subject In alignment device and a second captured image storage means for storing,
A first alignment step of obtaining a first corresponding position obtained by aligning a corresponding anatomical structure between the first standard density distribution model image and the first captured image;
A second alignment step of obtaining a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. And a corresponding position acquisition step for obtaining from the corresponding position of the corresponding anatomical structure, the first corresponding position, and the second corresponding position.

Further, the program of the present invention provides a computer,
A plurality of anatomical structures obtained by photographing a subject using the first photographing device are represented by using standard density values that appear on the image when the first photographing device is photographed. First model image reading means for reading out the first standard density distribution model image from first model image storage means for storing a first standard density distribution model image;
A plurality of anatomical structures obtained by photographing a subject using the second photographing device are represented by using standard density values appearing on the image when the second photographing device is photographed. Second model image reading means for reading out the second standard density distribution model image from second model image storage means for storing two standard density distribution model images;
The first photographed image is read out from first photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing device. First photographed image reading means;
The second photographed image is read out from a second photographed image storage means for storing a second photographed image including an anatomical structure obtained by photographing a predetermined subject using the second photographing apparatus. Second captured image reading means;
First alignment means for acquiring a first corresponding position obtained by aligning corresponding anatomical structures between the first standard density distribution model image and the first captured image;
Second alignment means for acquiring a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. It is made to function as a corresponding position acquisition means calculated | required from the corresponding position of a corresponding anatomical structure, a said 1st corresponding position, and a said 2nd corresponding position.

  The “first photographing apparatus” and the “second photographing apparatus” are different photographing apparatuses.

  The “first standard density distribution model image”, the “second standard density distribution model image”, the “first photographed image”, and the “second photographed image” are the photographing devices (the first photographing device or the second photographing image). Depending on the imaging device, anatomical structures such as bones and built-in tissues are composed of data having specific density values.

  Further, the “density value” is not limited to the optical density value as long as it represents the density of the image.

  The “standard concentration distribution model image” is a standard image that appears on an image obtained when an anatomical structure such as each organ or bone is imaged with an imaging device (eg, modality such as CT, PET, MRI). The model image is expressed by various density values, and the model image has a different density distribution depending on the type of the photographing apparatus. For example, a three-dimensional standard density distribution model for CT is different because a three-dimensional image photographed by a CT photographing apparatus and a three-dimensional image photographed by a PET photographing apparatus have different density values and contrasts of light and shade appearing even in the same organ. A model image different from the three-dimensional standard density distribution model image for PET is required.

  The “first standard density distribution model image” is a model image representing the distribution of standard density values that appear on the image when the image is taken by the “first imaging device”. The “distribution model image” is a model image representing a distribution of standard density values appearing on the image when the image is taken by the “second photographing apparatus”.

  The “corresponding position of the corresponding anatomical structure between the first standard density distribution model image and the second standard density distribution model image” means that two standard density distribution model images are already aligned. Are prepared in advance, the same coordinate position on each standard density distribution model image becomes the corresponding position.

  Further, the first or second imaging apparatus may be a CT imaging apparatus, a PET imaging apparatus, or an MRI imaging apparatus.

  According to the present invention, when aligning images captured with different modalities, standard density distribution model images having density distributions corresponding to the modalities are prepared, and corresponding positions between the standard density distribution model images are stored. In addition, the captured images taken with different modalities are once aligned with the standard density distribution model images, and based on the corresponding positions between the standard density distribution model images associated in advance, between the captured images taken with the different modalities. It is better to align the captured image with the standard density distribution model image corresponding to each modality rather than aligning the structures appearing on the captured image captured with different modalities. Since the density values are similar, accurate positioning can be performed.

  The alignment apparatus of the present invention will be described below with reference to the drawings.

