Patent Application
20060241387
1. Field of the Invention
The present invention relates to a patient positioning system and a patient positioning method which are used in radiation treatment for irradiating various kinds of radiations, such as an X-ray and a particle beam represented by a proton beam, to an affected part of the body for treatment.
2. Description of the Related Art
Recently, radiation treatment has been more widely practiced aiming at bringing tumor cells to death by irradiating various kinds of radiations to the tumor cells. While an X-ray is most prevalently used as the radiation in such treatment, the treatment using a particle beam, such as a proton beam, is also practiced.
One of important processes in the radiation treatment is to make precise positioning of a patient. The term “patient positioning” generally means a process performed by a medical engineer or a doctor through the steps of comparing a DRR (Digital Reconstructed Radiograph) image which is outputted from a treatment planning unit with a DR (Digital Radiograph) image which is taken before the start of radiation irradiation by using an X-ray image capturing device while a patient is lying on a treatment bed, computing a deviation between the patient position decided in the treatment planning and the current patient position on the treatment bed, and determining a shift amount of the treatment bed and moving the treatment bed such that the two images are matched with each other.
The DRR image is an image simulating the DR image and is produced from a CT image taken in the treatment planning stage. Hereinafter, the DRR image and the DR image are referred to also as a reference image and a comparative image, respectively, as required. Note that, in the patient positioning process, an image taken by an X-ray simulator, etc. is sometimes used as the reference image instead of the DRR image.
Hitherto, the so-called reference point (computing point) input method has been used as a method for computing a deviation between the current patient position and the position decided in the treatment planning stage. That method is performed through the steps of displaying the reference image and the comparative image on a monitor of a computer, designating corresponding points (which are selected as clearly recognizable points, e.g., bone portions, the designated points being called reference points) in both the images with an system operator by using an input device, e.g., a mouse, for computing a shift amount of the treatment bed, and deciding the shift amount so that an error between those corresponding points is minimized (see, e.g., Patent Document 1: JP,A 2000-510023 (FIGS. 6-7)).
Further, an image comparing method has also been proposed which comprises the steps of comparing pixel values of the reference image and the comparative image with each other, and deciding the shift amount so that, for example, the correlation coefficient or the square of the difference is minimized (see, e.g., Patent Document 2: JP,A 2004-267250 (FIGS. 8-12)).
With recent development of the imaging technology, DR images also tend to have higher resolutions and larger sizes. For example, a DR image of 2048×2048 pixels is outputted from an X-ray image capturing device. When such a large-sized image is displayed in its entirety on an ordinary monitor (e.g., a monitor with resolution of 1280×1024 pixels), the entire image cannot be displayed in size of 100% (i.e., without scaling-down). In the above example of trying to display the image of 2048×2048 pixels on the monitor with resolution of 1280×1024 pixels, an image scaled down to 50% at maximum is displayed on the monitor.
When an operator inputs reference points in accordance with the above-mentioned reference point input method while looking at the whole of the DR image displayed on the ordinary monitor, the reference points are inputted on the scaled-down image as described above. As a result, the following drawback occurs. If the image is displayed in size of 100%, a particular point of 1 pixel can be designated. In the above example, however, because the image is scaled down to 50% in height and width, 4 pixels are selected once. Practically, because buttons, a scroll bar, a menu, etc. for a graphical user interface are also displayed on a monitor screen according to the image display software used, there is a possibility that an actual scaling factor of the displayed image is further reduced and the number of pixels selected once is further increased. Hence, the correspondence between the reference point intended by the operator and the corresponding point on the image becomes ambiguous, thus resulting in a fear that a different point on the image from the intended one is set as the reference point and accuracy in positioning the patient is deteriorated.
On the other hand, with the known image comparing method described above, since the computation is executed regardless of the display scaling factor, it is possible to avoid the above-mentioned problem of a reduction in the positioning accuracy, which is caused by ambiguousness in setting of the reference point. However, because of the process of comparing pixel values over entire areas of the images to be compared, the known image comparing method requires a larger amount of computations and takes a longer time than the known reference point input method.
Thus, any of the known reference point input method and the known image comparing method cannot achieve an improvement in the positioning accuracy and a cut in the positioning time at the same time.
