JPH08173430A - Method of generating anatomic m- mode display - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method that allows the free selection of the position and direction of an Mmode line and provides good anatomical information, among methods for examining biological tissues in motion using an ultrasonic converter and generating an anatomical Mmode display. SOLUTION: A method of generating an anatomical M-mode display when an ultrasonic examination is performed for live biological tissues in motion, using an ultrasonic converter 21, comprises the stage of obtaining a time series of a two-dimensional or three-dimensional ultrasonic image 22 and the stage of arranging the time series so that a data set having at least one simultaneously-recorded virtual M-mode line 23 may be constituted. In addition, the method also has the stage of computing the data set so that interpolation may be made along the virtual M-mode line and the stage of displaying the anatomical M-mode display of the result of computation on a display unit.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention provides an anatomical M-mode display when performing live ultrasound examination of living living tissue, such as cardiac action, using an ultrasound transducer. On how to do.

[0002]

2. Description of the Related Art The present invention is 2D (two-dimensional)
Alternatively, a technique for extracting data from a 3D (three-dimensional) ultrasonic image and obtaining an M-mode display having anatomical significance is described. The conventional M-mode is acquired along one acoustic beam of an ultrasonic transducer, showing time along the x-axis, depth along the y-axis, and time-varying data on the display unit. Is displayed. The placement of M-mode lines in conventional M-mode is limited to the set of beam directions that can be produced (scanned) by the transducer.

In cardiology, the use of M-mode methods has been fairly standardized to cut the heart at reference positions and angles. The important criteria for successfully executing the M-mode measurement are as follows. 1. Video quality. The interface, contact surface, between the various tissues of the heart must be clearly visible. One of the most important factors for achieving this is to position the ultrasonic transducer at a point where the acoustic characteristics on the body are optimum. Such locations are often referred to as "acoustic windows." In elderly patients, these windows are scarce and difficult to detect. 2. Alignment. Standard M-mode measurements require that recordings be made along a specific angle, usually 90 ° to the tissue of the heart being examined. 3. motion. Since the heart moves in the chest while contracting and relaxing, the correct M-mode line position at one point in the heart cycle becomes incorrect at another point in the same heart cycle. Manually compensating for this is very difficult because the probe must be moved in synchronism with the heartbeat.
Therefore, many sonographers (acoustic recorders) orient the M-mode line or transducer beam in a fixed compromise direction. 4. Analysis of wall thickening. In coronary artery disease,
An important factor to observe is the thickening of the left ventricular muscle at various points.

In many cases there is a problem in getting the correct alignment in a good acoustic window. Good acoustic windows often give poor alignment and vice versa. Therefore, the sonographer or the user spends a lot of time and effort to optimize the image for two criteria (alignment, image quality). As mentioned above, techniques for M-mode imaging by mapping a single acoustic beam as a function of time are already known.

[0005]

SUMMARY OF THE INVENTION With the advent of high performance digital front end control of phased sequence conversion probes,
Very high frame rate (10ms or less for each 2D image)
There is a possibility to acquire a 2D image by using the. These 2D data are stored in the RAM of a computer with sufficient storage capacity to hold one or more full cardiac cycles worth recording as 2D data. An M-mode display can be generated with appropriate temporal resolution based on these recordings. According to the invention, the M-mode line can be positioned with complete flexibility. The present invention describes how this flexibility can be used to improve the anatomical information content in the excerpted M-mode display.

The present invention is also based on the time series of 3D images, and
It is also applied to extract the mode display. In 3D, it is possible to compensate for true 3D motion of the ventricles.
Based on 2D recordings, the operator would be limited to compensating for movements that could be measured in the image plane. The present invention also describes how the local M-mode information derived from the 3D acquisition material can be used to obtain a ventricular wall color encoding that provides information regarding wall thickening. .

The anatomical M-mode display can be generated in real time during scanning of 2D images or during real time spatial scanning.
In doing so, the present invention describes how multiple M-mode displays can be maintained with a live 2D or 3D image. These multiple M-mode displays can also be positioned freely and even follow the position and orientation of the heart chamber central chamber wall. The anatomical M-mode can also be used as a post-processing means, in which case the user obtains the course of the 2D / 3D image at a very fast frame rate without making an M-mode recording. . 2D
The anatomical M-mode can be used by the user to perform a later M-mode analysis, as long as the data includes appropriate cuts / illustrations through the heart.