  When tomography is performed using different modalities, the density values and contrasts of anatomical structures such as organs and bones appear differently on the image even when the same part of the same subject is imaged. For example, a CT image captured by a CT imaging apparatus is an image in which the position and shape of an organ appear accurately, but a PET image captured by a PET imaging apparatus is an image that emphasizes highly active tissue such as cancer cells. However, the spatial resolution is low and the shape of the organ does not appear clearly. Further, since density values appear differently between the CT image and the PET image, accurate alignment cannot be expected between the CT image and the PET image. Thus, in the present embodiment, the CT image and the PET image obtained by photographing the subject to be inspected are accurately aligned via the CT standard density distribution model image and the PET standard density distribution model image. The case where it performs is demonstrated concretely.

  As shown in FIG. 1, the alignment apparatus 1 of the present invention uses a CT standard density distribution model image 11 (first standard density distribution model image) representing a standard density distribution of an image captured by a CT imaging apparatus. First standard density distribution model image storage means 12 to be stored, and standard density distribution model image 13 for PET (second standard density distribution model image) representing a standard density distribution of an image photographed by a PET photographing apparatus are memorized. Corresponding position for storing the corresponding position 15 of the corresponding anatomical structure among the second standard density distribution model image storage means 14, the CT standard density distribution model image 11 and the PET standard density distribution model image 13. Storage means 10, first photographed image storage means 20 for storing a CT image 21 (first photographed image) obtained by photographing a subject to be inspected using a CT photographing apparatus, and PET photographing A second photographed image storage means 30 for storing a PET image 31 (second photographed image) obtained by photographing the same subject using a device, a CT image 21 and a CT standard density distribution model image 11 Between the first alignment means 22 for aligning the corresponding anatomical structures to obtain the first corresponding position 23, and the PET image 31 and the standard density distribution model image 13 for PET. The second alignment means 32 for aligning the anatomical structure to obtain the second corresponding position 33, and the corresponding position of the anatomical structure between the CT image 21 and the PET image 31 are used for CT. Corresponding position acquisition means 40 for obtaining from the corresponding position 15 between the standard density distribution model image and the standard density distribution model image for PET, the first corresponding position 23 and the second corresponding position 33 is provided.

  The CT image 21, the PET image 31, the CT standard density distribution model image 11, and the PET standard density distribution model image 13 are volume data composed of a large number of voxel data, hold information on density values for each voxel data, Anatomical structures such as bones and organs are represented by shades of voxel data.

  The first captured image storage means 20 and the second captured image storage means 30 are large-capacity storage devices such as an image server. A CT image 21 and a PET image 31 obtained by photographing a subject to be inspected using a CT photographing device and a PET photographing device are stored in an image server, and the CT image 21 and the PET image 31 of the subject can be searched as necessary. Is memorized.

  The standard density distribution model image is based on a large number of CT images 21 and PET images 31 taken in the past, and an anatomical structure in which a healthy person has a standard body shape and a tissue such as an internal organ exists in a standard size and position. It is a model image showing a structure by standard density value distribution. Since the density value that appears in each image differs depending on the modality that was taken, the CT standard density distribution model image 11 represented by the standard density value that appears on the image when taken with the CT imaging apparatus, and the PET imaging apparatus A standard density distribution model image 13 for PET represented by standard density values appearing on the image is prepared.

  These standard density distribution model images can acquire the average position and density value of each organ from a number of images taken in the past. In addition, a model image generated so that each voxel data included in the standard concentration distribution model image belongs to which anatomical structure (bone, viscera, fat, etc.) belongs is preferable. For example, when coordinate values (x, y, z) are given to the CT standardity distribution model image 11 (CT_Model), the density value and the tissue name can be understood as shown in the following expressions (1) and (2). 1 standard density distribution model image storage means 12.

Density value CT_Model_intensity (x, y, z) = 100 (1)
Organization name CT_Model_tissue (x, y, z) = Liver (2)
Similarly, when coordinate values (x, y, z) are given to the PET standardity distribution model image 13 (PET_Model), the density value and the tissue name can be understood as shown in the following equations (3) and (4). The second standard density distribution model image storage means 14 is stored.

