JP7291466B2 - medical image processor - Google Patents

Description

An embodiment of the present invention relates to a medical image processing apparatus.

Nuclear medicine diagnostic equipment such as SPECT (Single Photon Emission computed tomography) equipment selectively directs chemicals (blood flow markers, tracers) containing radioisotopes (hereinafter referred to as RI) to specific tissues and organs in the body. Gamma rays emitted from RI distributed in the living body are detected by a gamma ray detector arranged outside the living body by utilizing the property of being taken in. A nuclear medicine diagnostic apparatus can provide a functional image of internal organs by generating a nuclear medicine image in which the dose distribution of gamma rays detected by a gamma ray detector is imaged.

A collimator is detachably provided in the gamma ray detector. It is known that the collimator affects spatial resolution and sensitivity depending on the type of aperture shape, aperture arrangement, and the like. Therefore, it is preferable to change the type of collimator used in the gamma ray detector according to the inspection site and inspection method.

However, it is difficult to image the same subject with a plurality of types of collimators because the same subject must be imaged by the number of types of collimators.

U.S. Pat. No. 7,385,200

The problem to be solved by the present invention is to generate medical data based on a different type of collimator from the collimator used for imaging.

A medical image processing apparatus according to an embodiment includes an acquisition unit and a processing unit. The acquisition unit acquires first medical data based on a first collimator generated based on gamma rays emitted from a drug administered into the subject. The processing unit inputs the first medical data to a trained model that generates second medical data based on a second collimator of a type different from the first collimator, based on the first medical data. By doing so, the second medical data is generated.

1 is a block diagram showing a configuration example of a medical image processing system including a medical image processing apparatus according to a first embodiment; FIG. (a) is an explanatory diagram showing an example of medical data when an object is imaged using a high-energy parallel-hole collimator, and (b) is an example of medical data when an object is imaged using a pinhole collimator. An explanatory diagram showing . FIG. 4 is an explanatory diagram showing an example of a data flow during learning of a medical data generation function; FIG. 4 is an explanatory diagram showing an example of data flow during operation of the medical data generation function; FIG. 2 is a block diagram showing a configuration example of a medical image processing system including a medical image processing apparatus according to a second embodiment; FIG. FIG. 9 is an explanatory diagram showing an example of a data flow during operation of a plurality of trained models according to the second embodiment; (a) is an explanatory diagram showing an example of a selection accepting image, and (b) is an explanatory diagram showing another example of a selection accepting image. FIG. 11 is a block diagram showing a configuration example of a SPECT apparatus including a medical image processing apparatus according to a third embodiment;

Hereinafter, embodiments of a medical image processing apparatus will be described in detail with reference to the drawings.

The medical image processing apparatus according to this embodiment generates medical data based on a different type of collimator from the collimator used for imaging. As the medical data according to the present embodiment, data generated by a modality using a collimator can be used. X-ray CT data such as an X-ray CT image acquired by a (computed tomography) apparatus can be used. In the following description, an example of using nuclear medicine data acquired by a SPECT apparatus as medical data will be shown.

(First embodiment)
FIG. 1 is a block diagram showing a configuration example of a medical image processing system 1 including a medical image processing apparatus 10 according to the first embodiment. The medical image processing system 1 has a medical image processing apparatus 10, a SPECT apparatus 101, an image server 102, and an X-ray CT apparatus 103 connected to the medical image processing apparatus 10 via a network 100. FIG.

The medical image processing apparatus 10 has an input interface 11 , a display 12 , a memory circuit 13 , a network connection circuit 14 and a processing circuit 15 .

The input interface 11 is composed of general input devices such as a trackball, switch buttons, mouse, keyboard, numeric keypad, etc., and outputs operation input signals corresponding to user's operations to the processing circuit 15 . The display 12 is composed of a general display output device such as a liquid crystal display or an OLED (Organic Light Emitting Diode) display.

The storage circuit 13 has a configuration including a recording medium readable by a processor such as a RAM (Random Access Memory), a semiconductor memory device such as a flash memory, a hard disk, an optical disk, etc., and stores a program used by the processing circuit 15. , parameter data and other data. Some or all of the programs and data in the recording medium of the storage circuit 13 may be downloaded by communication via the network 100, or provided to the storage circuit 13 via a portable storage medium such as an optical disc. may

The storage circuit 13 may also store medical data of the subject acquired via the network 100 .

