Information processing device, display device, radiation therapy planning device, information processing method, and program

Abstract

To propose a novel method capable of supporting diagnosis, treatment, or the like in consideration of a side effect of radiation therapy, for example.SOLUTION: An information processing device includes a calculation unit that calculates a first value corresponding to positional information on the basis of a model and mathematical information. The model is generated using learning data related to a plurality of subjects and is for calculating the first value related to a side effect of radiation therapy. The mathematical information is related to an organ of a first subject and corresponds to the positional information.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an information processing apparatus and the like.

Background Art

[0002] Conventionally, there has been a technique (radiation therapy) for treating tumors such as lung cancer (for example, non-small cell lung cancer: NSCLC) by radiation therapy (for example, Patent Document 1). However, it is known that radiation pneumonia (hereinafter, appropriately referred to as "RP (radiation pneumonitis)") may occur as a side effect of radiation therapy (which may be regarded as an adverse event).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] For example, when treating lung cancer or the like by the above-described radiation therapy, various structures such as healthy lung tissue, esophagus, and trachea may be close to the tumor site (tumor region), and the treatment may be difficult. In addition, for example, RP has a high risk and occurrence frequency and is often fatal to patients, so the conventional prevention methods are insufficient, and there is a long way to go to overcome this situation.

[0005] The present invention has been made in view of such problems, and one object thereof is to propose a new method capable of assisting diagnosis, treatment, etc. in consideration of side effects of radiation therapy, for example.

Means for Solving the Problems

[0006] According to a first aspect of the present invention, an information processing apparatus includes a calculation unit that calculates a first value corresponding to position information based on a model for calculating a first value related to side effects of radiation therapy generated using learning data regarding a plurality of subjects, and a second value based on mathematical information regarding an organ of a first subject and corresponding to the position information. According to a second aspect of the present invention, an information processing method includes calculating a first value corresponding to position information based on a model for calculating a first value related to side effects of radiation therapy generated using learning data regarding a plurality of subjects, and a second value based on mathematical information regarding an organ of a first subject and corresponding to the position information. According to a third aspect of the present invention, a program for causing a computer to execute causes the computer to calculate a first value corresponding to position information based on a model for calculating a first value related to side effects of radiation therapy generated using learning data regarding a plurality of subjects, and a second value based on mathematical information regarding an organ of a first subject and corresponding to the position information.

Advantages of the Invention

[0007] According to the present invention, for example, diagnosis, treatment, etc. can be appropriately supported in consideration of side effects of radiation therapy.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Figure 14

Figure 15

Embodiments for Carrying Out the Invention

[0009] Hereinafter, an example of an embodiment for carrying out the present invention will be described with reference to the drawings. In the description of the drawings, the same elements may be denoted by the same reference numerals, and redundant descriptions may be omitted. In addition, the components described in this embodiment are merely examples, and are not intended to limit the scope of the present invention thereto.

[0010] [Embodiment] Hereinafter, an example of an embodiment for realizing the information processing technology of the present invention will be described.

[0011] FIG. 1 is a diagram showing an example of the configuration of an information processing system 1A according to an aspect of the present embodiment. The information processing system 1A includes, for example, a medical image diagnostic apparatus 100, a medical image database 200, a radiation treatment apparatus 300, a radiation treatment planning apparatus 400, a radiation treatment image database 500, and an information processing apparatus 10A which is an example of the information processing apparatus 10. These devices may be configured to communicate with each other via, for example, a bus or a communication unit. The communication method of the communication unit may be wired communication or wireless communication.

[0012] The medical image diagnostic apparatus 100 may include, for example, devices such as a simple X-ray apparatus, an X-ray CT (Computed Tomography) apparatus, an MRI (Magnetic Resonance Imaging) apparatus, etc. for acquiring and evaluating morphological information (information on anatomical structures), and devices such as a PET (Positron Emission Tomography) apparatus, a SPECT (Single Photon Emission Computed Tomography) apparatus, etc. (devices for nuclear medicine examinations) for acquiring and evaluating functional information (physiological function information). Further, it may be a diagnostic apparatus for medical images constituted by any one of these various devices or a combination of these various devices. It may be a PET-CT apparatus, a SPECT-CT apparatus, etc. The data of medical images (for example, volume data or image data obtained by imaging the volume data) by the medical image diagnostic apparatus 100 may be stored in the medical image database 200 in association with information such as identification information (ID, name, etc.) and date of a human (hereinafter, may be collectively referred to as "subject") such as a subject or a patient. A human may be an example of a subject. Further, in this specification, a subject may be a target for receiving a medical act such as radiation therapy. The data of medical images may include data of morphological information and data of functional information. Note that the medical image database 200 may be provided in the information processing apparatus 10A.

[0013] The radiation therapy device 300 is a device for performing radiation therapy. For example, it may be a device for IMRT (Intensity Modulated Radiation Therapy) including TomoTherapy for performing IMRT or a multi-leaf collimator (MLC), etc. VMAT (Volumetric Modulated Arc Therapy) may be included in IMRT. However, it is not limited to this. For example, a device for performing three-dimensional conformal radiation therapy (3D-CRT), stereotactic radiation therapy (SRT), etc. may be applied. Also, it may be configured by a combination of these.

[0014] Note that as radiation therapy, treatment using X-rays or γ-rays classified as electromagnetic waves may be applied, or treatment using α-rays, β-rays, electron beams, proton beams, heavy particle beams, neutron beams classified as particle beams may be applied. For example, proton beam therapy, heavy particle beam therapy, neutron beam therapy, etc. may be applied, and these devices may be applied as the radiation therapy device 300. Also, any combination of the above may be applied.

[0015] The radiation therapy planning device 400 may be, for example, a computer device (information processing device) that creates (generates) a radiation therapy plan that is a plan for radiation therapy by the radiation therapy device 300. It may also be referred to as a radiation therapy plan creation device (radiation therapy plan generation device). For example, data of the radiation therapy plan by the radiation therapy planning device 400 (for example, radiation irradiation data or image data obtained by imaging it) may be stored in the radiation therapy image database 500 as radiation therapy image data. Note that the radiation therapy image database 500 may be provided in the radiation therapy planning device 400 or the information processing device 10A.

[0016] The information processing apparatus 10A is, for example, a computer device including a data acquisition unit 11, a model generation unit 13, and a learned model processing unit 15 as functional units, and performs various processes based on data such as medical image data stored in the medical image database 200 and radiation therapy image data stored in the radiation therapy image database 500, which are acquired by the data acquisition unit 11. These functional units may be, for example, functional units (functional blocks) of a processing unit (processing device) or a control unit (control device) (not shown) of the information processing apparatus 10A, and may be configured to include a processing circuit such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), or an FPGA (Field Programmable Gate Array).

[0017] Based on the data acquired by the data acquisition unit 11, the model generation unit 13 generates, for example, by machine learning (e.g., supervised learning), a model related to the prediction of the occurrence of pneumonia (e.g., RP) associated with radiation therapy (hereinafter, appropriately referred to as the "pneumonia occurrence prediction model").

