JP5068336B2 - Medical image conversion apparatus and method, and program - Google Patents

Description

  The present invention relates to a medical image conversion apparatus and method for converting a time-series medical image for representing the motion of an organ such as the heart for volume rendering display, and a program for causing a computer to execute the medical image conversion method. Is.

  In recent years, high-quality three-dimensional images have been used for image diagnosis due to advances in medical equipment (for example, multi-detector CT). Here, since a three-dimensional image is composed of a large number of two-dimensional tomographic images and has a large amount of information, it may take time for a doctor to find and diagnose a desired observation site. Accordingly, the structure of interest is recognized by using a method such as a maximum value projection method (MIP method) and a minimum value projection method (MinIP method) from a three-dimensional image including the structure of interest, for example. By creating a 3D image and performing MIP display, 3D image volume rendering (VR) display, or CPR (Curved Planer Reconstruction) display, the entire structure, and further the structure Various techniques for improving the visibility of lesions contained in objects have been proposed.

  By the way, in the multi-detector type CT described above, a large number of tomographic images can be acquired at once using a plurality of detectors, and it is now possible to acquire a tomographic image exceeding 300 slices in one rotation. It has become. Further, since the time required for one rotation of the detector is about 0.3 seconds, it is possible to acquire a plurality of three-dimensional images in time series at short time intervals if only a specific organ is used. In this way, the organs of interest included in the three-dimensional image acquired in time series are displayed in time series order, that is, by displaying the three-dimensional time including the time in three dimensions, the moving image can be displayed as the moving organ of interest. It is possible to observe as seen (see Patent Document 1).

  Thus, by displaying a three-dimensional image four-dimensionally, heart analysis and the like can be performed particularly in the cardiovascular field. Moreover, when acquiring a three-dimensional image using a contrast agent as well as organs with motion such as the heart and lungs, a specific flow such as a liver due to a contrast agent effect is displayed by displaying the flow of the contrast agent in four dimensions. It becomes possible to diagnose organs.

  When a three-dimensional image is displayed in VR, the organ of interest is extracted, and the signal value of each pixel is determined according to the signal value (CT value in the case of a CT image) of each pixel position of the extracted three-dimensional image of the organ. A color (R, G, B) and opacity (opacity) are set for the three-dimensional display. The four-dimensional display of the VR image is performed by setting a color and opacity for each of a plurality of three-dimensional images, creating a VR image, and displaying the created VR images in chronological order.

JP 2005-322252 A

  By the way, in a three-dimensional image acquired by photographing a specific organ at a short time interval in chronological order, the signal value of the corresponding pixel position of the organ included in each of the three-dimensional images should be the same. In practice, however, signal values at corresponding pixel positions of the same organ often differ due to the influence of noise or the like during imaging. Thus, if the signal values of corresponding pixel positions of the same organ are different, when the VR image is displayed in four dimensions, the color and opacity of the same position of the organ will fluctuate together with the movement of the organ. . As described above, when the color and opacity of the organ fluctuate, the three-dimensional movement of the position is illusioned, and there is a possibility that accurate diagnosis cannot be performed.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to prevent variations in color and the like when displaying medical images in chronological order.

The medical image conversion apparatus according to the present invention includes an image acquisition means for acquiring a series of time-series medical images having different time phases for a specific organ,
Alignment means for aligning pixel positions in the series of time-series medical images between the series of time-series medical images;
The series of time-series medical images includes conversion means for converting a signal value at a corresponding pixel position of the specific organ into the same display pixel value.

“A series of time-series medical images with different time phases” is obtained by continuously capturing specific organs of the same subject at short time intervals , and displaying the movements of specific organs by displaying them in chronological order. Any image can be used if possible. Specifically, a three-dimensional image, a three-dimensional specific organ image extracted from the three-dimensional image, a two-dimensional image at a specific slice position including the specific organ in the three-dimensional image, or a specific acquired by simple X-ray imaging An organ image or the like can be used.

  The medical image conversion apparatus according to the present invention may further include a smoothing unit that smoothes the series of time-series images before the alignment.

  The medical image conversion apparatus according to the present invention may further include display means for displaying the converted series of time-series images in order of time series.

