Program, information processing device, information processing method

Abstract

We propose a new method for evaluating areas of interest. [Solution] The program causes the computer to perform the following actions in a medical image: setting multiple regions in which the region of interest is included and which have different angles from each other; calculating evaluation values ​​for evaluating the region of interest based on the multiple regions; and performing an evaluation of the region of interest based on the evaluation values.

Description

[Technical Field]

[0001] This invention relates to programs, etc. [Background technology]

[0002] For example, lung cancer is a leading cause of cancer-related deaths worldwide, and diagnosing lung cancer is a medical challenge. In recent years, attempts have been made to diagnose lung cancer using Radiomics, a technique that extracts numerous reproducible quantitative features from medical images, applied to computed tomography (CT) images (see Non-Patent Documents 1 and 2). [Prior art documents] [Non-patent literature]

[0003] [Non-Patent Document 1] Avanzo M et al. “Radiomics and deep learning in lung cancer”, Strahlenther Onkol 2020 Oct;196(10):879-887. [Non-Patent Document 2] Thawani R et al. “Radiomics and Radiogenomics in Lung Cancer” Lung Cancer 2017 DOI:10.1016 / j.lungcan.2017.10.015 [Non-Patent Document 3] Takuma Usuzaki, Kengo Takahashi, Kazuma Umemiya “A new radiomics feature:image frequency analysis”,arXiv:2111.05855v1[q-bio.QM]10 Nov 2021. [Overview of the project] [Problems that the invention aims to solve]

[0004] Radiomics-based diagnosis is well-suited to machine learning (including deep learning) and has been reported to have high accuracy. However, diagnosis using radiomics requires the preparation of mask images from which the contours of the affected area have been extracted. In the research stage, physicians often have to manually examine each CT image and outline the contours, which is a significant burden when considering its application in clinical practice.

[0005] From this perspective, the inventor of the present invention has proposed a method that does not require a mask image, as disclosed in Non-Patent Document 3, but a more advanced method was needed.

[0006] This invention was made in view of the above-mentioned problems, and one of its objectives is to propose a new method for evaluating areas of interest. [Means for solving the problem]

[0007] According to a first aspect of the present invention, the program causes a computer to perform the following actions in a medical image: to set up a plurality of regions in which the region of interest is included and which have different angles from each other; to calculate evaluation values ​​for evaluating the region of interest based on the plurality of regions; and to perform an evaluation of the region of interest based on the evaluation values. According to a second aspect of the present invention, the information processing device includes a processing unit that, in a medical image, sets a plurality of regions that include a region of interest and have different angles from each other, calculates an evaluation value for evaluating the region of interest based on the plurality of regions, and performs an evaluation of the region of interest based on the evaluation value. According to a third aspect of the present invention, the information processing method performed by a computer includes: setting a plurality of regions in a medical image that include a region of interest and have different angles from each other; calculating evaluation values ​​for evaluating the region of interest based on the plurality of regions; and performing an evaluation of the region of interest based on the evaluation values. [Effects of the Invention]

[0008] According to the present invention, it becomes possible to perform evaluations of a region of interest.

Brief Description of the Drawings

[0009] [Figure 1] A block diagram showing an example of the functional configuration of an information processing apparatus. [Figure 2] An explanatory diagram of the evaluation of the region of interest. [Figure 3] An explanatory diagram of the evaluation of the region of interest. [Figure 4] A flowchart showing an example of the flow of processing executed by the information processing apparatus. [Figure 5] An example of a radar chart showing the correlation array frequencies by angle. [Figure 6] A diagram showing an example of the functional configuration of an evaluation apparatus. [Figure 7] A diagram showing an example of a screen displayed on the evaluation apparatus.

Modes for Carrying Out the Invention

[0010] Hereinafter, an example of a mode 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. Also, the components described in this embodiment are merely examples, and are not intended to limit the scope of the present invention thereto.

[0011] [Embodiment] Hereinafter, an example of an embodiment for realizing the information processing technology of the present invention will be described.

[0012] FIG. 1 is a block diagram showing an example of the functional configuration of an information processing apparatus 1 according to an aspect of the present embodiment. The information processing apparatus is provided with, for example, an area-of-interest setting unit 110, a rectangular area setting unit 120, a correlation value calculation unit 130, a correlation array frequency calculation unit 140, and an area-of-interest evaluation unit 150. These can be functional units (functional blocks) possessed by, for example, a processing unit (processing device) or a control unit (control device) of the information processing apparatus 1 that is not shown in the figure, and are configured to include processing circuits such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), and an FPGA (Field Programmable Gate Array).