[0018] The learned model processing unit 15 is a processing unit of the learned pneumonia occurrence prediction model (hereinafter, appropriately referred to as the "learned pneumonia occurrence prediction model") generated by the model generation unit 13, and calculates (infers) predetermined quantities related to the occurrence of pneumonia using the learned pneumonia occurrence prediction model. The calculated quantities may be output from the learned model processing unit 15.

[0019] Here, the "output" of information and data (such as information) may include, for example, the output of information to other functional units in the own device (internal output), the output of information to a device other than the own device (external device) (external output), transmission (external transmission), display, sound output, and other concepts.

[0020] FIG. 2 is a diagram for explaining the outline of IMRT. In IMRT, it is configured such that radiation can be applied or not applied to a specific site, and the irradiation range (distribution, spread), for example, can be finely adjusted by a radiation treatment planning device 400 to create a radiation treatment plan. For example, when the site (area) indicated by the hatched circle is the tumor site of lung cancer, the irradiation range can be set in various patterns as shown in the figure. Note that the difference (high or low) in the dose of the radiation to be irradiated is roughly shown by applying four different hatchings.

[0021] As a method for preventing pneumonia in radiation therapy, for example, a method of simply reducing the exposure of lung tissue using a CT image and a method of using a functional image (lung function image) that visualizes the oxygen-carbon dioxide exchange efficiency and preferentially reducing the exposure of lung tissue with high exchange efficiency are conceivable. Radiation therapy using a functional image can be performed, for example, using a nuclear medicine image and a CT image. However, as a problem, there is little evidence that damage to a highly functional lung region (highly functional region) is equal to the occurrence of pneumonia, and there is also a problem that a plurality of functional regions cannot be unified and utilized.

[0022] In the present embodiment, as an example, a pneumonia occurrence prediction model that can numerically predict the likelihood (risk) of the occurrence of pneumonia as a side effect of radiation therapy is generated.

[0023] FIGS. 3 and 4 are diagrams for explaining the principle of generation of the pneumonia occurrence prediction model and the like in the present embodiment. Here, as an example, a case where 4DCT data is acquired by the data acquisition unit 11 is illustrated. The 4DCT data is, for example, time-series three-dimensional volume data and is dynamic data due to the respiration of the subject. Specifically, for example, it may be data composed of 3DCT data at two time points, an exhalation position (peak exhalation position, etc.) and an inhalation position (peak inhalation position) of the subject. In addition, when applying CT, not limited to 4DCT data, the processes described below may be performed based on 3DCT data or 2DCT data as well.

[0024] (1) Learning phase As shown in FIG. 3, for example, the model generation unit 13 includes a function value calculation unit 131, a percentile value conversion unit 133, a fractionation unit 135, a fractionation interval exposure dose calculation unit 137, and a pneumonia occurrence prediction model generation unit 139 as functional units.

[0025] (A) Quantification of function values The function value calculation unit 131 generates lung function data (lung function image) in which the lung ventilation volume is quantified by performing DIR (Deformable Image Registration) and quantitative evaluation (quantitative analysis) on the 4DCT data (4DCT image) acquired by the data acquisition unit 11, for example. The generated lung function data (lung function image) may be stored in the information processing device 10A or a database (not shown) external to it.

[0026] In this case, as a method of quantitative evaluation, for example, a method using HU (Hounsfield Unit) may be applied, and for example, the following formula (1) can be used.

Equation

[0027] In addition, as a method of quantitative evaluation, for example, a method (Jacobian-based metric) using the Jacobian matrix to analyze the degree of volume change of the lung from exhalation to inhalation may be used.

[0028] In addition, in the following processes, lung function information (for example, lung function values) may be used, and the acquisition method is not limited. For example, the lung function values (lung function images) obtained based on the nuclear medicine examination described above may be used.

[0029] (B) Percentile value conversion process The percentile value conversion unit 133 performs a percentile value conversion process of converting the lung function values quantified in the process of (A) (hereinafter, may be simply referred to as "function values"). The percentile value thus obtained is conveniently referred to as the "function percentile value".

[0030] Note that a functional unit for performing noise reduction processing (noise removal processing) on the function values (function images: for example, 4DCT ventilation images composed of function values quantified in HU) quantified in the process of (A) may be configured in the model generation unit 13. Specifically, for example, a filter processing unit that performs filter processing using a median filter (for example, a median filter with a width of 3×3×3 voxels) may be configured in the model generation unit 13. Further, the filter processing unit may perform filter processing using a smoothing filter such as a moving average filter or a Gaussian filter.

[0031] (C) Fractionation process The fractionation unit 135 fractionates (divides) the data (image) in which the position and the function percentile value are associated into a plurality of data (a plurality of images) based on a preset upper threshold value (cutoff upper limit value) and a lower threshold value (cutoff lower limit value) for the function percentile value obtained in the process of (B) (or filter processing). The threshold value may be set automatically by the information processing apparatus 10 (automatic setting), or may be set by the information processing apparatus 10 based on the input of the user (manual setting). The same shall apply hereinafter for various settings.

[0032] Specifically, for example, the cut-off lower limit value is set from the 0th to the 80th at intervals of 20, and the cut-off upper limit value is set from the 20th to the 100th at intervals of 20. Then, the functional percentile values (the functional percentile values associated with positions) included in the interval defined by the cut-off lower limit value and the cut-off upper limit value are extracted. As a result, in this example, data in which positions and functional percentile values corresponding to five intervals of "0 - 20th", "20 - 40th", "40 - 60th", "60 - 80th", and "80 - 100th" are associated are obtained. For convenience, each of the divided intervals is referred to as a "fractionated interval", and the number of fractionated intervals is referred to as the "number of fractionations".

[0033] The fractionated interval exposure dose calculation unit 137 calculates, for each fractionated interval, the exposure dose corresponding to that fractionated interval (hereinafter referred to as the "fractionated interval exposure dose") based on, for example, the data in which the positions and the functional percentile values obtained in the process (C) are associated (in the above example, five data, five functional percentile images).

[0034] Specifically, for example, by image segmentation, region division is performed for that fractionated interval, and the exposure dose for each region (for example, the average dose obtained by averaging the exposure doses for each region) is calculated.

[0035] Note that for each fractionated interval, the ratio (for example, V20Gy) of the lung volume irradiated exceeding a specific exposure dose (set threshold dose: for example, 20 Gy (gray)) to the whole lung may be calculated. And the calculated ratio of the lung volume may be output. Note that the threshold dose may be set to an arbitrary value (5 Gy, 40 Gy, etc.).

[0036] In this case, for example, by using data of a radiation treatment plan (which may also be referred to as data of exposure information (exposure data)) including a calculated exposure dose created by a radiation treatment planning device 400, and superimposing the region-divided image and the image of the radiation treatment plan, the exposure dose for each region may be calculated. However, the exposure dose calculated in this way is merely a calculated and estimated exposure dose (estimated exposure dose). In this example, in order to calculate the exposure dose for each region, the risk value of pneumonia occurrence described later also becomes a value for each region.

[0037] Note that for each fractionation interval, the average or the like of the exposure doses at positions having the same percentile value may be used. Also, for each fractionation interval, the exposure doses calculated for each position may be used.