  In the medical image conversion apparatus according to the present invention, the time-series medical image may be a three-dimensional medical image.

  In the medical image conversion apparatus according to the present invention, the specific organ may be a heart and / or a lung.

The medical image conversion method according to the present invention acquires a series of time-series medical images having different time phases for a specific organ,
Aligning pixel positions in the series of time-series medical images between the series of time-series medical images;
In the series of time-series medical images, the signal value of the corresponding pixel position of the specific organ is converted to the same display pixel value.

  The medical image conversion method according to the present invention may be provided as a program for causing a computer to execute the method.

  According to the present invention, a series of time-series medical images having different time phases for a specific organ are acquired, pixel positions in the series of time-series medical images are aligned between the series of time-series medical images, and a series of In the time series medical image, the signal value at the corresponding pixel position of the specific organ is converted to the same display pixel value. For this reason, even if a series of time-series medical images are displayed in time-series order, the display pixel value of a certain part in the specific organ does not fluctuate with the movement of the organ. Therefore, there is no illusion of the movement of the corresponding position in the specific organ, and as a result, the diagnosis using the medical images displayed in chronological order can be performed accurately.

1 is a schematic block diagram showing the configuration of a medical image conversion apparatus according to an embodiment of the present invention. The figure which shows the alignment result of the three-dimensional image of the heart Diagram showing color template A flowchart showing processing performed in the present embodiment The schematic block diagram which shows the structure of the medical image converter by other embodiment of this invention.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. FIG. 1 is a schematic block diagram showing a configuration of a medical image conversion apparatus according to an embodiment of the present invention. The configuration of the medical image conversion apparatus 1 shown in FIG. 1 is realized by executing a medical image conversion processing program read into the auxiliary storage device on a computer. At this time, the medical image conversion processing program is stored in a storage medium such as a CD-ROM or distributed via a network such as the Internet and installed in a computer.

  The medical image conversion apparatus 1 according to the present embodiment includes a volume data acquisition unit 10, a storage unit 20, a registration unit 30, a conversion unit 40, a display control unit 50, and an input unit 60.

  The volume data acquisition unit 10 is a three-dimensional volume data group 110 including a plurality of three-dimensional volume data 100 obtained by photographing a specific organ of a subject at a predetermined time interval Δt in a modality 2 such as a CT apparatus or an MRI apparatus. The communication interface function is acquired. The three-dimensional volume data group 110 is transmitted from the modality 2 via the LAN. In the present embodiment, the specific organ is the heart.

  Here, the three-dimensional volume data 100 is obtained by stacking two-dimensional tomographic image data obtained in order along the direction perpendicular to the tomographic plane on the heart to be diagnosed. Is generated by superimposing a plurality of tomographic images taken by a modality 2 such as a CT apparatus or an MRI apparatus. Note that the volume data acquired using the CT apparatus is data that stores the amount of X-ray absorption for each voxel (that is, pixel position), and one signal value for each pixel position (in the case of imaging with the CT apparatus) , A value indicating the amount of X-ray absorption).

  The three-dimensional volume data group 110 includes, for example, a series of three-dimensional volume data 100 acquired by photographing a subject at different time phases t1, t2,..., Tn having a constant time interval Δt. The

  The three-dimensional volume data 100 is appended with incidental information defined by the DICOM (Digital Imaging and Communications in Medicine) standard. The incidental information includes, for example, an image ID for identifying a three-dimensional image represented by each three-dimensional volume data 100, a patient ID for identifying a subject, an examination ID for identifying an examination, and image information. Unique ID (UID) assigned, examination date when the image information was generated, examination time, type of modality used in the examination to obtain the image information, patient information such as patient name, age, gender, Examination site (imaging site, heart in this embodiment), imaging conditions (contrast agent used, radiation dose, etc.) Information such as series number or collection number when multiple images are acquired in one examination Can be included.

  The storage unit 20 is a large-capacity storage device such as a hard disk, and stores a three-dimensional volume data group 110. Note that the storage unit 20 stores a plurality of three-dimensional volume data groups 110 having different subjects (that is, different patients) or having the same subject and different photographing times.