In principle, computer processing of datasets containing operations or steps which are also used in the practice of the present invention is already known from the literature such as: [1] J.
D. Foley, A Van Dam, S .; K. Sei
ner, J.N. F. Hughes's "Computer Figure Processing: Principles and Practice" Addition Wesley USA1
990. The drawing algorithm is also described in this document. Therefore, such computer processes, operations, and steps will not be described in detail below. Other references related to the technology of particular interest here are: [2] B. Olstad's "Maximizing Video Variation in Spatial Dataset Formation" Journal of El
electronic Imaging, 1: 245-26
5, July 1992 issue. [3] E. Steen, B.M. Report on the 8th Scandinavian Conference on Image Analysis of "Spatial Display in Medical Ultrasound Images" by Olstad. Tromsφ, Norway 1
May 993. [4] G. Borgfors, "Distance Conversion in Digital Images," Computer Video, Graphic Processing, 34, 198.
6, pp. 344-371. [5] Peter Seitz, "Optical Super-Resolution Using Solid-State Camera and Digital Signal Processing," Optical Engineering 27 (7) July 1988.

Against the background of the known art, the present invention takes the established clinical means of utilizing M-mode imaging, starting from the conventional method of calculating M-mode. The present invention is 2D
Or, it describes a new technique for calculating anatomical M-mode display based on time series of 3D ultrasound images.
Anatomical M-modes are derived as virtual M-mode measurements along any or virtual title M-mode line. The novel and special features of the method of the present invention are set forth in more detail in the appended claims.

The advantages of the present invention can be summarized as follows. 1. There are two M-mode displays that can be arbitrarily located.
It can be calculated based on D or 3D acquisition material. 2. The position of the M-mode line is not limited to the scanning terrain and can be freely positioned. 3. Global heart motion can be compensated by moving the M-mode line in response to the motion of the heart during the cardiac cycle. 4. Wall thickening analysis is improved by the ability to keep the M-mode line normal to the heart wall during the entire cardiac cycle. 5. The reference point in the region of interest can be fixed on a given y-axis of the M-mode display, thus improving the visualization of the relative movement / thickening phenomenon. 6. The 3D acquisition material can be visualized by mapping the features extracted from the local M-mode line with the color encoding of the ventricular wall.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described in detail below with respect to various embodiments with reference to the accompanying drawings. FIG. 1 illustrates conventional M-mode visualization. An ultrasound transducer 11 is shown schematically in connection with an ultrasound image 12 obtained by angularly scanning the acoustic beam of the transducer. In this conventional method, M-
The mode line or corresponding acoustic beam 13 is fixed at a given point and the ultrasonic signal along the beam is mapped onto the M-mode display 14 as a function of time. In the present prior art, an extreme temporal resolution is achieved because a new time sample can be generated as soon as one beam of data has been collected. On the other hand, this prior art of M-mode imaging suffers from the limitation of locating the M-mode line 13 in correspondence with the acoustic window and the scanned figure.

Title M-Mode Line The present invention describes how an M-mode image is 2D or 3D.
It relates to what can be generated by extracting an interpolated display from a time series of images. The concept of the "Title" M-Mode display 24 is shown in FIG. in this case,
The "virtual" M-mode line 23 is free to move and need not coincide with the acoustic beam (transducer 21) originating at the top of the 2D image 22.

Multiple M-Mode Lines FIG. 3 shows two title M-mode displays 34A, 34B.
However, using the virtual and title M-mode lines shown in 33A and 33B, respectively, a single 2D sequence or image 3
The example calculated from 2 is shown. 2D or 3D
Since the M-mode display is generated based on the image, it is possible to generate the M-mode display with an arbitrary number of areas, which enables analysis of various sizes from the same heart beat. Thus, the acquisition time series indicated by 1, 2, 3, and 4 in FIG. 2 are arranged so as to form a data set, and at least one virtual M-mode line 23 or 33A, 3 in FIG.
3B is provided and registered with the dataset, then
These are computer processed with interpolation along the associated virtual M-mode line. The importance of interpolation will be further explained below.