[0038] Hereinafter, for convenience,[[]] Fractionation interval exposure dose corresponding to the "80 - 100th" fractionation interval = D1 Fractionation interval exposure dose corresponding to the "60 - 80th" fractionation interval = D2 Fractionation interval exposure dose corresponding to the "40 - 60th" fractionation interval = D3 Fractionation interval exposure dose corresponding to the "20 - 40th" fractionation interval = D4 Fractionation interval exposure dose corresponding to the "0 - 20th" fractionation interval = D5 It is expressed as such.

[0039] The pneumonia occurrence prediction model generation unit 139 generates, for example, a pneumonia occurrence prediction model for inferring a value (hereinafter referred to as "pneumonia occurrence risk value") representing the likelihood (risk) of pneumonia occurrence associated with radiation treatment. The pneumonia occurrence risk value may be an example of a first value regarding the side effects of radiation treatment, and may be defined, for example, as a unitless value of 0 or more. Specifically, for example, a pneumonia occurrence prediction model is generated by machine learning using a plurality of data sets, where one data set is a combination of the fractional interval exposure dose calculated by the fractional interval exposure dose calculation unit 137 and the pneumonia occurrence risk value obtained in advance as epidemiological data on the occurrence / non-occurrence of pneumonia (side effect). The pneumonia occurrence prediction model generated in this way is referred to as a "trained pneumonia occurrence prediction model".

[0040] In this embodiment, the pneumonia occurrence prediction model for which the pneumonia occurrence prediction model generation unit 139 performs learning is, for example, a model represented by the following formula (2).

Equation

[0041] In the learning phase, a combination of the fractional interval exposure dose of each fractional interval calculated by the fractional interval exposure dose calculation unit 137 and the pneumonia occurrence risk value is regarded as one piece of data, and for example, data (data sets) for a plurality of subjects may be used as learning data for learning.

[0042] The learning data may be, for example, training data, and the training data may also be referred to as learning data or training data. Note that, in contrast, the learning data may be data including training data and verification data. The verification data may be, for example, data for confirming the evaluation of the model based on a score.

[0043] In this embodiment, the pneumonia occurrence prediction model generation unit 139 generates a pneumonia occurrence prediction model by, for example, Lasso regression. Lasso regression is a method that simultaneously determines the regression coefficient "α" and feature selection (feature selection) based on an objective function that introduces the L1 norm. Using the above learning data, a hyperparameter (λ of the L1 regularization term) representing the magnitude of the penalty and the above regression coefficients "α1" to "α5" can be calculated. Since Lasso regression itself is well-known, detailed explanations are omitted. One of the reasons for using Lasso regression is to eliminate unnecessary features through feature selection. This is more effective when the number of fractions is increased.

[0044] Note that, unlike this embodiment, for example, regression analysis methods such as Ridge regression, Elastic Net regression, Bayesian linear regression, robust regression, maximum likelihood estimation method, and gradient descent method may be used. Also, for example, machine learning methods such as deep learning, which is an advanced form of neural networks, and MLP (Multi Layer Perceptron) (however, here deep learning and MLP are also regarded as a type of machine learning) may be applied. A genetic algorithm may be used.

[0045] Figure 4 is a diagram for explaining the learning process of the model described with reference to Figure 3 using an image. As described above, in reality, the above processing can be performed using 3D volume data, but here it is explained using a 2D image for easier understanding.

[0046] On the left side of the figure, an example of a lung function image (4DCT ventilation image) of a certain subject is shown. Five segmented images obtained by performing the above processing on this function image are shown on the right side of the figure. Note that in this figure, what is shown is the result of binarizing the pixel values after classification based on the functional percentile value. For example, in the "80 - 100th" image, the region shown in white indicates a region with a high function value, that is, a high-function region. From the perspective of images, learning of the model is performed using a plurality of data sets of combinations of respective fractional exposure doses "D1" to "D5" calculated based on superimposing each of the fractionated images thus fractionated and the IMRT image, and, for example, a pneumonia occurrence risk value "Y" based on the above-described epidemiological data.

[0047] (2) Inference phase As shown in FIG. 5, the learned model processing unit 15 has, for example, a function value calculation unit 131, a percentile value conversion unit 133, a fractionation unit 135, a fractional exposure dose calculation unit 137, and a learned pneumonia occurrence prediction model processing unit 159 as functional units.

[0048] The function value calculation unit 131, the percentile value conversion unit 133, the fractionation unit 135, and the fractional exposure dose calculation unit 137 may be the same as those in FIG. 3.

[0049] The learned pneumonia occurrence prediction model processing unit 159 infers a pneumonia occurrence risk value according to Equation (2) using, for example, the fractional exposure doses "D1" to "D5" calculated by the fractional exposure dose calculation unit 137 and the regression coefficients "α1" to "α5" obtained in the learning phase. The inferred pneumonia occurrence risk value is referred to as the "inferred pneumonia occurrence risk value".

[0050] FIG. 6 is a diagram schematically showing the flow of inference of the pneumonia occurrence risk value by an image. The upper part shows the binarized images for each fractional interval shown in FIG. 4. Using the same method as the above-described method, for each fractional interval, fractional exposure doses "D1" to "D5" are calculated. Then, using the calculated fractional exposure doses "D1" to "D5" and the regression coefficients "α1" to "α5" obtained in the learning phase, a pneumonia occurrence risk value is inferred according to Equation (2).

[0051] Based on the association between the position and the inferred pneumonia occurrence risk value, the inferred pneumonia occurrence risk value visualized in the form of a map (visualized as an image) is referred to as a "pneumonia occurrence hazard map". At the bottom of the figure, a two-dimensional pneumonia occurrence hazard map corresponding to the image shown in FIG. 4 is shown, and FIG. 7 shows an enlarged view of this two-dimensional pneumonia hazard map.

[0052] In the two-dimensional pneumonia occurrence hazard map of FIG. 7, the hatched area of the dots indicates the area with the highest risk of pneumonia occurrence, for example, the area of "very dangerous" (very high risk). Also, the lightly shaded area with a drawn contour indicates the area with the next highest risk of pneumonia occurrence, for example, the area of "dangerous" (high risk). For easy understanding, a part of the "dangerous" area in the right lung area facing the drawing surface is marked as "dangerous" with a drawn line. The area of "very dangerous" (very high risk) can be considered to represent the so-called red zone, and the area of "dangerous" (high risk) can be considered to represent the so-called yellow zone.

[0053] In fact, in these two types of areas, there are also multiple areas according to the risk. Since it is a hazard map, for example, the risk can be classified into multiple levels based on the inferred pneumonia occurrence risk value. For example, in this embodiment, since the number of fractions is set to "5", the risk can be classified into five levels accordingly. For example, it can be classified into "very high risk", "high risk", "medium risk", "low risk", "very low risk", etc. in descending order of risk. In this case, for example, stepwise thresholds can be set as thresholds (cutoff values) for the inferred pneumonia occurrence risk value, and threshold determination for the inferred pneumonia occurrence risk value can be performed to classify the risk. Note that the number of levels of risk can be set and changed as appropriate.