The network connection circuit 14 implements various information communication protocols according to the form of the network 100 . The network connection circuit 14 connects with other electric devices via the network 100 according to these various protocols. The network 100 means a general information communication network using telecommunication technology, and includes a wireless/wired LAN such as a hospital backbone LAN (Local Area Network), an Internet network, a telephone communication network, an optical fiber communication network, and a cable communication network. Including networks and satellite communication networks.

A medical image processing apparatus 10 is connected to a SPECT apparatus 101, an image server 102, and an X-ray CT apparatus 103 via a network 100 so as to be able to transmit and receive data.

The processing circuit 15 implements a function of centrally controlling the medical image processing apparatus 10 . The processing circuit 15 is a processor that reads out and executes a program stored in the storage circuit 13 to generate medical data based on a collimator of a different type from the collimator used for imaging. . Here, medical data refers to medical images, medical image data, or raw data. An example of handling nuclear medicine data acquired by the SPECT apparatus 101 as medical data will be described below. The nuclear medicine data are nuclear medicine images, nuclear medicine image data, or raw data.

Here, the relationship between the type of collimator used in the radiation detector and the medical data will be described using the type of collimator used in the gamma ray detector as an example.

FIG. 2A is an explanatory diagram showing an example of medical data when an object is imaged using a parallel hole collimator for high energy, and FIG. FIG. 4 is an explanatory diagram showing an example of medical data;

A collimator used in a gamma ray detector is made of a material such as lead or tungsten that does not easily transmit radiation, and is provided with an opening for regulating the direction in which photons fly.

The spatial resolution and sensitivity of the SPECT apparatus 101 change according to the type of collimator used for the gamma ray detector. Here, the type of collimator differs depending on the thickness of partition walls, the number of openings, the field of view, the focal position, and the like. Various types of collimators are known, such as parallel hole collimator, fan beam collimator, cone beam collimator, astigmatic collimator, pinhole collimator, diverging collimator, converging collimator, and slant hole collimator. Parallel-hole collimators are sometimes classified into high-energy parallel-hole collimators with thick partition walls and low-energy collimators with thin partition walls.

If it is possible to image the same subject with multiple types of collimators and observe multiple nuclear medicine images based on each collimator, compared to observing only nuclear medicine images based on one type of collimator, the subject's It is expected that the diagnostic accuracy can be greatly improved.

However, it is difficult to image the same subject with a plurality of types of collimators because the same subject must be imaged by the number of types of collimators. This is because it is necessary to replace the first collimator with the second collimator after imaging with the first collimator.

Since the collimator is made of heavy materials such as lead and tungsten, the replacement work is very troublesome and imposes a heavy burden on the user. In addition, since the RI administered to the subject attenuates according to the time required for the collimator replacement work, the imaging using the first collimator and the imaging using the second collimator must be performed under the same conditions. is extremely difficult to implement.

Therefore, the medical image processing apparatus 10 uses the first collimator to image the subject, based on the first medical data based on the first collimator, which is acquired by imaging the subject, without imaging the subject. generating second medical data based on a second collimator of a different type than the collimator of the first;

Thus, as shown in FIG. 1, the processor of the processing circuitry 15 implements the organizing function 21, the acquiring function 22, and the medical data generating function 23. As shown in FIG. Each of these functions is stored in the storage circuit 13 in the form of a program.

The organizing function 21 acquires training data obtained by classifying medical data into a plurality of types according to parameter information related to medical data. Raw data is data before image reconstruction, and sinogram data, for example, can be used.

Information on parameters related to nuclear medicine data includes, for example, the dose of RI to the subject, the acquisition time of gamma rays during imaging, information on the physique of the subject, information on the imaging target region of the subject, and the like.

Note that the medical image processing apparatus 10 does not have to have the organizing function 21 . When the medical image processing apparatus 10 has the sorting function 21, training data classified according to parameter information related to nuclear medicine data can be used. In this case, a plurality of trained models can be generated according to the classification.

Acquisition function 22 acquires first medical data based on the first collimator. In the examples shown in the embodiments below, the acquisition function 22 acquires first medical data based on a first collimator generated based on gamma rays emitted from RI administered into the subject. The acquisition function 22 is an example of an acquisition unit.