[0013] The region of interest setting unit 110 sets a region of interest for the input medical image. The medical image may be acquired, for example, from a medical image diagnostic apparatus that is not shown in the figure. The medical image diagnostic apparatus may include apparatuses such as an X-ray CT (Computed Tomography) apparatus and an MRI (Magnetic Resonance Imaging) apparatus that can acquire and evaluate morphological information (information on anatomical structures) regarding organs. In the present embodiment, the region of interest may be, for example, a region detected from a medical image and including at least a nodule portion. Note that a region of peripheral tissue of the nodule portion may also be included in the region of interest.

[0014] The rectangular region setting unit 120 sets a region including the region of interest set by the region of interest setting unit 110, and sets, for example, a plurality of rectangular regions (hereinafter simply referred to as "a plurality of rectangular regions") having different angles from each other. The rectangle may include a square, and the rectangular region may be, for example, a square region of a predetermined size (for example, 256 pixels × 256 pixels). Also, the plurality of rectangular regions may be, for example, 360 / n rectangular regions obtained by rotating one rectangle by n° each. In this case, as one method, the angle at which the subject (including the patient) lies during imaging such as CT may be set to 0°, and the angle may be defined in the coronal plane with the counterclockwise direction being positive.

[0015] The correlation value calculation unit 130 calculates the correlation value between adjacent rows or columns for each of the multiple rectangular regions set by the rectangular region setting unit 120. For example, it divides two opposing sides into rows at equal intervals and other two opposing sides into columns at equal intervals.

[0016] The correlation array frequency calculation unit 140 performs a predetermined frequency analysis (spectral analysis), such as Fourier analysis, on the array of correlation values ​​corresponding to each of the multiple rectangular regions calculated by the correlation value calculation unit 130, and calculates the frequency corresponding to each of the arrays of correlation values ​​for each of the multiple rectangular regions (for convenience, referred to as the "correlation array frequency"). For example, the amplitude for each frequency is obtained, and the frequency at which the amplitude is maximum (where a peak appears) is taken as the correlation array frequency. This yields one correlation array frequency corresponding to one rectangular region.

[0017] The region of interest evaluation unit 150 evaluates the region of interest based on the correlation array frequency calculated by the correlation array frequency calculation unit 140 and outputs the evaluation result. This may include, for example, determining whether the nodules included in the region of interest are benign or malignant (benign / malignant) and determining the probability of them being benign or malignant.

[0018] Furthermore, "output" of information and data, including evaluation results (judgment results), may include, for example, display (display output), output to other functional units within the device (internal output), output to devices other than the device itself (external devices) (external output), and transmission (external transmission).

[0019] Figures 2 and 3 are explanatory diagrams illustrating the evaluation of the region of interest in this embodiment. As shown in Figure 2, the region of interest setting unit 110 sets a region of interest as shown in the figure through image processing of the input medical image. This figure shows the state in which the location of a pulmonary nodule (pulmonary nodule area) has been detected from a CT image of the lung. In this case, the region of interest may be set to include at least the entire pulmonary nodule area (the region of the pulmonary nodule area may also be the region of interest). Then, the rectangular region setting unit 120 sets (360 / n) rectangular regions that include the region of interest, for example, obtained by rotating one rectangle by n degrees. Specifically, for example, (360 / 5) rectangular regions that include the region of interest may be set, for example, obtained by rotating a rectangle by 5 degrees (n=5) with respect to the center point of the region of interest. Then, for each of these multiple rectangular regions, the correlation array frequency is calculated by the correlation array frequency calculation unit 140.

[0020] In this case, for each of the multiple rectangular regions, the correlation value between adjacent columns may be calculated using, for example, the pixel values ​​of two adjacent columns. Specifically, for one rectangular region, the correlation value between the x-th column and the y-th column may be calculated, for example, according to equation (1) below. Correlation value r i,i+1 =(Covariance of pixel values ​​in column i and column i+1) / [(Standard deviation of pixel values ​​in column i) × (Standard deviation of pixel values ​​in column i+1)] ···(1) In this case, for example, the correlation value r shown in the graph in the center of Figure 3 is... i,i+1 An array is obtained. The horizontal axis represents the column number "0 to 255", and the vertical axis represents the correlation value r i,i+1 These are shown respectively.

[0021] Correlation value r for all adjacent columns i,i+1 Calculate the correlation value r i,i+1 Obtain an array of these. This process is performed for each of the multiple rectangular regions (for each of the multiple angles), and the correlation value r is calculated for each of the multiple rectangular regions. i,i+1 Obtain an array of these.

[0022] Next, for each of the multiple rectangular regions, the correlation value r i,i+1 Frequency analysis (e.g., Fourier analysis) is performed on the sequence, and the correlation value r is obtained. i,i+1 Calculate the correlation array frequency corresponding to the array. As a result, as shown in Figure 2, we can obtain, for example, the correlation array frequency corresponding to "0°" (0°), the correlation array frequency corresponding to "5°" (5°), the correlation array frequency corresponding to "10°" (10°), ..., the correlation array frequency corresponding to 355° (355°).