[0054] In this case, after setting colors corresponding to the classified risks, a pneumonia occurrence hazard map with the areas color-coded can be generated and displayed. For example, in the case of the above five levels, for example, a pneumonia occurrence hazard map may be generated and displayed such that "very high risk = red", "high risk = orange", "medium risk = yellow", "low risk = green", and "very low risk = blue".

[0055] The risk of pneumonia occurrence can also be considered to be represented by the regression coefficient "α". This is because the area is divided for each fraction interval, and the fraction interval exposure dose is calculated for each area. The area divided into the fraction interval of "α = 0" may be considered an area unrelated to the occurrence of pneumonia, in other words, a "safe" area. As will be described later, in this embodiment, by using lasso regression, in the above example, the regression coefficient "α5" of the fraction interval "0 - 20th" and the regression coefficient "α4" of the fraction interval "20 - 40th" became "0". Therefore, the area divided into "0 - 20th" and the area divided into "20 - 40th" may be considered "safe" areas.

[0056] However, when there is not much difference in the values of the regression coefficient "α" corresponding to each fraction interval, just because the calculated regression coefficient "α" is small, it cannot be simply said that the area divided into that fraction interval is "safe". However, at least the area divided into the fraction interval of "α = 0" may be considered an area unrelated to the occurrence of pneumonia.

[0057] Note that the criterion for "safe" can be set and changed as appropriate. For example, an area where the inferred pneumonia occurrence risk value is less than (or equal to) the smallest threshold among the set step - by - step thresholds may be regarded as "safe".

[0058] Note that instead of color - coding, different hatching may be used for distinction and display. Also, it may be displayed in any of monochrome, black - and - white, grayscale, or color.

[0059] In addition, data in which a position or region is associated with an inferred pneumonia occurrence risk value (numerical value) may be displayed, for example, in the form of a table or a chart.

[0060] In addition, based on a user input such as clicking or touching (when configured with a touch screen) an arbitrary position or region in the pneumonia occurrence hazard map displayed on the display device 600, control may be performed to display the inferred pneumonia occurrence risk value (numerical value) of the corresponding position or region, so that the user can check detailed data.

[0061] In the present embodiment, since the above processing is performed using three-dimensional volume data, the pneumonia occurrence hazard map is generated as a three-dimensional map as shown in FIG. 8, for example. Note that the same applies because three-dimensional lung function data can also be obtained using 4DCT data. In this case, control may be performed to display maps of different arbitrary viewpoints or cross-sections on the display device 600 based on the user input.

[0062] FIG. 9 is a diagram showing an example of the configuration of an information processing system 1B capable of realizing the generation and output (display, etc.) of the above-described pneumonia occurrence hazard map. The information processing system 1B includes, for example, a medical image diagnostic apparatus 100, a medical image database 200, a radiation treatment apparatus 300, a radiation treatment planning apparatus 400, a radiation treatment image database 500, an information processing apparatus 10B which is an example of an information processing apparatus, and a display device 600.

[0063] In the information processing system 1B, the information processing apparatus 10B includes, in addition to a data acquisition unit 11, a model generation unit 13, and a learned model processing unit 15, for example, a pneumonia occurrence hazard map generation unit 17 and a display control unit 19.

[0064] The pneumonia occurrence hazard map generation unit 17 generates a pneumonia occurrence hazard map based on, for example, the inferred pneumonia occurrence risk value output from the learned model processing unit 15. The pneumonia occurrence hazard map generation unit 17 may be said to be a functional unit that generates data for visualizing the inferred pneumonia occurrence risk value as a map.

[0065] The display control unit 19 performs control to cause the pneumonia occurrence hazard map generated by the pneumonia occurrence hazard map generation unit 17 to be displayed on the display device 600.

[0066] The display device 600 may be an output device (display device) that displays various types of information according to the control of the display control unit 19. The display device 600 may be, for example, a component of a device including the information processing device 10B. For example, the display device 600 may be a component (display unit) of the radiation treatment planning device 400. However, it is not limited thereto, and it may be a component (display unit) of a device other than the radiation treatment planning device 400. Also, the display control unit 19 may be configured as a functional unit of a device including the display device 600. In this case, the display control unit 19 may cause the pneumonia occurrence hazard map received from the information processing device 10B by a communication unit (not shown) to be displayed on the display unit, for example. Also, the display device 600 may be a component of the information processing device 10B.

[0067] In the information processing device 10A shown in FIG. 1, the display control unit 19 is configured at a subsequent stage of the learned model processing unit 15, and in the information processing system 1A shown in FIG. 1, a display device 600 that displays information according to the control by the display control unit 19 is configured. Then, the display control unit 19 may perform control to cause the inferred pneumonia occurrence risk value output from the learned model processing unit 15 to be displayed on the display device 600.

[0068] <Process> FIG. 10 is a flowchart showing an example of the procedure of information processing in the present embodiment, and is a flowchart showing an example of the flow of processing related to the generation, verification, and evaluation of a model. The processing in this flowchart may be realized, for example, by a processing unit (control unit) of the information processing apparatus 10B reading a program stored in a storage unit (not shown) into a RAM (not shown) and executing it.

[0069] In the following, each symbol S in the flowchart means a step. Note that the flowchart described below only shows an example of the procedure of information processing in the present embodiment, and other steps may be added, some steps may be deleted, or some steps in the flowchart may be executed after being swapped.

[0070] First, the model generation unit 13 performs pneumonia occurrence prediction model generation processing. In the pneumonia occurrence prediction model generation processing, the model generation unit 13 performs learning processing (S11). Specifically, for example, a plurality of subjects are divided into three groups (Group 1 to Group 3), and using the dataset of the subjects in Group 1 as a training dataset, the model is learned by the method described above.

[0071] Next, the model generation unit 13 performs verification processing for verifying the generated model (S13). Specifically, for example, using the dataset of the subjects in Group 2 as a verification dataset, the generated pneumonia occurrence prediction model is verified according to a predetermined verification method (verification algorithm).

[0072] Next, the model generation unit 13 determines whether the verification result is OK or NG (S15). If it is NG (S15: NG), the process returns to S11. In this case, as an example, the above Group 1 may be changed and re-learning may be performed.

[0073] On the other hand, if the verification result is OK (S15: OK), the model generation unit 13 ends the pneumonia occurrence prediction model generation processing.

[0074] Next, the processing unit performs an evaluation process for evaluating the generated model (S17). Specifically, for example, using the dataset of the subjects in the third group as a test dataset, the generated pneumonia occurrence prediction model is evaluated according to a predetermined evaluation method (evaluation algorithm). As the evaluation method, for example, cross-validation may be used. Then, the processing unit ends the process.

[0075] FIG. 11 is a flowchart showing an example of the information processing procedure in the present embodiment, and is a flowchart showing an example of the flow of processing related to the prediction of the occurrence of pneumonia in a patient using the pneumonia occurrence prediction model generated in FIG. 10. The processing in this flowchart may be realized, for example, by a processing unit (control unit) of the information processing apparatus 10B reading a program stored in a storage unit (not shown) into a RAM (not shown) and executing it.