The medical data generation function 23 applies the first medical data to a trained model that generates second medical data based on a second collimator of a type different from the first collimator, based on the first medical data. Inputting data generates second medical data. The medical data generation function 23 is an example of a processing unit.

FIG. 3 is an explanatory diagram showing an example of a data flow during learning of the medical data generation function 23. As shown in FIG.

The medical data generation function 23 sequentially updates the parameter data 32 by performing deep learning with a large amount of training data input.

Training data, nuclear medicine data 411, 412, 413, ..., based on the first collimator (hereinafter referred to as collimator A) constituting the training input data group 41, corresponding to each nuclear medicine data, the A teacher data group 51 composed of ideal nuclear medicine data 511, 512, 513, . Become.

Training data is generated by simulation, for example. In the simulation, the position information of the gamma ray detector and the radiation source are prepared on the computer. As the position information of the radiation source, for example, SPECT images or PET images that are actually captured can be used. This is because the pixel values of these images contain radiation count information, so the position of the radiation source can be simulated based on these images. Further, the positional information of radiation may be RI distribution information included in a human body model created by a user on a computer. Then, based on the position information of the radiation source, a simulation is performed so that the radiation emitted from the position is detected by the gamma ray detector via the collimator, and the nuclear medicine data 411, 412, 413, . . . and nuclear medicine data 511, 512, 513, .

The medical data generation function 23 processes the nuclear medicine data 411, 412, 413, . So-called learning is performed by updating the parameter data 32 so as to approach the data 511, 512, 513, . . . In general, when the change rate of the parameter data 32 converges within a threshold value, it is determined that learning has ended. Hereinafter, the parameter data 32 after learning is particularly referred to as learned parameter data 32t.

Note that the type of training input data and the type of input data during operation shown in FIG. 4 should match. For example, when nuclear medicine images are used as nuclear medicine data for training input data, nuclear medicine images are also used as input data during operation.

FIG. 4 is an explanatory diagram showing an example of a data flow during operation of the medical data generation function 23. As shown in FIG. During operation, the medical data generation function 23 receives the nuclear medicine data 61 based on the collimator A generated based on the gamma rays emitted from the RI administered into the subject, and generates the collimator B using the trained model 30AB. Generate nuclear medicine data 71B based on.

The neural network 31 and the learned parameter data 32t constitute a learned model 30AB. The neural network 31 is stored in the memory circuit 13 in the form of a program. The learned parameter data 32 t may be stored in the storage circuit 13 or in a storage medium connected to the medical image processing apparatus 10 via the network 100 . When the trained model 30AB (the neural network 31 and the trained parameter data 32t) is stored in the storage circuit 13, the medical data generation function 23 as an example of a processing unit realized by the processor of the processing circuit 15 is stored in the storage circuit 13, the trained model 30AB is read out and executed to generate nuclear medicine data 71B based on the collimator B based on the nuclear medicine data 61 based on the collimator A.

Note that each of the trained models according to the present embodiment, including the trained model 30AB, may be constructed by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array).

In addition, when the medical image processing apparatus 10 has an organizing function 21, the organizing function 21 acquires training data classified according to parameter information (for example, RI dose), and a trained model is generated for each classification. generated. In this case, the type of nuclear medicine data is matched between the training input data and the operational input data, and a trained model corresponding to the classification of the parameter information to which the operational input data belongs is used during operation. By preparing a plurality of trained models 30AB according to parameter information related to medical data and using the trained model 30AB according to the parameter information of the nuclear medicine data 61 based on the collimator A, the collimator B can be more accurately based nuclear medicine data 71B can be obtained.

The medical image processing apparatus 10 according to the present embodiment inputs the nuclear medicine data 61 based on the collimator A into the trained model 30AB, so that the nuclear medicine data 61 based on the collimator B can be obtained without actually performing imaging using the collimator B. Medical data 71B can be generated. Therefore, the user can compare and confirm both the nuclear medicine data 61 based on the collimator A and the nuclear medicine data 71B based on the collimator B, for example, simply by imaging the subject using the collimator A. Therefore, according to the medical image processing apparatus 10 according to the present embodiment, it is possible to greatly improve the diagnostic accuracy of the subject compared to the case of confirming only nuclear medicine data based on one type of collimator.