With the view of solving the above-described problems in the related art, an object of the present invention is to provide a patient positioning system and a patient positioning method which can realize an improvement in the positioning accuracy and a cut in the positioning time at the same time.
To achieve the above object, the patient positioning system of the present invention is featured by comprising a display capable of displaying, in scaled-down size, at least one of first image information and second image information that are given as image information containing an affected area in the body of a patient; an input device for designating particular regions in the first image information and the second image information displayed on the display with input operation; an enlarged region setting processor for setting an enlarged region obtained by scaling-up the designated particular region depending on a scaling-down factor of the display; a representative region setting processor for setting a representative region in the set enlarged region as a part thereof; and a processing unit for producing positioning information used for positioning the patient based on position information of the representative region in the first image information and position information of the representative region in the second image information. Thus, the particular region is enlarged depending on the scaling-down factor of the display, the representative region is set in the enlarged region, and the positioning information is produced such that a position deviation between the corresponding the representative region is minimized. It is therefore possible to avoid the problem of a reduction in positioning accuracy even when an operator designates the particular region (reference point) in a scaled-down display state. Further, because of not comparing pixel values over an entire image unlike the known image comparing method, an amount of computations is greatly reduced and a positioning time can be shortened. As a result, an improvement in the positioning accuracy and a cut in the positioning time can be realized at the same time.
According to the present invention, it is possible to realize an improvement in the positioning accuracy and a cut in the positioning time at the same time.
Preferred embodiments of the present invention will be described below with reference to the drawings.
As shown in
Further, as shown in
While the above description is made as transmitting the comparative image directly from the X-ray image capturing device 3 to the patient positioning computer 1, the image data may be transmitted to and stored in the medical image server 4 from the X-ray image capturing device 3, and the stored image data may be read, as the comparative image, out of the medical image server 4 by the patient positioning computer 1.
First, the reference image prepared in the treatment planning stage and stored in the medical image server 4 is read into, e.g., a not-shown storage (memory) (step 21).
Then, the comparative image of the patient 12, which is captured by the X-ray image capturing device 3 in the state of the patient 12 lying on the bed 6, is read into, e.g., a not-shown storage (memory) from the X-ray image capturing device 3 (step 22). The comparative image is also transmitted from the X-ray image capturing device 3 (or the patient positioning computer 1) to the medical image server 4 and stored therein.
The above steps 21 and 22 may be replaced in order or may be executed in parallel.
After the completion of read of both the reference image and the comparative image in the above steps 21 and 22, those images are displayed on the monitor 9 of the patient positioning computer 1 (step 23). At that time, the reference image and the comparative image are displayed in scaled-down sizes, as required, such that the whole of both the images is contained in an image display screen of the monitor 9. In this embodiment, it is assumed, for example, that the reference image and the comparative image are each an image having a size of 2048×2048 pixels (see
Then, it is determined whether reference points (particular regions) have been inputted by the operator (step 24). More specifically, it is determined whether the operator has inputted reference points, which are used for the patient positioning, at respective corresponding positions in the reference image and the comparative image by using the input device 10 on the screen of the monitor 9 on which the whole of the reference image and the comparative image is displayed in size scaled down to 25%. The reference points are preferably set on the images in an entirely distributed way instead of being located in a portion. For that reason, the reference points are usually inputted by the operator in such a state that the whole of the affected area in the body of the patient is displayed. Clearly recognizable portions, e.g., tips of bones, are selected as positions to be inputted as the reference points.
Then, it is determined whether an image scale-up instruction has been inputted from the operator (step 25). More specifically, it is determined whether the operator has inputted an instruction for scaling up the image displayed on the monitor 9 by using the input device 10. If the scale-up instruction has been inputted, the reference point in the form of the circle point is scaled up and displayed as a computing region (enlarged region) in the form of, e.g., a rectangular box (step 26). This scaled-up state is shown in
Then, it is determined whether the set computing region has been corrected by the operator (step 27). In this embodiment, the operator is able to, e.g., expand/contract, move and rotate the computing region for correction thereof. If any correction has been inputted by the operator using the input device 10, the corrected computing region is reflected at once (step 28). While the computing region has a rectangular shape in this embodiment, the shape of the computing region is not necessarily limited to a rectangle and can be defined as any other suitable one including a circle or an ellipse.