[0076] First, the data acquisition unit 11 acquires data related to the living body (organ) of a patient (an example of the first subject) who is the subject of diagnosis or treatment, for example (S21).

[0077] Next, the function value calculation unit 131 performs a function value calculation process for calculating a function value based on the data acquired in S21 (S23).

[0078] Thereafter, the percentile value conversion unit 133 performs a percentile value conversion process for converting the function value calculated in S23 into a percentile value (S25).

[0079] Next, the fractionation unit 135 performs a fractionation process for fractionating the percentile value converted in S25 (S27).

[0080] Next, the fractional interval exposure dose calculation unit 137 performs a fractional interval exposure dose calculation process for calculating the fractional interval exposure dose for each fractional interval in S27 (S29).

[0081] After that, the trained pneumonia occurrence prediction model processing unit 159 performs pneumonia occurrence risk value calculation processing for calculating a pneumonia occurrence risk value according to formula (2) based on the fractionated interval exposure dose calculated in S29 and the stored model parameters (regression coefficients) (S31).

[0082] Next, the processing unit determines whether to display the pneumonia occurrence hazard map (S33). If it is determined to display it (S33: YES), the pneumonia occurrence hazard map generation unit 17 performs pneumonia occurrence hazard map generation processing for generating a pneumonia occurrence hazard map (S35).

[0083] Then, the display control unit 19 performs control to display the pneumonia occurrence hazard map generated in S35 on the display device 600 (S37).

[0084] Note that the pneumonia occurrence hazard map may be generated as a two-dimensional map or may be generated as a three-dimensional map. Also, based on the user's input, various maps may be switched and displayed.

[0085] After that, the processing unit determines whether to end the processing (S39). If it is determined to continue the processing (S39: NO), the processing returns to S1. On the other hand, if it is determined to end the processing (S39: YES), the processing unit ends the processing.

[0086] Note that for the processing of parts other than the generation and display control of the pneumonia occurrence hazard map, the processing unit (control unit) of the information processing apparatus 10A shown in FIG. 1 may perform the processing.

[0087] <Model evaluation result> FIG. 12 is a diagram showing an example of the evaluation result of the generated model. The inventor of the present application evaluated the generated pneumonia occurrence prediction model by nested five-fold cross-validation using a dataset of a certain number of patients with advanced lung cancer cases treated in the hospital of the patent applicant of the present application and cooperating medical institutions. Here, as an evaluation index, the case where the area under the curve (hereinafter referred to as the "AUC (Area Under Curve) value") of a function of the exposure dose according to the above-described fractionation interval (dose function) is used is shown. The AUC value is a value in the numerical range of "0 to 1", and the larger the value, the higher the discrimination ability.

[0088] The average AUC value (reference value) calculated based on the method of simply reducing the lung volume to be exposed, which was implemented by the inventor of the present application, was about "0.61", whereas in the test data set independent of the training data set and the verification data set, the average AUC value calculated based on the pneumonia occurrence prediction model generated by the method of the present embodiment was about "0.81", and it was confirmed that a high discrimination ability was shown.

[0089] FIG. 13 shows an example of a table in which the horizontal axis is the relative regression coefficient RRC (Relative Regression Coefficients) and various values are shown on the vertical axis for the generated pneumonia occurrence prediction model. The relative regression coefficient RRC on the horizontal axis is an index value indicating the relative value of the above-described regression coefficient "α", and is calculated, for example, according to the following formula (3).

Equation

[0090] Also, on the vertical axis, those denoted as "fMLD80 - 100th", "fMLD60 - 80th", "fMLD40 - 60th", "fMLD20 - 40th", "fMLD0 - 20th" respectively correspond to the above-described fractionation intervals "80 - 100th", "60 - 80th", "40 - 60th", "20 - 40th", "0 - 20th".

[0091] As described above, in the method focusing on lung function, there was a problem that a plurality of functional regions could not be unified and utilized. On the other hand, in the method of this embodiment, although the relative regression coefficient RRC of the fractionation interval "80 - 100th" corresponding to the high - performance region is the highest at about "0.5", the relative regression coefficient RRC corresponding to the fractionation interval "60 - 80th" and the relative regression coefficient RRC corresponding to the fractionation interval "40 - 60th" are also relatively large values to some extent. It can be seen that information in regions other than the high - performance region is also used. On the other hand, by feature selection using lasso regression, the relative regression coefficients RRC (regression coefficient α) of the fractionation intervals "0 - 20th" and "20 - 40th" become "0" (the fractionation interval exposure dose "D4" (20 - 40th) and the fractionation interval exposure dose "D5" (0 - 20th) are excluded), and it can be seen that this information was not used. Note that when using other regression analysis methods (described above), the fractionation interval exposure dose "D4" (20 - 40th) and the fractionation interval exposure dose "D5" (0 - 20th) may also affect the results.

[0092] <Application Example> In the method of this embodiment, based on the inferred pneumonia occurrence risk value, for example, non - invasively and immediately, the occurrence of pneumonia associated with radiotherapy can be predicted (non - invasiveness, immediacy). For example, there is no need to use special pharmaceuticals, no need to wait for the reaction of pharmaceuticals to be sufficiently obtained, and without additional examinations, the occurrence of pneumonia can be easily predicted from information about the patient's body. Moreover, the method of this embodiment basically does not require additional equipment and can be directly introduced into existing radiotherapy.

[0093] The information processing apparatus 10 of this embodiment may be incorporated into a radiotherapy planning apparatus 400, for example, as shown in FIG. 14. Note that the information processing apparatus 10 of this embodiment may also be a separate apparatus that can communicate with the radiotherapy planning apparatus 400. In this case, the radiotherapy planning apparatus 400 may receive information from the information processing apparatus 10 by, for example, a communication unit.

[0094] The radiation treatment planning device 400 may perform processes related to radiation treatment planning (various processes related to radiation treatment planning).

[0095] FIG. 15 is a diagram for explaining a radiation treatment plan, and illustrates the diagram of IMRT shown in FIG. 2. On the left side of the figure, an example of an image of a radiation treatment plan generated by the radiation treatment planning device 400 is shown. Although the illustration of the pneumonia occurrence hazard map is omitted, the area surrounded by the white line on the right side in the lung region is the "dangerous" area in the pneumonia occurrence hazard map, but there are areas where the dose is high. That is, there are areas where the dose of dangerous lung tissue is high.

[0096] Therefore, as one method, it is conceivable that the user performs an operation to adjust the irradiation range. Specifically, the radiation treatment planning device 400 displays, for example, a pneumonia occurrence hazard map and an image of a radiation treatment plan on the display unit. These images may be displayed on the same screen or on different screens. Also, the pneumonia occurrence hazard map may be displayed on another display device 600. Then, the user can operate the operation unit while looking at the pneumonia occurrence hazard map to adjust the irradiation range so that the dose of dangerous lung tissue becomes low (center of the figure → right side). That is, based on the image of the radiation treatment plan generated by the radiation treatment planning device 400, the user may be able to adjust the irradiation range.