Since the motion-correcting heart moves while contracting and relaxing in the chest, the correct M-mode line position at one point in the cardiac cycle is not correct at other points in the same cardiac cycle. This is very difficult to correct manually and the probe must be moved in synchronism with the heartbeat. The anatomical M-mode of the present invention can compensate for this movement. FIG.
Explains this concept. User is M-mode line 43
At the different points of the cardiac cycle,
The 2D video loop is determined by scrolling and fixing the new M-mode line position. Suitable computer operating methods or software for this purpose are available to the person skilled in the art, as indicated in the above-referenced document, where "fixed"
The M-mode line is interpolated between the M-mode lines 43A and 43B to produce an M-mode display 44, in which each vertical line is extracted corresponding to the position-specified M-mode line. ing. Clearly in this method, the position and / or orientation of the virtual M-mode line is
It is possible to move in response to a rhythmic movement in a biological tissue or body other than the heart beat mentioned in the description of.

Motion Reference Point In studying the size of an organ over time in a living body, it is possible to know the relative size of various tissues with respect to each other without observing the displacement of the entire organ within the body. Often desired. This is of particular interest when observing contraction and relaxation of the ventricles of the heart, where thickening of muscle tissue is an important factor to observe.

In an embodiment of the present invention, the user can set a reference point on the "fixed" M-mode line described in the section on motion compensation to emphasize relative changes. Typically this would correspond to an easily designated clinical organization. 5 and 6 illustrate the generation of M-modes with and without a given reference point 66 fixed within the imaging area 62. Thus, based on the reference point 66 associated with the interpolated M-mode line positions from 63A to 63B shown in FIG. 6, an M-mode display 64 is generated where the point 66 is a straight line 67 ( No motion), ie as the selected vertical coordinate point in the display. Instead, for a given y coordinate point,
The M-mode display is tracked within the M-mode display and the M-mode display is reproduced by sliding the position of the M-mode line at various points in time so that the tracked video structure appears as a horizontal structure in the final M-mode display. It can also be done.

Wall Thickening Analysis In coronary artery disease, an important factor to observe is the thickening of the left ventricular muscle at various locations. By combining the techniques described in the preceding paragraph, the present invention provides
As shown in, it provides a particularly useful tool for left ventricular wall thickening analysis. Each M-mode display 74A, 74B, 7
4C shows localized thickening and contraction of part of the ventricle 70,
The respective units correspond to the corresponding virtual M-mode lines 73A and 73A.
It is penetrated by B and 73C. FIG. 7 shows the left ventricle 70.
Short-axis view of 3D and three anatomical M-mode displays 74A, 74B, 74C generated using the technique described in the preceding paragraph.
Is shown.

The sequence of running 2D / 3D frames is stored in a scanner / computer to be used as a 3D or 4D array of ultrasound samples, or a data set. This array has different terrain characteristics, depending on the probe topography of the transducer used and whether the image was scan converted to a rectangular format prior to storage.
For purposes of illustration, in the configuration of FIG. 8, 2D area data has been scan converted (typically using an ultrasound scanner hardware converter) and has a rectangular data set format, [x , Y, t] stored as a 3D array of samples in the disk / storage.

To generate the M-mode display 84, a 3D
The data set 82 is cut at the plane 88 and the data is interpolated and resampled to fit the desired display rectangle 84. The motion compensation technique described above modifies the cutting plane 88 into a curved surface that linearly intersects the [x, y] plane. It is of primary importance to apply suitable interpolation techniques both spatially and temporally. Such interpolation can, to some extent, compensate for the inferior resolution compared to conventional M-modes along the acoustic beam produced by a transducer as shown in FIG. In accordance with one embodiment of the present invention, it is advantageous to perform known image processing of edge enhancement on the results of computer processing including interpolation to produce a computer processed anatomical M-mode display. It is a stage.

