JP4387758B2 - SPECT apparatus and SPECT image reconstruction method - Google Patents

Description

  The present invention is a SPECT for performing tomography called single photon emission computed tomography (SPECT) among nuclear medicine diagnostic apparatuses that obtain images by detecting radiation emitted from nuclides administered to a subject. The present invention relates to an apparatus, and more particularly, to a SPECT apparatus that executes an image reconstruction method that eliminates or suppresses resolution degradation depending on a distance from a collimator provided in a detector.

  Coupled with recent technological advances in hardware and software, nuclear medicine diagnostic equipment as medical equipment has also made great progress, especially with an emission CT (ECT) device that obtains tomographic images. . One form of emission CT is single photon ECT (SPECT).

  A SPECT apparatus for performing SPECT basically passes a collimator that controls the incident direction of γ rays to detect γ rays emitted from a nuclide (RI) administered to a subject, and the collimator. A scintillator that converts γ rays into an optical signal, a photomultiplier tube that converts an optical signal output from the scintillator into an electrical signal, and a data collection unit that accumulates data obtained from the photomultiplier tube as projection data A gamma camera provided; and a reconstruction unit that reconstructs projection data output from the gamma camera to generate a distribution image of nuclides in the subject as a SPECT image.

  In this SPECT apparatus, as the distance from the collimator to the nuclide increases, a phenomenon occurs in which the resolution of the position of the nuclide deteriorates. For this reason, correction (this is called collimator aperture correction) is necessary to prevent such resolution degradation. Conventionally, several methods have been used as a reconstruction method with collimator aperture correction.

  One of them is a method called “circular fan beam high-resolution direct reconstruction method” as shown in Patent Document 1. When reconstructing the projection data by the fan beam collimator by the direct reconstruction method, a region not called a circular backprojection is set in the reconstruction region far from the collimator surface of the gamma camera, This is a method of reconstructing using only data with less blur due to the collimator opening from the remaining area.

As shown in Non-Patent Documents 1 and 2, another reconstruction method is “OS-EM (Ordered subsets-expectation maximization) reconstruction method (or ML-EM (maximum likelihood expectation maximization) reconstruction method)” It is a technique called. This is a reconstruction method that incorporates the effect due to the collimator aperture. Specifically, it incorporates a blur function that assumes the effect due to the collimator aperture in the process of the successive approximation calculation, and calculates the distance from the collimator surface. Reconfiguration including the function of the dependent correction, that is, the collimator aperture correction is performed.
JP-A-4-370784 Video information (M) September 1998 Vol. 30 No. 18, pp. 1181-1123, Hiroyuki Shinohara et al., “Image Reconstruction Method Outline and Significance of ML-EM Algorithm” Hudson HM and Larkin RS: Accelerated image reconstruction using ordered subsets of projection data IEEE Trans Med Imag 1994: MI-13: 601-609

  However, among the above-described collimator aperture corrections, the “non-circular fan beam high-resolution direct reconstruction method” is a method of reconstructing projection data by fan beam collection, and projection data by parallel beam collection is not taken into consideration. It wasn't.

  Further, since the “non-circular fan beam high-resolution direct reconstruction method” directly reconstructs projection data based on the fan beam, it is difficult to correct the attenuation of γ rays of the human body.

  On the other hand, the OS-EM reconstruction method (or ML-EM reconstruction method) described above can incorporate attenuation correction, but instead, the amount of computation is enormous and the processing time required for reconstruction is long. It becomes unsuitable for routine inspection.

  Therefore, the present invention has been made in view of the current state of the SPECT apparatus described above, and the “non-circular fan beam high-resolution direct reconstruction method” is applied to the projection data obtained by parallel beam acquisition at the time of image reconstruction. It is a first object of the present invention to provide a SPECT apparatus and a SPECT image reconstruction method that perform collimator aperture correction using a method for limiting the reconstruction area in this way and prevent the resolution of the reconstructed image from deteriorating. .