After the computing region has been decided in such a way, a representative point (representative region) is set in the decided computing region. A method of computing the representative point will be described below with reference to
For the sake of simplifying the explanation, the computing region is assumed here to be a rectangular region made up of 9 pixels with one side containing 3 pixels, as shown in
In this embodiment, it is assumed that the computing point is given by each of the pixels contained in the computing region, and the computing region is a set of 9 computing points. Weights are allocated to the computing points, respectively, and one representative point is computed based on the computing points contained in the computing region while taking into account the weights. The weights used for computing the representative point are provided by pixel values in the computing region. The reason is as follows. Generally, in the patient positioning process using an X-ray radiographic image, a definite structure, e.g., a bone, is set as the computing point for the patient positioning in many cases, and the definite structure, e.g., the bone, has a higher brightness (larger pixel value) on the X-ray radiographic image. Therefore, the magnitude of the pixel value can be regarded as an index representing the definite structure, e.g., the bone.
More specifically, the weight of the computing point where the pixel value is maximized in the computing region is set to 1, and the weights of the other computing points are set by normalizing their pixel values based on the maximum pixel value. Then, the representative point is computed by multiplying the weights of the computing points by the respective coordinate values of the same computing points and summing the multiplied results. The coordinate values of the computing points used at that time are defined with the center of the computing region set as the origin. By converting the pixel values in the computing region to weights and computing the representative point in consideration of a resultant weight distribution as described above, it is possible to reproduce the point which has been intended by the operator to input. Incidentally, such a weight distribution can be said as being a probability distribution that represents certainty of the input information (namely, the point having the maximum pixel value and given with the weight 1 is the point inputted by the operator with the highest probability).
Thus, with the computing method described above, the representative point is computed by automatically executing the weighting process. In this embodiment, the weights can be corrected by the operator as required. In other words, it is determined prior to computing the representative point whether a correction has been inputted by the operator on the weights of the computing points automatically set from the pixel value distribution in the computing region (step 29). More specifically, the weight distribution in the computing region is graphically displayed on the monitor 9 of the patient positioning computer 1, as shown in
After the weights have been decided in such a way, the representative point is computed per the computing region by the above-described computing method (step 31). Then, computation for the patient positioning is executed (step 32). More specifically, an image shift amount for minimizing an error between the corresponding representative points in the reference image and the comparative image is computed by the least square method using the computed representative points.
After the end of the computation for the patient positioning, the comparative image is rotated and/or moved in accordance with the computed shift amount, and this result is reflected on the image displayed on the monitor 9 (step 33). It is then determined whether the end of the patient positioning has been instructed from the operator (step 34). If the operator has confirmed the result of the patient positioning on the image displayed on the monitor 9 and has instructed the end of the patient positioning by using the input device 10 based on the determination that there are no problems, a shift amount (positioning information) for the patient 12 (or the bed 6) is transmitted to the treatment bed controller 5 via the network cable 2 (step 35). Responsively, the treatment bed controller 5 executes control of the bed driving unit 7 such that the bed 6 is moved for precise positioning of the patient 12.
In the above description, the reference image and the comparative image correspond respectively to first image information and second image information that are stated in claims as image information containing an affected area in the body of the patient.
The above-described patient positioning system of this embodiment can provide advantages as follows.
When the known reference point input method is used to input the reference points in the state of the reference image and the comparative image being displayed on the monitor 9 in scaled-down size as in the embodiment described above, pixels contained in each enlarged reference point (i.e., an enlarged computing region) on the scaled-up image are all selected, some one of those pixels is set as a computing point, and a shift amount is computed by the least square method so that an error between the corresponding computing points in the reference image and the comparative image is minimized. With the known method, therefore, the pixel selected as the computing point is not definite, thus resulting in a risk that a different point on the image from the point intended by the operator is selected as the computing point and the positioning accuracy is deteriorated.