  The alignment unit 30 aligns corresponding pixel positions in the heart portion between the three-dimensional volume data 100 for each of the three-dimensional volume data 100. Specifically, `` WM Wells III, P. Viola, H. Atsumi, S. Nakajima, and R. Kikinis, Multi modal volume registration by maximization of mutual information, Med.Image Anal., Vol.1, no.1 , pp. 35 51, 1996. (Reference 1), “Rueckert, D., Sonoda, LI, Hayes, C., Hill, DLG, Leach, MO, Hawkes, DJ, Nonrigid registration using free-form deformations: application to breast MR images. IEEE Transactions on Medical Imaging, vol.18, pp.712? 721, 1999. (Reference 2), “Masumoto J, Sato Y, Hori M, Murakami T, Johkoh T, Nakamura H, `` Tamura S: A similarity measure for nonrigid volume registration using known joint distribution of target tissue: Application to dynamic CT data of the liver, Medical Image Analysis, vol.7, no.4, pp.553-564, 2003. '' Reference 3), “Yongmei Wang, Lawrence H. Staib, Physical model-based non-rigid registration incorporating statistical shape information, Medical Image Analysis (2000) volume 4, number 1, pp 7-21” (reference 4) Corresponding pixel positions can be associated using the described technique.

  The technique described in Reference 1 performs alignment using a rigid registration technique, and the pixel position is set so that the mutual information amount is maximized between three-dimensional medical images acquired by different modalities. This is a method of aligning pixels by adjusting the direction and direction. The method described in Reference 2 performs registration using a non-rigid registration method, and by using a deformation estimation method based on a B-spline function called “free from deformation” (FFT), This is a technique for performing alignment. The technique described in Reference 3 performs registration using a non-rigid registration technique, and in a CT image acquired in time series order, a joint distribution of a target tissue such as a liver (joint distribution). Is used to measure and align the similarity by sliding the tissue at the boundary between the target tissue and the non-target tissue. The technique described in Reference 4 performs alignment using a non-rigid registration technique. The shape of an object is given in advance, and the object is deformed so as to obtain the shape. This is a method of aligning the positions.

  Moreover, it is also possible to use the technique described in the Japanese translations of PCT publication No. 2005-528974 gazette and the Japanese translations of PCT publication 2007-516744 gazette. The technique described in JP 2005-528974 A acquires first and second image data sets from a region of interest of a target, creates a model of physiological motion such as respiration and heart motion with respect to the region of interest, Fit a physiological model to the first image data set, apply an object-specific physiological phantom to the second image data set for conversion, and apply the conversion to the first image data set This is a technique for performing alignment.

  The technique described in JP-T-2007-516744 is a technique for aligning images by aligning the positions of the landmarks based on the similarity of the positions of the landmarks between the two images. .

  The alignment unit 30 performs alignment between the hearts included in the three-dimensional image represented by the three-dimensional volume data 100 using these methods. Thus, pixels representing the same position are associated with each other between the hearts included in the three-dimensional image.

  The alignment method is not limited to the above-described various methods, and any known method can be used. Alternatively, the three-dimensional volume data 100 may be sequentially displayed on the monitor 4, and the operator may perform alignment by input from the input unit 60. Further, not only the heart portion but also all pixel positions of the three-dimensional volume data 100 may be aligned.

  FIG. 2 is a diagram showing the alignment result of the three-dimensional image of the heart. FIG. 2 shows the alignment result of the heart included in the three three-dimensional volume data 100A, 100B, and 100C. As shown in FIG. 2, the pixel P1 on the heart included in the three-dimensional volume data 100A is associated with the pixels P2 and P3 of the three-dimensional volume data 100B and 100C, respectively.

  The conversion unit 40 converts the signal value at each pixel position of the three-dimensional volume data 100 into a display pixel value for volume rendering (VR) display. FIG. 3 is a diagram showing a color template for converting the three-dimensional volume data 100. Note that a plurality of color templates are prepared according to the part to be extracted and displayed in VR among the parts included in the three-dimensional volume data 100. In the present embodiment, the color template T0 for VR display of the heart is prepared. Shall be selected. As shown in FIG. 3, the color template T0 is a one-dimensional lookup table in which the horizontal axis represents the signal value of the three-dimensional volume data 100, and the vertical axis represents color (R, G, B) and opacity. Although only one color template is shown in FIG. 3, actually, four color templates are prepared for each of R, G, and B colors and opacity.