  Furthermore, the present invention can perform attenuation correction at the time of reconstruction simultaneously with the above-described object, and the processing time is shorter than that of the OS-EM reconstruction method (or ML-EM reconstruction method). A second object is to provide a SPECT apparatus and a SPECT image reconstruction method that can be completed.

  In order to achieve the first object described above, according to the SPECT image reconstruction method according to one aspect of the present invention, a radiation detector emits radiation emitted from a nuclide administered into a subject through a collimator. In the SPECT image reconstruction method for reconstructing the image of the area by back projecting the projection data obtained by this detection onto the reconstruction area based on the FBP method, Backprojecting the projection data onto the reconstructed region using a weight function that prioritizes the semicircular region closer to the collimator when the center of the region is the origin than the remaining semicircular region It is characterized by. Further, according to the present invention, a SPECT apparatus having a configuration that exhibits the same function as this is also provided.

In order to achieve the second object described above, according to the SPECT image reconstruction method according to another aspect of the present invention, a radiation detector emits radiation emitted from a nuclide administered into a subject via a collimator. In the image reconstruction method for SPECT in which the projection data obtained by this detection is reconstructed based on the Chang successive approximation method, the center of the region of the reconstruction region is determined. A semicircular area closer to the collimator at the time of the origin is obtained with a weight function that gives priority over the remaining semicircular area, and a limited area attenuation correction map is obtained from the projection data using the weight function and the attenuation correction map. The image of the reconstruction area is reconstructed based on the Chang sequential approximation method using the limited area attenuation correction map. Further, according to the present invention, a SPECT apparatus having a configuration that exhibits the same function as this is also provided.

  According to the present invention, with the configuration according to the above-described one aspect, the projection data obtained by parallel beam acquisition is reconstructed as “non-circular fan beam high-resolution direct reconstruction method” at the time of image reconstruction. Collimator aperture correction can be performed using a method for limiting the region, and deterioration of the resolution of the reconstructed image can be prevented.

  In addition to the above-described effect, the configuration according to another aspect described above can perform attenuation correction at the time of reconstruction, and compared with the OS-EM reconstruction method (or ML-EM reconstruction method). The processing time can be shortened.

  Embodiments of a SPECT apparatus and a SPECT image reconstruction method according to the present invention will be described below.

(First embodiment)
FIG. 1 shows an outline of the configuration of the SPECT apparatus according to the first embodiment. As shown in the figure, the SPECT apparatus includes a gantry 11, a preprocessing device 12, an image reconstruction device 13, a display device 14, an input device 15, and a data collection control device 16. In FIG. 1, reference numeral F indicates a virtual focus.

  Among these, the gantry 11 includes a gamma camera 20 having a detection surface for detecting a signal corresponding to γ (gamma) rays emitted from the nuclide (RI) administered to the subject P as projection data. The gamma camera 20 has a two-dimensional detection unit using, for example, the above-described scintillator, detects γ-rays incident on the detection surface, processes the detection signal, and projects a two-dimensional pixel region (detection surface). Data can be collected. The gamma camera 20 may be configured to include a detection unit in which semiconductor cells are arranged in a two-dimensional array.

  A collimator 21 is attached to the detection surface of the gamma camera 20. The gamma camera 20 and the collimator 21 are supported so that γ rays can be detected from the direction of 360 degrees while rotating around the subject integrally. Note that one gamma camera is not necessarily required, and a plurality of gamma cameras are provided, and 360 degrees directions are shared by each gamma camera and detected from 360 degrees as a whole. I can do it.

  The collection of projection data by the gamma camera 20 is controlled by the data collection controller 16. The data collection control device 16 can control the data collection operation by the gamma camera 20 based on desired data collection conditions given from the input device 15.

  Projection data corresponding to γ rays detected two-dimensionally by the gamma camera 20 is sent to the preprocessing device 12 and temporarily stored in its built-in memory. The preprocessing device 12 performs predetermined preprocessing (uniformity correction, rotation center correction, preprocessing filtering, etc.) on the stored projection data, and then sends the data to the image reconstruction device 13. The image reconstruction device 13 performs a reconstruction process described later on the received projection data to obtain a SPECT image. This SPECT image is displayed on the display 14 in an appropriate manner. The operator can operate the input device 15 to give various conditions regarding information necessary for SPECT imaging, such as data collection, preprocessing, reconstruction, and back projection, to the apparatus.