In contrast, with the patient positioning system of this embodiment described above, the reference point inputted by the operator is enlarged and displayed as the computing region depending on the display scaling-down factor of the monitor 9, and the representative point is computed by the computing method in consideration of the pixel value distribution (weight distribution) in the computing region. In other words, the representative point can be set taking into account the probability distribution that represents certainty of the input information entered by the operator in the relevant computing region. Therefore, the desired point can be reproduced which has been intended by the operator at the time of inputting the reference point on the scaled-down image. As a result, by computing a deviation amount of the patient position using the thus-set representative point, it is possible to eliminate uncertainty in the input information of the reference point, which has been caused due to the image scaling-down in the known reference point input method, and to increase the accuracy in the patient positioning.
Further, with this embodiment, as described above in connection with the step 27 in
In addition, the representative point computing method of this embodiment requires just the operation process of multiplying the pixel value by the coordinate value for each of pixels in the computing region which is made up of only several to several tens pixels, and summing the multiplied results. The remaining operation is to compute, similarly to the known reference point input method, the shift amount by the least square method so that an error between the corresponding representative points in the reference image and the comparative image is minimized. Accordingly, the amount of computations required in this embodiment is greatly reduced in comparison with that required in the known image comparing method in which pixel values are compared for all of pixels in image areas of the reference image and the comparative image, each of the image areas having a predetermined size enough to recognize the feature specific to the images. As a result, it is possible to execute the computation for the patient positioning at a higher speed and to precisely position the patient in a shorter time than those required in the patient positioning process based on the known image comparing method.
Thus, this embodiment enables an improvement in the positioning accuracy and a cut in the positioning time to be realized at the same time.
While the first embodiment represents the case where the operator inputs the reference point, this second embodiment represents the case where the operator inputs the computing region. More specifically, when the reference image and the comparative image are displayed on the monitor 9 at a scaling factor of 100% or larger (such as 150% or 200%), for example, there is a possibility that a pointed portion of the skeletal structure or the like, which can be definitely recognized in the image displayed in scaled-down size, is rounded due to the scaling-up and cannot be definitely recognized. This second embodiment is applied to that case. In other words, when the operator designates a region having a certain size and including a point intended by the operator, a representative point is computed from the designated region and the computation for the patient positioning is executed using the representative point.
Steps 21-23 are the same as those shown in
Then, it is determined whether computing regions have been inputted by the operator (step 24′). More specifically, it is determined whether the operator has inputted computing regions (each made up of plural pixels) in the form of rubber bands or the like, which are used for the patient positioning, at respective corresponding positions in the reference image and the comparative image so as to surround the intended portion by using the input device 10 (e.g., the mouse) on the screen of the monitor 9 on which a part of the reference image and/or the comparative image is displayed, while moving or changing over the images with manipulation of the input device 10.
After the inputting of the computing regions, it is determined whether an image scale-up or -down instruction has been inputted from the operator (step 25′). More specifically, it is determined whether the operator has inputted an instruction for scaling down (or up) the image by using the input device 10 in order to confirm, e.g., the positional relationship of the inputted computing region on the entire image. If the scale-down (or -up) instruction has been inputted, it is then determined whether the computing region has been corrected by the operator (step 27).
Subsequent steps 28-35 are the same as those shown in
The other processes in this second embodiment than described above are the same as those in the first embodiment, and hence a description thereof is omitted here.
With the second embodiment described above, by designating a region having a certain size and including a portion intended by the operator, the representative point is computed by the computing method in consideration of the pixel value distribution (weight distribution) in that region. Therefore, even when the reference image and/or the comparative image is displayed in scaled-up size on the monitor 9 and recognizability of the intended portion is deteriorated, it is possible to reproduce a point intended by the operator to be designated as the reference point. Accordingly, the accuracy in the patient positioning can be increased. Further, since the amount of computations is greatly reduced as in the first embodiment, a time required for positioning the patient can be cut. As a result, this embodiment also enables an improvement in the positioning accuracy and a cut in the positioning time to be realized at the same time.
While the second embodiment has been described in connection with the case of displaying the image on the monitor 9 in size of 100% or larger, the present invention is not limited to such an example. Namely, as a matter of course, the operator may designate the computing region in the state where the image is displayed in scaled-down size on the monitor 9. Such an application is very useful, for example, when the operator has poor eyesight and cannot properly input the reference points on the image displayed in scaled-down size on the monitor 9.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2005-071291 | Mar 2005 | JP | national |