  The converter 40 refers to the color template T0 and converts the signal value of each pixel of the three-dimensional volume data 100 into a display pixel value composed of R, G, B, and opacity. At this time, the conversion unit 40 selects one reference time phase B serving as a reference among the time phases of the plurality of three-dimensional volume data 100. A signal value at each pixel position of the three-dimensional volume data 100 in a time phase other than the reference time phase is used as a corresponding pixel position of the three-dimensional volume data in the reference time phase B (hereinafter referred to as reference three-dimensional volume data 120). To the signal value of.

  The selection of the reference time phase B is, for example, the acquisition of the three-dimensional volume data 100 with the least time, the first time phase, the intermediate time phase, the last time phase, or the noise in the time phase of the three-dimensional volume data group 110. A predetermined time phase such as a time phase may be selected. In addition, selection of which time phase is set as the reference time phase B may be accepted by input from the input unit 60. At this time, the operator may select a time phase or the like at which the three-dimensional volume data 100 with the least noise is acquired as the reference time phase B.

  Then, the conversion unit 40 converts the signal value at each pixel position in all the three-dimensional volume data 100 into the signal value at the corresponding pixel position in the reference three-dimensional volume data 120, and then uses the color template T0 to convert each 3 The signal value of the dimension volume data 100 is converted into a display pixel value. As a result, the three-dimensional volume data 100 represents a VR image from which the heart is extracted.

  As a result, the corresponding pixel positions P1, P2, and P3 of the three three-dimensional volume data 100A, 100B, and 100C shown in FIG. Even if it is, it will be converted into display pixel values of the same color and opacity. For example, among the three three-dimensional volume data 100A, 100B, and 100C, when the three-dimensional volume data in the reference time phase B is the three-dimensional volume data 100B, the signal values at the pixel positions P1 and P3 are converted to 110, respectively. As a result, the color and opacity corresponding to 110 signal values in the color template T0 are converted.

  The display control unit 50 displays the VR images represented by the converted series of three-dimensional volume data 100 on the display 4 in chronological order. In the present embodiment, since the heart is displayed, a state in which the heart beats is displayed on the display 4.

  The input unit 60 includes a known input device such as a keyboard and a mouse.

  Next, processing performed in the present embodiment will be described. FIG. 4 is a flowchart showing processing performed in the present embodiment. A plurality of three-dimensional volume data groups 110 of the heart are acquired by the volume data acquisition unit 10 and stored in the storage unit 20. When the operator operates the input unit 60 to select a three-dimensional image to be displayed (Yes in step ST1), the alignment unit 30 selects the three-dimensional volume data group 110 corresponding to the selected three-dimensional image. Is read from the storage unit 20, and the pixel positions between the three-dimensional volume data 100 constituting the three-dimensional volume data group 110 are aligned (step ST2). Thereby, the pixel position between the three-dimensional volume data 100 is matched.

  Then, the conversion unit 40 selects the reference time phase B (step ST3), and converts the signal values at the respective pixel positions of all the three-dimensional volume data 100 to the corresponding pixels of the reference three-dimensional volume data 120 in the reference time phase B. The signal value of the three-dimensional volume data 100 is converted into a display pixel value using the selected color template T0 (step ST5). Then, the display control unit 50 displays the VR images of the heart represented by the converted three-dimensional volume data 100 on the display 4 in chronological order (step ST6), and ends the process.

  Thus, in the present embodiment, the alignment unit 30 associates the pixel positions of the three-dimensional volume data 100 constituting the three-dimensional volume data group 110 by aligning the three-dimensional volume data 100, The conversion unit 40 selects the reference time phase B, converts the signal values of the respective pixel positions of all the three-dimensional volume data 100 into the signal values of the corresponding pixel positions of the reference three-dimensional volume data 120 in the reference time phase B, Further, the color template T0 is used to convert the signal value of each three-dimensional volume data 100 into a display signal value composed of R, G, B and opacity.