Density value PET_Model_intensity (x ', y', z ') = 80 (3)
Tissue name PET_Model_tissue (x ', y', z ') = liver (4)
Corresponding positions of the anatomical structures of the CT standardity distribution model image 11 and the PET standardity distribution model image 13 are stored in the corresponding position storage means 10 in correspondence with each voxel data. That is, a relationship represented by the following formula (5) is obtained in advance.

CT_Model (x, y, z) = PET_Model (x ', y', z ') (5)
This corresponding position may be obtained manually.

  Hereinafter, the alignment method of the present embodiment will be described based on the flowchart of FIG.

  First, the first captured image reading unit 24 reads the CT image 21 obtained by capturing the subject to be inspected from a storage device (first captured image storage unit 20) such as an image server (S100), and the second captured image. The reading unit 34 reads a PET image 31 obtained by photographing the same subject from a storage device (second photographed image storage unit 30) such as an image server (S101).

  Next, the first alignment means 22 aligns the CT image 21 and the CT standardity distribution model image 11 (S102). Images taken with the same modality are more accurate than the alignment of images taken with different modalities because the concentration distribution of anatomical structures such as organs and tissues shown in the image are similar. Can be combined.

  For example, the body surface is recognized from the CT image 21 based on the pixel value, and the shape of the body surface of the CT standardity distribution model image 11 read by the first standardity distribution model image reading means 16 is substantially the same. As described above, alignment is performed using a technique such as warping. Alternatively, characteristic points with little difference due to individual differences are extracted and aligned so that the characteristic points coincide with each other. Specifically, first, the CT image 21 is translated and rotated so that the shape and feature points of the body surface of the CT standardity distribution model image 11 and the CT image 21 as shown in FIG. When rough alignment is performed by performing three-dimensional affine transformation such as enlargement or reduction, a CT image 21 ′ shown in FIG. 3B is obtained. A small region R2 (for example, a cubic shape) is set on the CT image 21 ′ subjected to affine transformation, and slightly above the small region R2 at a position on the CT standardity distribution model image 11 corresponding to the position of the small region R2. A wide range search region R1 is set, and a place where the density values of the CT image 21 most closely match is searched on the CT standardity distribution model image 11 to accurately correspond to the anatomical structures. Get the corresponding position.

  Alternatively, a specific organ is automatically extracted from the CT image 21 (for details, see, for example, “Mito Watanabe,“ Automatic extraction of liver region from a simple and contrasted CT image by region expansion method ”, MEDICAL IMAGING TECHNOLOGY (21st (Refer to the Abstracts of the 1st Annual Meeting of the Japan Medical Imaging Society (CD-ROM version), 20, 4, OP6-31 (July 2002)))) Also good.

  In this way, the CT image 21 (CT_Data) and the CT standardity distribution model image 11 obtain the first corresponding position as in the following equation (6).

CT_Model (x1, y1, z1) = CT_Data (x, y, z) (6)
Similarly, the second alignment means 32 aligns the PET image 31 (PET_Data) and the PET standardity distribution model image 13 read by the second standardity distribution model image reading means 17. (S103), a corresponding position as in the following formula (7) is obtained.

PET_Model (x2, y2, z2) = PET_Data (x ', y', z ') (7)
Corresponding position acquisition means 40 calculates coordinate values (x ′, y ′,...) Of PET image 31 corresponding to coordinate values (x, y, z) of CT image 21 from the above equations (5), (6), (7). z ′) can be obtained. By obtaining this for all the voxel data, the CT image 21 and the PET image 31 are aligned (S104).

  In the above description, the corresponding positions of the CT and PET standard density distribution model images are stored in advance and aligned. However, two standard density distribution model images prepared in advance have already been registered. When the same coordinate position on each standard density distribution model image corresponds, the CT image is aligned with the standard density distribution model image for CT, and the PET image is aligned with the standard density distribution model image for 2 for PET. By simply doing, the CT image and the PET image can be aligned.

  As described above in detail, it is possible to perform accurate alignment by once aligning an image captured by each modality with the standard density distribution model image for each modality.