[0097] Also, as another method, the radiation treatment planning device 400 may set the irradiation range based on the inferred pneumonia occurrence risk value and generate an image of the radiation treatment plan. Specifically, the processing unit (control unit) of the radiation treatment planning device 400 may set the irradiation range so that, for example, the dose at a position where the inferred pneumonia occurrence risk value exceeds a threshold or is equal to or higher than the threshold is relatively lower than the dose at other positions. The threshold may be set, for example, to a value used for classifying as "dangerous" in the pneumonia occurrence hazard map. Also, based on the image of the radiotherapy plan generated in this way, the user may be able to further adjust the irradiation range.

[0098] In conventional radiotherapy, a large amount of lung tissue that is highly likely to induce pneumonia also received a large dose of radiation. In contrast, with the method of this embodiment, by basing on the inferred pneumonia occurrence risk value and the pneumonia occurrence hazard map, the exposure of lung tissue with a high risk of inducing pneumonia can be reduced.

[0099] In addition to the above, the processing unit of the radiotherapy planning device 400 may perform processing to change, for example, the inferred pneumonia occurrence risk value. In this case, the processing unit may perform the following processing, for example. · Determine a specific site from volume data or the like · Change the inferred pneumonia occurrence risk value of the site excluding the specific site to a value that does not exceed the threshold (or a value that does not become equal to or higher than the threshold)

[0100] The specific site may be, for example, a site in the lung region that includes the site to be treated by radiotherapy (hereinafter referred to as the "target site") and is set as a site in a predetermined range in the vicinity thereof. The user may be able to set the range of the predetermined vicinity range.

[0101] In this case, · Site in the predetermined vicinity range: At positions where the inferred pneumonia occurrence risk value does not exceed the threshold, the dose becomes relatively high, and at positions where the threshold is exceeded, the dose becomes relatively low · Other sites (including sites of organs different from the lung): Since the threshold is not exceeded, the dose becomes relatively high It can be made like this. Basically, it is necessary to irradiate the tumor with radiation, but there may be a position with a high risk in the vicinity. In that case, the dose at that position can be made relatively low.

[0102] However, in this case, since at least a part of an organ different from the lungs is included in the irradiation range, specific side effects may occur in that organ. Generally, the side effects of radiotherapy tend to occur at the site where the radiation is applied. They may occur during or immediately after the treatment, or may occur after a certain period of time has elapsed. That is, when irradiating with radiation, it may affect normal organs. For this reason, for example, there is an idea of preventing radiation from being applied to normal organs (especially important organs) different from the organ in which the tumor exists.

[0103] On the other hand, if the position has a low risk of side effects occurring, even if it is a different organ, it is also possible to have the idea of applying radiation to it. In this case, for example, based on the method of this embodiment, for some or all of the different organs, a risk value similar to the pneumonia occurrence risk value (an example of the first value regarding the side effects of radiotherapy) is inferred, and not only the pneumonia occurrence risk value but also the risk values inferred for different organs are used to perform similar processing. The risk values for different organs may be inferred, for example, by quantifying the functional information based on the morphological information obtained for that organ and performing the same processing as above. When the functional information can be directly obtained by nuclear medicine examinations or the like, that may be used. Also, similar to the pneumonia occurrence hazard map, for different organs, a side effect occurrence hazard map indicating the risk of side effects occurring in that organ may be generated and output based on the inferred risk values.

[0104] As ideas, broadly speaking, · including at least a part of different organs in the irradiation range · not including different organs in the irradiation range these two can be considered.

[0105] Similarly to the above, the processing unit may change the inferred risk value based on, for example, the inferred pneumonia occurrence risk value, the risk values inferred for different organs, and the corresponding positions. In this case, the processing unit may perform, for example, the following processing. · Determine a specific site from volume data or the like · Change the inferred risk values of the sites excluding the specific site to values that do not exceed the threshold (or values that do not become equal to or higher than the threshold).

[0106] The specific site may be, for example, a site within a predetermined range in the vicinity and a site of a specific organ different from the lung. In this case, · Site within a predetermined range in the vicinity and site of a specific organ different from the lung: For positions where the inferred pneumonia occurrence risk value does not exceed the threshold, the dose is relatively high, and for positions where the threshold is exceeded, the dose is relatively low. · Other sites: The dose becomes relatively high because the threshold is not exceeded. It can be made like this. Regarding a specific organ, since the dose can be made relatively low at positions where the risk value exceeds the threshold, the exposure of positions with a high risk of side effects can be reduced. For example, if the specific organ is an important organ, the exposure of positions with a high risk of side effects among the important organs can be reduced.

[0107] Also, if it is based on the idea of making different organs not included in the irradiation range, the processing unit may, for example, change the inferred risk value for a specific organ to a value exceeding the threshold. For example, if the specific organ is an important organ, the exposure of the important organ can be minimized. Also, the irradiation range may be set so as to avoid a specific organ without inferring the risk value or regardless of the inferred risk value. These may be regarded as methods based on the idea of avoiding irradiation of radiation to specific organs.

[0108] Further, for example, with respect to an organ set as an organ in which side effects are less likely to occur or, even if side effects occur, the risk is considered low among different organs, the processing unit may be configured such that at least a part of the organ is included in the irradiation range.

[0109] Also, it is possible to adopt the idea of avoiding irradiation of radiation to a specific site within an organ including the target site. In this case, for example, within the organ including the target site, the processing unit may change the inferred risk value to a value that does not exceed a threshold value, excluding sites within a predetermined vicinity range. In this case, within the organ including the target site, · Sites within a predetermined vicinity range: At positions where the inferred risk value does not exceed the threshold value, the dose becomes relatively high, and at positions where the threshold value is exceeded, the dose becomes relatively low. · Other sites: Since the threshold value is not exceeded, the dose becomes relatively high. This can be achieved.

[0110] <System configuration, etc.> The method of this embodiment may be implemented by two or more devices. For example, a system may be configured that includes a first information processing device that generates a pneumonia occurrence prediction model and a second information processing device that infers a pneumonia occurrence risk value using the pneumonia occurrence prediction model generated by the first information processing device. Also, for example, a system may be configured that includes a second information processing device that infers a pneumonia occurrence risk value and a third information processing device that generates a pneumonia occurrence hazard map based on the pneumonia occurrence risk value inferred by the second information processing device. That is, the method of this embodiment may be implemented by a system including two or more devices.

[0111] Further, the information processing device may, for example, infer a pneumonia occurrence risk value based on a pneumonia occurrence prediction model generated in advance and stored in a storage medium or the like. Further, for example, the information processing apparatus may generate a pneumonia occurrence prediction model according to a pneumonia occurrence prediction model generation program stored in advance in a storage medium or the like. The same may apply to the generation and display of a pneumonia occurrence hazard map, settings related to the plan of radiation therapy, and the like.

[0112] <Mathematical information> In the present embodiment, the information used as the basis for performing the processing may be any mathematical information (mathematical information) related to the organs of the subject, and it does not matter what information of the organ it represents. This mathematical information may include, for example, morphological information and functional information. Further, the mathematical information may be, for example, information expressed in the real space related to the organs of the subject (3D information, 4D information when considering the time axis). Note that it may also be information expressed on a real plane (2D information).