  Next, the reconstruction process executed in this embodiment will be described.

  This reconstruction processing is a case where attenuation correction is not performed, and a high-resolution SPECT image is obtained by applying a back projection method (limited region back projection method) in which the region to be reconstructed is limited to the projection data of parallel beam acquisition. It is to provide.

  Such reconstruction is executed under the procedure shown in FIG. 2 in cooperation with the preprocessing device 12 and the image reconstruction device 13. Note that the functions executed by the preprocessing device 12 and the image reconstruction device 13 may be provided in a computing device such as one computer device.

  First, the preprocessing device 12 and the image reconstruction device 13 set processing conditions based on operation information given from the operator via the input device 15 (step S1). This processing condition includes setting of a processing range for reconstruction, parameter setting of a preprocessing element such as a preprocessing filter, and the like.

  Next, the preprocessing device 12 reads the projection data collected by the gamma camera 12 and stored in the memory of the device 12 (step S2).

  Next, the preprocessing device 12 preprocesses the projection data (step S3). By this preprocessing, uniformity correction, rotation center correction, preprocessing filtering, and the like are performed.

  Next, sinogram data of the collected data is created by the image reconstruction device 13 (step S4), and convolution with a reconstruction filter (filter such as Ramachandran, Shepp & Logan) is executed (step S5).

  Next, the image reconstruction device 13 creates a weight function used during limited area backprojection (step S6). This weighting function has a high weight in the semicircular area on the side closer to the collimator (projection data side) around the center position of the reconstructed area of the projection data, and is opposed to that area. In the semicircular region on the opposite side (opposite projection data side), the weight is low.

  For example, as shown in FIG. 3 (a), the center of the reconstruction area has a weight of 0.5, and the weight is predetermined from the center to weights 1 and 0 on the projection data side and the opposite projection data side, respectively. It is also possible to use a linear weight function that changes linearly with an inclination of. Also, as shown in FIG. 5B, the center of the reconstruction area has a weight of 0.5, and the weight is cosine from the center to weights 1 and 0 on the projection data side and the opposite projection data side, respectively. It may be a cos type weight function that changes along the curve.

  Further, limited image backprojection is executed by the image reconstruction device 13 using the weight function created as described above (step S7). The concept of this limited area backprojection is shown in FIG. As will be described with reference to the figure, for each projection direction, for each pixel in the reconstruction area, the backprojection data is multiplied by a weighting function, and the multiplication result is added to each pixel in the reconstruction area. As a result, the pixels in the semicircular area close to the projection data (ie, the collimator) (ie, the pixels in the area close to the body surface) are back-projected while being weighted so as to be strongly influenced by the projection data, and reconstructed. A SPECT image is obtained.

  The processes in steps S4 to S7 are repeated until the reconstruction based on the limited region backprojection of the SPECT image for the designated number of slices is completed (step S8).

  As described above, the SPECT apparatus according to the present embodiment utilizes the fact that the projection data from the subject part close to the collimator has more information than the projection data from the subject part farther than the collimator. The superiority of the projection data from the near subject area is extracted by a weight function. Unlike the conventional normal FBP method, the same handling is not performed regardless of the amount of information (that is, the weight in the conventional case is constant), and the weight is changed according to the region to be reconfigured. As a result, it is possible to prevent the image resolution from being deteriorated due to the collimator opening.

(Second Embodiment)
Next, a second embodiment of the present invention will be described with reference to FIGS.

  The SPECT apparatus according to the second embodiment is configured in the same manner as the apparatus shown in FIG. 1 described above in terms of hardware. However, the contents of processing executed in cooperation by the preprocessing device 12 and the image reconstruction device 13 are different.

  In the present embodiment, a process of reconstructing projection data using the “Chang's successive approximation method” as the “limited region backprojection method” described above is executed.