3D Ultrasound Imaging The techniques described herein all apply to both 2D and 3D series of ultrasound images. The 3D acquisition material further improves the ability of the motion correction, since the true 3D motion of the heart can be evaluated. In addition to actually generating the M-mode representation, the techniques of the present invention can be used to extract the anatomical M-mode for all points across the endocardial surface in the left ventricle. This configuration is illustrated in FIG.
The 4-dimensional ultrasound data set 92 is assumed to consist of m3 short-axis planes and n3D cubes recorded during the cardiac cycle. For simplicity, the figure shows M-mode display 94A, 94.
Although only three virtual M-mode lines 93A, 93B, 93C associated with B and 94C respectively are shown,
A similar M-mode display should be present for every point or position on the endocardial surface of ventricle 90.

Independent M-mode displays 94A, 94B,
94C ,. ． ． , Are then color-encoded to make the relevant location of the ventricular wall visible. The mapping policy is shown in FIG. 9, which is described in [2],
It is similar to the method of [3]. This featured method is performed on an anatomical M-mode display to produce a single value or color index that reflects the physiological characteristics guided in the M-mode image. One of these properties is to determine the amount of wall thickening by assessing the variation in wall thickening during the cardiac cycle. In this case, the anatomical M-mode displays 94A, 94B, 94C are analyzed. The wall is M-
In the mode display, using the prior art method [5], M
Positioned to obtain super-resolution peripheral concentration at different instants of the mode display, the thickness variation is used for the quantification of said wall thickening. The second characteristic is characterized by the temporal signal characteristic at a given spatial coordinate or range of spatial coordinates in the M-mode display 94A, 94B, 94C.

A second alternative is to use only two cubes that are close to each other in time or located in the systole and diastole of the heart. In this case, M
The mode reduces to only two samples in the time direction. This method is easier to calculate and provides gradual thickening information of the ventricular wall when the cube is near in time. The wall thickening analysis in this case is to compare two one-dimensional signals and evaluate the thickening for super-resolution concentration using the method [5] of the prior art.

The color encoding for 3D visualization described above also applies to 2D visualization, where the color encoding is associated with the boundaries of blood regions of the 2D visualization. FIG. 7 shows such a 2D video sequence. The figure shows three virtual M-mode lines 73A, 73 respectively associated with associated M-mode displays 74A, 74B, 74C.
Although only B, 73C is shown, a similar M-mode display should be present for every point or position on the endocardial surface of ventricle 70. Next, each independent M-mode display 7
4A, 74B, and 74C are the corresponding M- in the case of three dimensions.
It is processed using the same techniques described above to obtain the mode indications 94A, 94B, 94C. The M-mode line in this example is associated with a point or position on the surface of the ventricle wall, the direction being perpendicular to the wall of the ventricle. The direction of the local M-mode is calculated as the direction obtained by performing a two-dimensional or three-dimensional distance transformation on a binary two-dimensional or three-dimensional image showing the position of a point on the ventricle wall. See reference [4] for information on distance transforms.

In summary, the invention described above provides a method of calculating an anatomical M-mode display based on a time series of 2D or 3D ultrasound images. The method is used to examine living tissue during exercise, such as cardiac action. Used mainly in hospitals. Anatomical M
-Mode display is calculated in real time at the time of video acquisition,
It is calculated as a post-processing of a 2D or 3D video loop. The anatomical M-mode is derived as a virtual M-mode measurement along any title M-mode line. Multiple simultaneous M-mode lines and M-mode displays can be identified. The ability to position the M-mode line arbitrarily makes it independent of the acoustic window that limits the positioning of the M-mode in the prior art, allowing for anatomically meaningful M-mode measurements. The position of the M-mode line can be moved as a function of time to compensate for global motion. In this way, the M-mode line can be positioned perpendicular to the heart wall during the entire cardiac cycle. This property increases the value of the M-mode in wall thickening analysis, since misunderstandings of thickening caused by oblique measurements are avoided. Furthermore, the reference points of the image area can be fixed in the M-mode display so that the relative changes are better visualized. In the 3D image loop, M-modes can be calculated locally at all points on the ventricular wall along the M-mode line normal to the endocardial surface. These local M-
The modes are used to assess wall thickening and to utilize these measurements for color encoding of the endocardial surface.

[Brief description of drawings]

FIG. 1 is a schematic diagram showing the calculation of a prior art M-mode display.

FIG. 2 is a schematic diagram illustrating the inventive concept of an anatomical or virtual title M-mode line for the calculation of the corresponding M-mode representation.