[0113] Further, the mathematical information may be obtained by any method. A dedicated device may be used, but the mathematical information may also be obtained without using a dedicated device. For example, 2D information may be pseudo-3D converted to obtain morphological information such as 3DCT information. In addition to the method described in the present embodiment, for example, a deep learning model such as a CNN (Convolutional Neural Network) may be used to obtain functional information. For example, "A deep learning method for translating 3DCT to SPECT ventilation imaging: First comparison with 81mKr-gas SPECT ventilation imaging" (https: / / doi.org / 10.1002 / mp.15697) discloses a method for generating functional information from 3DCT information using a deep learning model, and for example, this method may be applied. In addition, techniques using a GAN (Generative Adversarial Network), a diffusion model, or the like may be applied.

[0114] Moreover, mathematical information may be obtained using different types of information. For example, mathematical information may be obtained using function information acquired by some method and morphological information acquired by some method. For example, a deep learning model that outputs mathematical information with function information and morphological information as inputs is generated. In this case, if the model is learned to prioritize function information over morphological information, for example, although it does not appear to be information representing a function on the surface, it can be made to output information as if it represents a function. That is, information that does not appear to be function information on the surface but can be regarded as (hypothetically) function information can be obtained. For example, the information obtained in this way may be used as function information to perform the above processing.

[0115] <Actions and effects of the embodiment> In this embodiment, the information processing apparatus 10 includes a pneumonia occurrence prediction model (an example of a model) for calculating a pneumonia occurrence risk value (an example of a first value related to the side effects of radiation therapy) generated using learning data related to a plurality of subjects (an example of a subject), and a predetermined value (an example of a second value) based on information such as a lung function value corresponding to the position information of the lungs (an example of an organ) of the first subject (an example of a first subject), which is mathematical information related to the organ and an example of mathematical information corresponding to the position information. The learned model processing unit 15 (an example of a calculation unit) calculates a pneumonia occurrence risk value (an example of a first value) corresponding to the position information based on the above. Thereby, based on a model for calculating a first value related to side effects when performing radiation therapy, generated using learning data related to a plurality of subjects, and a second value based on mathematical information related to the organ of the first subject and corresponding to the position information, the first value corresponding to the position information can be calculated. As a result, using the calculated first value, it becomes possible to appropriately support diagnosis, treatment, etc., taking into account side effects of radiation therapy, for example.

[0116] Note that the generation of the model may be realized, for example, by learning the model by machine learning or the like using learning data. In addition, the learning of the model may include, for example, learning using a single training data or learning (relearning) using another training data based on the verification result of the model using verification data.

[0117] The functional information, which is an example of mathematical information, may be obtained, for example, based on morphological information regarding the organs of the first subject (an example of the first subject) and morphological information corresponding to the position information (information on anatomical structure). Thereby, based on the morphological information regarding the organs of the first subject and morphological information corresponding to the position information, the functional information can be obtained simply and appropriately.

[0118] Further, the above-mentioned predetermined value (an example of the second value) may be the exposure dose based on the above-mentioned mathematical information and the exposure information regarding radiation therapy. Thereby, based on the exposure dose based on the above-mentioned mathematical information and the exposure information regarding radiation therapy, and the model for calculating the first value regarding the side effects in the case of performing radiation therapy generated using the learning data regarding a plurality of subjects, the first value corresponding to the position information can be appropriately calculated.

[0119] In addition, the exposure dose may include a plurality of exposure doses divided based on the above-mentioned mathematical information. Thereby, using the learning data including a plurality of exposure doses divided based on the above-mentioned mathematical information, a model for calculating the first value regarding the side effects of radiation therapy can be appropriately generated.

[0120] In addition, the learning data may be data of a set of the pneumonia occurrence risk value (an example of the first value) and a plurality of exposure doses. Thereby, using the data of the set of the first value and a plurality of exposure doses, a model for calculating the first value regarding the side effects of radiation therapy can be generated.

[0121] Further, the above model may be a pneumonia occurrence prediction model (an example of the model) represented by the weighted sum value of the above multiple exposure doses. Thereby, using the model represented by the weighted sum value of the above multiple exposure doses, the first value corresponding to the position information can be appropriately calculated.

[0122] Further, the information processing apparatus 10 may include a pneumonia occurrence hazard map generation unit 17 (an example of an image generation unit) that generates a pneumonia occurrence hazard map (an example of an image related to side effects) based on the position information and the pneumonia occurrence risk value (an example of the first value) calculated using the pneumonia occurrence prediction model (an example of the model). Thereby, an image related to side effects can be generated based on the position information and the first value calculated using the model. As a result, it becomes possible to visualize the image related to side effects and appropriately support diagnosis, treatment, etc.

[0123] Further, the information processing apparatus 10 may include a display control unit 19 that performs control to display the pneumonia occurrence hazard map generated by the pneumonia occurrence hazard map generation unit 17 on a display device. Thereby, the image related to side effects can be displayed on the display device so that the user can view it.

[0124] Alternatively, a display device 600 including a display unit that displays the pneumonia occurrence hazard map acquired from the information processing apparatus 10 may be configured. Thereby, the user can view the image related to side effects.

[0125] Alternatively, a radiation treatment planning apparatus 400 including the information processing apparatus 10 and including a display unit that displays a pneumonia occurrence hazard map and an image of the lung (an image of the radiation treatment plan) in which the irradiation range of radiation in radiation treatment is associated may be configured. Thereby, the user can compare the image related to side effects with the image of the radiation treatment plan.

[0126] In this case, the irradiation range may be made changeable based on a user operation. This enables the user to adjust the irradiation range of the radiation while viewing an image related to side effects.

[0127] Further, a radiation treatment planning apparatus 400 may be configured to include a processing unit that performs processing related to a radiation treatment plan based on the position information corresponding to the mathematical information acquired from the information processing apparatus 10 and the pneumonia occurrence risk value (an example of a first value) calculated using a pneumonia occurrence prediction model (an example of a model). This makes it possible to configure a radiation treatment planning apparatus that performs processing related to a radiation treatment plan based on the position information corresponding to the mathematical information acquired from the information processing apparatus and the first value calculated using the model.

[0128] Also, the pneumonia occurrence risk value is a value indicating that the higher the value, the higher the risk of side effects. The processing unit may set the irradiation range of the radiation such that the dose at a position where the inferred pneumonia occurrence risk value exceeds a threshold or is equal to or higher than the threshold is relatively lower than the dose at other positions. This makes it possible to set the irradiation range so as to reduce the exposure at positions where the risk of side effects occurring is high.

[0129] Also, the pneumonia occurrence risk value is a value indicating that the higher the value, the higher the risk of side effects. The processing unit may change the inferred pneumonia occurrence risk value so as not to exceed a threshold or not to become equal to or higher than the threshold, except for a specific site. This makes it possible to, for example, make the dose at positions excluding a specific site relatively higher.

[0130] In this case, the specific site may include at least a site within a predetermined range in the vicinity including the target site of the radiation treatment. This makes it possible to, for example, make the dose at positions excluding at least a site within a predetermined range in the vicinity including the target site of the radiation treatment relatively higher.