(Second embodiment)
FIG. 5 is a block diagram showing a configuration example of a medical image processing system 1A including a medical image processing apparatus 10A according to the second embodiment. The medical image processing apparatus 10A shown in the second embodiment differs from the medical image processing apparatus 10 shown in the first embodiment in that a plurality of collimator types are provided according to the learned model. Other configurations and actions are substantially the same as those of the medical image processing apparatus 10 shown in FIG.

The processor of the processing circuit 15A of the medical image processing apparatus 10A implements a sorting function 21, an acquisition function 22, a medical data generation function 23A, a selected image generation function 24, and a selection function 25. FIG. Each of these functions is stored in the storage circuit 13 in the form of a program.

The configuration and action of the sorting function 21 and the acquisition function 22 are the same as those of the first embodiment, so description thereof will be omitted.

FIG. 6 is an explanatory diagram showing an example of a data flow during operation of a plurality of trained models AB, AC, . . . , AG according to the second embodiment. FIG. 7A is an explanatory diagram showing an example of the selection accepting image 75, and FIG. 7B is an explanatory diagram showing another example of the selection accepting image 75. As shown in FIG.

FIG. 6 shows an example in which there are six collimator types B, C, . . . , G after conversion by the learned model. Any number is fine. A trained model is prepared for each type of collimator after conversion.

The medical data generation function 23A generates nuclear medicine data 71B based on the collimator B from the nuclear medicine data 61 based on the collimator A using the trained model 30AB according to the method described in the first embodiment (Fig. 3, See Figure 4). In a similar manner, the medical data generation function 23A converts the nuclear medicine data 61 based on the collimator A from the nuclear medicine data 61 based on the collimator A using the trained models 30AC, . Data 71C, . . . , 71G are generated.

In this way, the medical data generation function 23A converts the nuclear medicine data 61 based on the collimator A from the nuclear medicine data 61 based on the collimators B, C, . 71B, 71C, . . . , 71G.

The selection image generation function 24 generates a selection acceptance image 75 for the user to select at least one collimator from other collimators B, C, . . . For example, the selection acceptance image 75 may be an image for the user to directly select at least one of the other collimators B, C, . . . Alternatively, the selection reception image 75 may be an image for selecting conditions for the medical data to be generated by the user, as shown in FIG. 7B. In this case, for example, if sensitivity priority is selected, it is assumed that collimator C is selected. . The selected image generation function 24 is an example of a selected image generation section.

A selection function 25 selects at least one collimator from other collimators B, C, . Specifically, the selection function 25 receives information on the collimator selected by the user based on the selection reception image 75 via the input interface 11 . The selection function 25 is an example of a selection unit.

Then, the medical data generation function 23A uses the learned model corresponding to the collimator selected by the selection function 25 to generate medical data based on the selected collimator, as shown in FIG. Note that when two or more collimators are selected by the selection function 25, the medical data generation function 23 generates medical data corresponding to each of the selected two or more collimators by sequential or parallel processing. Just do it.

The medical image processing apparatus 10A according to the second embodiment also provides the same effects as the medical image processing apparatus 10 according to the first embodiment. In addition, the medical image processing apparatus 10 according to the second embodiment allows the user to easily select a desired collimator from among a plurality of conversion destination collimators by using the selection reception image 75 .

(Third embodiment)
FIG. 8 is a block diagram showing a configuration example of a SPECT device 80 including a medical image processing device according to the third embodiment.

The SPECT apparatus 80 includes an imaging device 81 that performs nuclear medicine imaging of a subject, and a console device 82 as an example of a medical image processing device. The SPECT apparatus 80 shown in the third embodiment differs from the medical image processing apparatus 10 shown in the first embodiment in that it can use medical data 61 based on the collimator A, which is obtained by imaging the subject itself. Other configurations and actions are substantially the same as those of the medical image processing apparatus 10 shown in FIG. Further, since the processing related to the generation of the learned model 30AB is the same as that of the medical image processing apparatus 10 shown in FIG. omitted.

The imaging device 81 has an imaging system for performing nuclear medicine imaging of a subject using a gamma ray detector set with a collimator A, and displays nuclear medicine data 61 based on the collimator A of the subject obtained by imaging on a console device. Give to 82.

An acquisition function 22n of a processing circuit 15n of a console device 82 as an example of a medical image processing device acquires nuclear medicine data 61 of a subject from an imaging device 81 . The medical data generation function 23n generates nuclear medicine data 71B based on the collimator B based on the nuclear medicine data 61 based on the collimator A of the subject using the trained model 30AB.