  Such reconstruction is executed under the procedure shown in FIGS. 5 and 6 in cooperation with the preprocessing device 12 and the image reconstruction device 13. Note that the functions executed by the preprocessing device 12 and the image reconstruction device 13 may be provided in a computing device such as one computer device.

  First, the preprocessing device 12 and the image reconstruction device 13 set processing conditions based on operation information given from the operator via the input device 15 (step S11 in FIG. 5). This processing condition includes setting of a reconstruction processing range, setting of parameters of a preprocessing element such as a preprocessing filter, selection of an attenuation map (an image representing a distribution of attenuation coefficients of γ rays in a measurement target), and the like.

  Next, the preprocessing device 12 reads the projection data collected by the gamma camera 12 and stored in the memory of the device 12 (step S12).

  Next, the preprocessing device 12 performs preprocessing on the projection data for parallel beam collection (step S13). By this preprocessing, uniformity correction, rotation center correction, preprocessing filtering, and the like are performed.

  Next, sinogram data is created by the image reconstruction device 13 (step S14).

  Further, the image reconstruction device 13 creates a weight function for determining a limited region for reconstruction in the same manner as in the first embodiment (step S15: see FIG. 8).

  Further, the image reconstruction device 13 creates a limited area attenuation correction map from the selected attenuation map (step S16). For this creation, 360-degree projection data is used, but the attenuation of only the image reconstruction area for a semicircle close to the body surface of the subject is obtained.

  That is, the attenuation with respect to the pixel position of the attenuation map is calculated for each projection angle in the same manner as when the attenuation correction map of the entire area of the reconstruction area is obtained. FIG. 7A shows an attenuation distribution at a projection angle of 0 degree with respect to the cylindrical phantom. By multiplying this attenuation distribution by the weight function shown in FIG. 8 (see FIG. 7B), the attenuation distribution for the limited region of reconstruction is obtained (see FIG. 7C). In this way, an attenuation distribution for 360 degrees is obtained, and as shown schematically in FIG. 9 (a), the attenuation distribution of the cross section is calculated by averaging each other. When this attenuation distribution is obtained, an attenuation correction map is obtained by calculation from this attenuation distribution.

  When this series of preparations is completed, the image reconstruction device 13 executes limited area image reconstruction with attenuation correction using the Chang's successive approximation method (step S17).

  An outline of the image reconstruction process is shown in FIG. First, reconstruction of 0th order approximation is performed (step S17A). This reconstruction is performed by doubling the pixel value of the image reconstructed by limited angle region back projection without performing attenuation correction and multiplying this image by the limited region attenuation map.

  Next, projection data is created (step S17B). This is executed by creating projection data for 360 degrees in consideration of attenuation based on the attenuation map from the nth-order approximate image. In creating this projection data, the data of all the reconstruction areas are added with the same weight as in the ordinary Chang successive approximation method.

  Next, error data is calculated (step S17C). That is, the calculation is performed by subtracting the projection data calculated as described above from the original projection data for each pixel.

  Thereafter, the error data is convolved using the convolution function used in the reconstruction (step S17D). Further, error data is reconstructed (step S17E). That is, the pixel value of the reconstructed image reconstructed by the limited angle region back projection without attenuation correction is doubled, and the reconstructed image is multiplied by the limited region attenuation correction map.

  Further, the data obtained by reconstructing the error data is added to the reconstructed data of the immediately previous order (step S17F). The processing of steps S19B to S19F is repeated for the designated order (step S17G), and when the processing for the designated order is completed, the final SPECT reflecting the signal components up to the designated order of each slice is performed. An image is obtained.

  Note that the processing of steps S14 to S17 in FIG. 5 is repeatedly executed for the designated number of slices, and a SPECT image in which attenuation correction of each slice is performed is obtained (step S18).

  As described above, by applying the limited area back projection method using the weight function according to the present invention to the Chang successive approximation method, in addition to the function and effect obtained in the first embodiment described above, it is possible to accurately. Attenuation correction can be performed, and a high-quality SPECT image can be provided. For this reason, it is possible to overcome the situation where it is difficult to perform attenuation correction, which has been faced by the conventional non-circular fan beam high resolution SPECT reconstruction method. In addition, there is an advantage that the processing time can be shortened as compared with the OS-EM reconstruction method (or ML-EM reconstruction method).