[0131] In addition, in this case, in addition to the site within the predetermined range in the vicinity described above, the specific site may also include a site of a specific organ different from the organ including the target site. Thereby, for example, the dose at a position excluding the site within the predetermined range in the vicinity including the target site of radiotherapy and the site of the specific organ can be made relatively high.

[0132] Also, the pneumonia occurrence risk value is a value indicating that the higher the value, the higher the risk of side effects. The processing unit may change the inferred pneumonia occurrence risk value so as not to exceed or not become equal to or higher than the threshold value within the organ including the target site of radiotherapy, excluding the site within the predetermined range in the vicinity including the target site. Thereby, for example, within the organ including the target site of radiotherapy, the dose at a position excluding the site within the predetermined range in the vicinity can be made relatively high.

[0133] <First Modification Example of the Embodiment> The method of the above embodiment may be similarly applied to medical acts such as examinations and diagnoses using radiation in addition to radiotherapy. That is, it may be applied to the whole of radiation medicine including treatment, examination, diagnosis, etc. In addition, the method of the above embodiment is not limited to the treatment of lung cancer, and can be similarly applied as long as it performs a treatment involving exposure, etc. In addition, the subject is not limited to humans and may be an animal (excluding humans). In this case, as an example, if it is a dog, the risk value is calculated using a model generated using learning data related to a plurality of dogs, and if it is a cat, the risk value is calculated using a model generated using learning data related to a plurality of cats. For example, the subject may be of the same type.

[0134] In addition, when performing radiation therapy for cancers of organs other than the lungs, etc., by the same method as described above, for that organ, a risk value similar to the pneumonia occurrence risk value may be inferred and output. Also, based on the inferred risk value, for that organ, a side effect occurrence hazard map may be generated and output. Further, for that organ, the settings regarding the radiation therapy plan described above may be made.

[0135] <Second Modified Example of the Embodiment> The pneumonia occurrence risk value in the above embodiment may be defined and calculated as a value in a numerical range such as "0 to 1" or "0 to 100". In this case, the pneumonia occurrence risk value may be the pneumonia occurrence risk degree representing the risk of pneumonia occurrence as a probability. Also, the pneumonia occurrence risk value may be defined and calculated such that the smaller the value, the higher the risk of pneumonia occurrence.

[0136] Also, instead of the pneumonia occurrence risk value, a pneumonia occurrence safety value representing the difficulty (safety) of pneumonia occurrence (for example, a value indicating that the higher the value, the higher the safety) may be defined and calculated. And based on the pneumonia occurrence safety value, a pneumonia safety map may be generated.

[0137] For example, the pneumonia occurrence risk value and the pneumonia occurrence safety value may be examples of the first value regarding the side effects of radiation therapy. Whichever value is used, for example, by introducing a threshold value (cutoff value), it is possible to determine occurrence / non-occurrence, risk / safety, etc.

[0138] Note that the same may apply when processing organs other than the lungs.

[0139] <Third Modified Example of the Embodiment> In the above embodiment, the fractionation interval was divided into five, but it is not limited to this. The number of divided sections can be set as appropriate. For example, numbers such as "10", "15", "20" can be set. This setting may be automatically performed by the information processing apparatus (automatic setting), or the information processing apparatus may set it based on the user's input (manual setting). However, if the number of divided sections is too small, information will be lost. Therefore, for example, a value of about "3" to "10" is appropriate. Since the inventor of the present application obtained the finding that good results can be obtained when the number of divided sections is "5", the number of divided sections was described as "5" in the above embodiment.

[0140] Note that the same may apply when processing organs other than the lungs.

Explanation of Signs

[0141] 1(1A, 1B) Information processing system 10(10A, 10B) Information processing apparatus 100 Medical image diagnostic apparatus 200 Medical image database 300 Radiation treatment apparatus 400 Radiation treatment planning apparatus 500 Radiation treatment image database 600 Display device

Claims

1. A model for calculating a first value related to side effects of radiation therapy, generated using learning data regarding a plurality of subjects, and a second value based on mathematical information regarding an organ of a first subject and corresponding to position information, based on which a calculation unit for calculating the first value corresponding to the position information is provided. An information processing apparatus.

2. The second value is a radiation dose based on the mathematical information and radiation exposure information regarding the radiation therapy. The information processing apparatus according to Claim 1.

3. The radiation dose is a plurality of radiation doses divided based on the mathematical information. The information processing apparatus according to Claim 2.

4. The learning data is data of a set of the first value and the plurality of radiation doses. The information processing apparatus according to Claim 3.

5. The model is a model represented by a weighted sum value of the plurality of radiation doses. The information processing apparatus according to Claim 4.

6. An image generation unit for generating an image regarding the side effect is provided based on the position information and the first value calculated by the calculation unit. The information processing apparatus according to Claim 1.

7. A display unit for displaying the image regarding the side effect acquired from the information processing apparatus according to Claim 6 is provided. A display apparatus.

8. The information processing apparatus according to Claim 6, a display unit for displaying the image regarding the side effect and an image of the organ of the first subject in which the irradiation range of radiation in the radiation therapy is associated, A radiation treatment planning apparatus comprising the same.

9. Based on the position information corresponding to the mathematical information and the first value calculated by the calculation unit, acquired from the information processing apparatus according to any one of Claims 1 to 6, a processing unit for performing processing regarding the radiation therapy plan is provided. A radiation treatment planning apparatus.

10. The first value is a value indicating that the higher the value, the higher the risk of the side effect. The processing unit sets an irradiation range of radiation such that the dose at a position where the first value calculated by the calculation unit exceeds a threshold or is equal to or higher than the threshold is relatively lower than the dose at other positions. The radiation treatment planning apparatus according to Claim 9.

11. The first value is a value indicating that the higher the value, the higher the risk of the side effect. The processing unit changes the first value calculated by the calculation unit so as not to exceed a threshold or not to be equal to or higher than the threshold, except for a specific site. The radiotherapy planning apparatus according to claim 9.

12. The specific site includes a site within a predetermined range in the vicinity including at least the target site of the radiotherapy. The radiotherapy planning apparatus according to claim 11.

13. The specific site includes a site of a specific organ different from the organ. The radiotherapy planning apparatus according to claim 12.

14. The first value is a value indicating that the higher the value, the higher the risk of the side effect. The processing unit changes, within the organ, the first value calculated by the calculation unit so as not to exceed or not to be equal to or higher than a threshold value, excluding a site within a predetermined range in the vicinity including the target site of the radiotherapy. The radiotherapy planning apparatus according to claim 9.

15. Calculating the first value corresponding to the position information based on a model for calculating a first value related to a side effect of radiotherapy generated using learning data related to a plurality of subjects, and a second value based on mathematical information related to an organ of a first subject and corresponding to the position information. Information processing method.

16. Causing a computer to calculate the first value corresponding to the position information based on a model for calculating a first value related to a side effect of radiotherapy generated using learning data related to a plurality of subjects, and a second value based on mathematical information related to an organ of a first subject and corresponding to the position information. Program.

Images