FIG. 3 is a diagram showing a configuration having a large number of M-mode lines according to an embodiment of the present invention.

FIG. 4 shows that moving the position of the M-mode line as a function of position during the cardiac cycle can be used for motion compensation.

FIG. 5 shows an anatomical M-mode with no identified reference points.

FIG. 6 shows the anatomical M-mode line when the reference point is pinned and fixed at a given vertical position in the display of the anatomical M-mode.

FIG. 7 shows a thickening analysis in a configuration with 3 simultaneous anatomical M-mode displays.

FIG. 8 shows how the anatomical M-mode display is calculated with the position of the M-mode line fixed during the cardiac cycle.

FIG. 9 is a schematic diagram showing how the color encoding of the ventricular wall indicating wall thickening is calculated in 4D ultrasound imaging.

[Explanation of symbols]

11,21 Ultrasonic transducer 12 Ultrasonic image 13 Acoustic beam 14,24,34,74,94 M-mode display 23,33,43,63,73,93 M-mode line 66 Reference point 88 Cut plane 92 data set

Continued Front Page (72) Inventor Bjorn Orstad Osbegen 39 G, Lanheim, Norway

Claims (11)

[Claims] 1. A method of producing an anatomical M-mode display in an ultrasound examination performed using an ultrasound transducer (11, 12) on a living biological tissue in motion, eg cardiac action. 2D (2D) or 3D (3D) ultrasound image (2D
2, 32), arranging the time series to form a dataset, and registering at least one virtual M-mode line (23, 33A, 33B) with the dataset. ), And performing computer processing on the data set such that an interpolation operation is performed along the virtual M-mode line based on the at least one virtual M-mode line,
And the anatomical M-mode display (2
4, 34A, 34B) on the display unit. 2. A method according to claim 1, characterized in that the virtual M-mode line (s) can be chosen such that they do not coincide with the direction of the ultrasonic beam of the transducer. Method. 3. The method according to claim 1, wherein the position and / or the direction of the virtual M-mode line can move in response to a rhythmic movement of biological tissue, particularly a rhythmic movement in a cardiac cycle. A method characterized by. 4. A method according to any one of claims 1 to 3, wherein a reference point relating to the 2D or 3D ultrasound image is selected in the resulting anatomical M-mode display. The method is characterized as being a basis for fixing corresponding reference points on a defined vertical axis. 5. The method according to claim 1, which is used for examination of the wall of the left ventricle of the heart.
Or the anatomical M-mode associated with each position on the left ventricular wall surface in the 3D image is first calculated to present a staggered evolution of the cardiac cycle, and then the calculated anatomical M-mode.
The method, wherein each mode is characterized as being color encoded at each of the locations on the left ventricular wall surface. 6. The method of any of claims 1-4, wherein the anatomical M-mode associated with each position on the left ventricular wall surface in a 2D or 3D image is first a systole. Time frame and diastolic frame, the calculated anatomical M-modes are each calculated to be limited to the difference between the two video frames, and then the left ventricle wall surface is calculated. A method characterized in that each location is characterized as being color encoded. 7. The method of claim 1, wherein the anatomical M-mode associated with each position on the left ventricular wall surface in a 2D or 3D image is first of all a whole heart. Calculated to present a salient time interval, such as a cycle, and then each calculated anatomical M-mode is characterized to be color encoded at the surface of the left ventricle wall. A method characterized by the following. 8. The method according to claim 5, wherein the local or global thickening of the wall of the left ventricle is measured along the virtual M-mode line, and the measurement result is A method characterized by being used for color encoding. (Figs. 7, 9) 9. The method according to claim 5, wherein the temporal intensity variation is measured along the virtual M-mode line and the result of the measurement is used for color encoding. How to characterize. (Figs. 7, 9) 10. The method according to any one of claims 5 to 9, wherein the local direction of the virtual M-mode line is two-dimensional from any position to the closest position on the left ventricular wall. Alternatively, the method is calculated as a direction determined in a three-dimensional distance conversion. 11. The method according to claim 1, wherein the result of the computer processing involving the interpolation is:
Known image processing for edge enhancement is performed, thus producing an anatomical M-mode representation of the calculated result.