Configuration diagram showing the configuration of the alignment device Flow chart showing the flow of alignment Diagram for explaining alignment

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Positioning apparatus 10 Corresponding position storage means 11 First standard density distribution model image 12 First standard density distribution model image storage means 13 Second standard density distribution model image 14 Second standard density distribution model image storage means 15 Corresponding position of first standard density distribution model image and second standard density distribution model image 16 First standard degree distribution model image reading means 17 Second standard degree distribution model image reading means 20 First photographed image storage means 21 First captured image 22 First alignment unit 23 First corresponding position 24 First captured image reading unit 30 Second captured image storage unit 31 Second captured image 32 Second alignment unit 33 Second 2 corresponding position 34 second captured image reading means 40 corresponding position obtaining means

Claims (4)

A plurality of anatomical structures obtained by photographing a subject using the first photographing device are represented by using standard density values that appear on the image when the first photographing device is photographed. First model image storage means for storing a first standard density distribution model image;
A plurality of anatomical structures obtained by photographing a subject using the second photographing device are represented by using standard density values appearing on the image when the second photographing device is photographed. Second model image storage means for storing two standard density distribution model images;
First photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing device;
Second photographed image storage means for storing a second photographed image including an anatomical structure obtained by photographing the predetermined subject using the second photographing device;
First alignment means for obtaining a first corresponding position obtained by aligning corresponding anatomical structures between the first standard density distribution model image and the first captured image;
Second alignment means for acquiring a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. An alignment apparatus comprising: a corresponding position acquisition unit that is determined from a corresponding position of a corresponding anatomical structure, the first corresponding position, and the second corresponding position.   The alignment apparatus according to claim 1, wherein the first or second imaging apparatus is any one of a CT imaging apparatus, a PET imaging apparatus, and an MRI imaging apparatus. A plurality of anatomical structures obtained by photographing a subject using the first photographing device are represented by using standard density values that appear on the image when the first photographing device is photographed. First model image storage means for storing a first standard density distribution model image;
A plurality of anatomical structures obtained by photographing a subject using the second photographing device are represented by using standard density values appearing on the image when the second photographing device is photographed. Second model image storage means for storing two standard density distribution model images;
First photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing device;
An alignment apparatus comprising: a second photographed image storage unit that stores a second photographed image including an anatomical structure obtained by photographing the predetermined subject using the second photographing apparatus. ,
A first alignment step of obtaining a first corresponding position obtained by aligning a corresponding anatomical structure between the first standard density distribution model image and the first captured image;
A second alignment step of obtaining a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. A registration method, comprising: a corresponding position acquisition step that is obtained from a corresponding position of a corresponding anatomical structure, the first corresponding position, and the second corresponding position. Computer
A plurality of anatomical structures obtained by photographing a subject using the first photographing device are represented by using standard density values that appear on the image when the first photographing device is photographed. First model image reading means for reading out the first standard density distribution model image from first model image storage means for storing a first standard density distribution model image;
A plurality of anatomical structures obtained by photographing a subject using the second photographing device are represented by using standard density values appearing on the image when the second photographing device is photographed. Second model image reading means for reading out the second standard density distribution model image from second model image storage means for storing two standard density distribution model images;
The first photographed image is read out from first photographed image storage means for storing a first photographed image including an anatomical structure obtained by photographing a predetermined subject using the first photographing device. First photographed image reading means;
The second photographed image is read out from a second photographed image storage means for storing a second photographed image including an anatomical structure obtained by photographing a predetermined subject using the second photographing apparatus. Second captured image reading means;
First alignment means for obtaining a first corresponding position obtained by aligning corresponding anatomical structures between the first standard density distribution model image and the first captured image;
Second alignment means for acquiring a second corresponding position obtained by aligning a corresponding anatomical structure between the second standard density distribution model image and the second captured image;
The corresponding position of the anatomical structure corresponding to between the first captured image and the second captured image is determined between the first standard density distribution model image and the second standard density distribution model image. A program that functions as a corresponding position acquisition unit that is obtained from a corresponding position of a corresponding anatomical structure, the first corresponding position, and the second corresponding position.

Images