  For this reason, when the three-dimensional volume data group 110 is displayed in chronological order, the color and opacity at a certain pixel position in the heart do not fluctuate with the movement of the heart. Therefore, there is no illusion of the three-dimensional movement of the heart due to fluctuations in the color and opacity in the heart, and as a result, accurate diagnosis using the three-dimensional volume data group 110 displayed in time series is performed. Can be done.

  In the above embodiment, a three-dimensional volume data group of the heart is used, but a three-dimensional volume data group of the lung may be used. In this case, the VR image is displayed in four dimensions so as to represent the three-dimensional movement of the lungs due to respiration. Further, a three-dimensional volume data group of the circulatory system including both the heart and the lung may be used. In this case, the VR image is displayed four-dimensionally so as to represent both the three-dimensional movement due to the heartbeat and the three-dimensional movement of the lung due to breathing.

  Further, in the above embodiment, as shown in FIG. 5, a smoothing unit 70 is provided so that each three-dimensional volume data 100 is smoothed before the pixel positions between the three-dimensional volume data 100 are aligned. Also good. Specifically, the average value of the signal values at the respective pixel positions of the three-dimensional volume data 100 is calculated using a smoothing filter of a predetermined size (for example, 3 × 3 × 3), whereby each three-dimensional volume data 100 is calculated. Can be smoothed. Thereby, when performing alignment, since the influence of the noise contained in the three-dimensional volume data 100 can be reduced, alignment can be performed more accurately.

  In the above-described embodiment, the case is described in which the three-dimensional volume data group 110 of the heart is displayed in VR in chronological order, but 2 representing a cross section of the heart on the slice plane at the same position of each three-dimensional volume data 100. Of course, the present invention can also be applied to a case where a three-dimensional image is extracted from each of the three-dimensional volume data 100 and the density and / or color of the extracted two-dimensional image is converted and displayed in time series. Further, the time series images are not limited to the three-dimensional volume data 100, and an image group including a series of images acquired at a predetermined time interval by simple X-ray imaging can be used.

DESCRIPTION OF SYMBOLS 1 Medical image conversion apparatus 2 Modality 4 Display 10 Volume data acquisition part 20 Storage part 30 Positioning part 40 Conversion part 50 Display control part 60 Input part 70 Smoothing part 100 Three-dimensional volume data 110 Three-dimensional volume data group

Claims (7)

A series of images in which the heart and / or lungs of the same subject can be replayed by displaying the heart and / or lung movements in time-sequential order obtained by continuously taking images at short time intervals. Image acquisition means for acquiring time-series medical images;
Alignment means for aligning pixel positions in the series of time-series medical images between the series of time-series medical images;
A medical image conversion apparatus comprising: conversion means for converting a signal value at a corresponding pixel position of the heart and / or lung into the same display pixel value in the series of time-series medical images.   The medical image conversion apparatus according to claim 1, further comprising a smoothing unit that smoothes the series of time-series images before the alignment.   3. The medical image conversion apparatus according to claim 1, further comprising display means for displaying the converted series of time-series images in order of time series.   The medical image conversion apparatus according to claim 1, wherein the time-series medical image is a three-dimensional medical image. The conversion means selects one reference time phase as a reference from each time phase of the series of time-series medical images, and in the series of time-series medical images, corresponding pixels of the heart and / or lungs the signal value of the position, the medical image conversion apparatus according to any one of the four claims 1, characterized in that the means for converting the display pixel values in the time-series medical images of the reference time phase. A series of images in which the heart and / or lungs of the same subject can be replayed by displaying the heart and / or lung movements in time-sequential order obtained by continuously taking images at short time intervals. Acquire time series medical images,
Aligning pixel positions in the series of time-series medical images between the series of time-series medical images;
A medical image conversion method, comprising: converting a signal value of a corresponding pixel position of the heart and / or lung into the same display pixel value in the series of time-series medical images. A series of images in which the heart and / or lungs of the same subject can be replayed by displaying the heart and / or lung movements in time-sequential order obtained by continuously taking images at short time intervals. A procedure for acquiring time-series medical images;
A procedure for aligning pixel positions in the series of time-series medical images between the series of time-series medical images;
In the series of time-series medical images, the computer has a procedure for converting a signal value at a corresponding pixel position of the heart and / or lung into the same display pixel value. Program for.

Images