Also with the SPECT apparatus 80 according to the third embodiment, similarly to the medical image processing apparatus 10 according to the first embodiment, the collimator B is used by inputting the nuclear medicine data 61 based on the collimator A into the learned model 30AB. The nuclear medicine data 71B based on the collimator B can be generated without actually performing the imaging. Also, the processing circuit 15n of the console device 82 may have the same function as the processing circuit 15A of the medical image processing apparatus 10A according to the second embodiment. In this case, the SPECT device 80 has the same effect as the second embodiment.

According to at least one embodiment described above, it is possible to generate medical data based on a different type of collimator from the collimator used for imaging.

In the above embodiment, the word "processor" is, for example, a dedicated or general-purpose CPU (Central Processing Unit), a GPU (Graphics Processing Unit), or an application specific integrated circuit (ASIC), Circuits such as programmable logic devices (eg, Simple Programmable Logic Devices (SPLDs), Complex Programmable Logic Devices (CPLDs), and FPGAs) shall be meant. The processor implements various functions by reading and executing programs stored in the storage medium.

Further, in the above embodiments, an example of a case where a single processor of the processing circuit realizes each function is shown, but a processing circuit is configured by combining a plurality of independent processors, and each processor realizes each function. good too. Further, when a plurality of processors are provided, a storage medium for storing programs may be provided individually for each processor, or a single storage medium may collectively store programs corresponding to the functions of all processors. good too.

It should be noted that although several embodiments of the invention have been described, these embodiments are provided by way of example and are not intended to limit the scope of the invention. These embodiments can be implemented in various other forms, and various omissions, replacements, and modifications can be made without departing from the scope of the invention. These embodiments and their modifications are included in the scope and spirit of the invention, as well as the scope of the invention described in the claims and equivalents thereof.

10, 10A medical image processing apparatus 12 displays 15, 15A, 15n processing circuits 22, 22n acquisition functions 23, 23A, 23n medical data generation function 24 selection image generation function 25 selection functions 30AB, 30AC, . 61 medical data 75 selection reception image

Claims (7)

an acquisition unit that acquires first medical data based on a first collimator generated based on gamma rays emitted from a drug administered into a subject;
inputting the first medical data to a trained model that generates second medical data based on a second collimator of a type different from the first collimator, based on the first medical data; a processing unit that generates the second medical data by
with
displaying on a display first medical data based on the first collimator and second medical data based on the second collimator ;
Medical image processing equipment. The first collimator and the second collimator are
at least one of partition thickness, number of apertures, field of view, and focal position are different from each other;
The medical image processing apparatus according to claim 1. The acquisition unit
further obtaining parameter information related to the first medical data;
The processing unit is
A trained model for generating the second medical data based on the second collimator based on the first medical data, wherein, among a plurality of trained models prepared according to the parameter information, generating the second medical data based on the second collimator based on the learned model corresponding to the obtained parameter information and the input of the first medical data;
The medical image processing apparatus according to claim 1 or 2. The information of said parameters is
the dose of the drug to the subject, the collection time of the gamma rays when the first medical data is generated, information about the physique of the subject, and the site of the subject included in the first medical data; is at least one
The medical image processing apparatus according to claim 3. a selection unit that selects at least one collimator from other collimators, including the second collimator, of a type different from the first collimator;
further comprising
The processing unit is
A trained model for generating medical data based on the other collimator of a type different from the first collimator based on the first medical data, wherein the trained model is prepared for each type of the other collimator. generating medical data based on the selected collimator based on the trained model corresponding to the selected collimator among the models and the input of the first medical data;
The medical image processing apparatus according to any one of claims 1 to 4. a selection image generation unit that generates a selection reception image for a user to select at least one collimator from the other collimators including the second collimator and displays it on a display;
further comprising
The selection unit
Information on the collimator selected by the user based on the selection acceptance image is received via an input unit, and the collimator selected by the user is the selected collimator.
The medical image processing apparatus according to claim 5. an acquisition unit that acquires first medical data based on the first collimator;
inputting the first medical data to a trained model that generates second medical data based on a second collimator of a type different from the first collimator, based on the first medical data; a processing unit that generates the second medical data by
with
displaying on a display first medical data based on the first collimator and second medical data based on the second collimator ;
Medical image processing equipment.

Images