The block diagram explaining the outline | summary of the SPECT apparatus which concerns on embodiment of this invention. The flowchart explaining the outline | summary of the reconstruction process which concerns on 1st Embodiment. The graph which illustrates a weight function. The figure explaining the back projection which concerns on 1st Embodiment. The flowchart explaining the outline | summary of the reconstruction process which concerns on 2nd Embodiment. 12 is a flowchart for explaining in more detail attenuation correction reconstruction performed in the reconstruction processing according to the second embodiment. The figure explaining creation of a limited area attenuation correction map. The graph explaining an example of a weight function. Another figure explaining creation of a limited area attenuation correction map.

Explanation of symbols

11 Stand 12 Pre-processing device 13 Image reconstruction device 14 Display device 15 Input device 21 Gamma camera 22 Collimator P Subject

Claims (8)

Radiation emitted from the nuclide administered into the subject is detected by a radiation detector via a collimator, and projection data obtained by this detection is back-projected onto a reconstruction area based on the FBP method. In the SPECT image reconstruction method for reconstructing an image of
Of the reconstruction area, the projection data is reconstructed using a weight function that gives priority to the semicircle area closer to the collimator when the center of the area is the origin than the remaining semicircle area. A SPECT image reconstruction method comprising backprojecting to a region. The weighting function is a weighting function set so that the weight gradually decreases from a semicircular region closer to the collimator with respect to the origin of the image reconstruction region to the remaining semicircular region. Item 2. The SPECT image reconstruction method according to Item 1. 3. The SPECT image reconstruction method according to claim 2, wherein the weight function is either a weight function in which the degree of weight reduction changes linearly or a weight function that changes along a cosine curve. . SPECT which detects radiation emitted from the nuclide administered into the subject with a radiation detector via a collimator, and reconstructs an image of the reconstruction area based on the projection data obtained by this detection based on the Chang sequential approximation method In the image reconstruction method for
Among the reconstructed regions, a semicircular region closer to the collimator when the center of the region is used as the origin is obtained as a weight function that gives priority over the remaining semicircular region ,
Obtain a limited area attenuation correction map from the projection data using the weight function and attenuation correction map ,
An image reconstruction method for SPECT, wherein the image of the reconstruction area is reconstructed based on a Chang successive approximation method using the limited area attenuation correction map. The weighting function is a weighting function set so that the weight gradually decreases from a semicircular region closer to the collimator with respect to the origin of the image reconstruction region to the remaining semicircular region. Item 5. The SPECT image reconstruction method according to Item 4. A radiation detector that detects the radiation emitted by the nuclide administered into the subject through a collimator is provided, and projection data detected by the radiation detector is back-projected onto a reconstruction area based on the FBP method. In a SPECT device configured to reconstruct an image of a region,
A means for creating a weighting function that gives priority to the semicircular area closer to the collimator when the center of the area is the origin of the reconstruction area than the remaining semicircular area, Means for using the projection data to project back the projection data onto the reconstruction area. The weighting function is a weighting function set so that the weight gradually decreases from a semicircular region closer to the collimator with respect to the origin of the image reconstruction region to the remaining semicircular region. Item 7. The SPECT apparatus according to Item 6. A radiation detector that detects the radiation emitted by the nuclide administered into the subject through a collimator, and the projection data detected by the radiation detector is reconstructed based on the Chang sequential approximation method. In the SPECT apparatus designed to do
Means for obtaining a weighting function that prioritizes the semicircular area closer to the collimator when the center of the area is the origin of the reconstructed area over the remaining semicircular area ;
Means for obtaining a limited area attenuation correction map from the projection data using the weighting function and attenuation correction map ;
A SPECT apparatus comprising: means for reconstructing an image of the reconstructed area based on a Chang successive approximation method using the limited area attenuation correction map.

Images