[0023] The region of interest evaluation unit 150 uses the correlation array frequencies calculated for multiple rectangular regions with different angles as described above to evaluate the region of interest. Specifically, (i) for example, the average value of the correlation array frequencies calculated for multiple rectangular regions is calculated, and the calculated average value is compared with a pre-set cutoff value for the correlation array frequency to determine whether the nodule included in the region of interest is benign or malignant. Here, the cutoff value can be, for example, a correlation array frequency value appropriate for distinguishing between benign and malignant, as shown in Figure 5.

[0024] Furthermore, (ii) as a cutoff value, based on the database of correlation array frequencies (benign, malignant), the probability of each correlation array frequency included in the database being benign and the probability of it being malignant are obtained. Then, for each correlation array frequency, the true positive rate, false positive rate, false negative rate, and true negative rate are calculated, respectively, when that correlation array frequency is used as the cutoff value. Then, (sensitivity + specificity - 1) is calculated, and the correlation array frequency that yields the maximum value can be adopted as the cutoff value (Youden Index).

[0025] Here, the region of interest evaluation unit 150 may be composed of, for example, a machine learning model (artificial intelligence model) including deep learning. For example, when it receives correlation array frequencies calculated for multiple rectangular regions with different angles as input, it may calculate a pre-learned probability (e.g., "0" to "1") that serves as a guideline for the benign or malignant nature of the nodule, according to a pre-learned artificial intelligence model. Alternatively, it may determine whether the nodule is malignant or benign (e.g., "0" or "1") based on this probability and output it. For example, when applying a neural network model, its output layer may have multiple output nodes to enable multi-class classification, or it may have a single output node to enable binary classification (e.g., malignant / benign).

[0026] <Processing> Figure 4 is a flowchart showing an example of the information processing procedure in this embodiment. The process shown in the flowchart of Figure 4 is realized, for example, by the processing unit of the information processing device 1 reading the program code stored in a memory unit (not shown) into a RAM (Random Access Memory) (not shown) and executing it.

[0027] In a flowchart, each symbol S represents a step. The flowchart described below is merely an example of the information processing procedure in this embodiment, and other steps may be added or some steps may be deleted. Furthermore, some of the steps in the flowchart may be rearranged before execution.

[0028] First, the region of interest setting unit 110 performs region of interest setting processing (S110). Specifically, it performs image processing such as segmentation, edge detection, and contour detection on the acquired medical image to set the region of interest, including the nodule.

[0029] Next, the rectangular region setting unit 120 performs rectangular region setting processing (S120). Specifically, it sets multiple rectangular regions with different angles, including the region of interest set in S110.

[0030] Subsequently, the correlation value calculation unit 130 performs correlation value calculation processing (S130). Specifically, for each of the multiple rectangular regions set in S120, the unit divides the space between two opposing sides at equal intervals to form multiple rows, and the space between other two opposing sides at equal intervals to form multiple columns, and calculates the correlation value between adjacent rows or columns.

[0031] Next, the correlation array frequency calculation unit 140 performs correlation array frequency calculation processing (S140). Specifically, it applies a predetermined frequency analysis (e.g., Fourier analysis) to the array of correlation values ​​corresponding to each of the multiple rectangular regions calculated in S130 to calculate the correlation array frequency corresponding to each of the multiple rectangular regions with different angles.

[0032] Subsequently, the region of interest evaluation unit 150 performs region of interest evaluation processing (S150). Specifically, based on the correlation array frequencies corresponding to each of the multiple rectangular regions with different angles calculated in S140, it determines whether the nodule included in the region of interest is malignant or benign, and calculates the probability thereof, for example, according to the method described above.

[0033] Then, the region of interest evaluation unit 150 outputs the evaluation result of S150 (S160).

[0034] Next, the control unit of the information processing device 1 determines, for example, whether to terminate the process based on user input (user operation, etc.) (S190). If it determines to continue the process (S190: NO), it returns to, for example, S110. If it determines to terminate the process (S190: YES), the control unit of the information processing device 1 terminates the process.

[0035] <Experimental Results> Figure 5 shows an example of experimental results from an experiment conducted by the inventor of the present invention to investigate how the correlation array frequencies for each angle differ depending on whether the nodule is benign or malignant. This experimental result is presented in radar chart format, showing whether or not there was a statistically significant difference in the average correlation array frequency calculated for a subset of subjects from 244,527 CT images of lung nodules from 1,010 subjects included in the publicly available dataset, The Lung Image Database Consortium and Image Database Resource Initiative (LIDC-IDRI).

[0036] In this example, rectangles (squares) were extracted from CT images of the lungs of LIDC-IDRI patients, with a margin of approximately 20% relative to the contour of the lung nodules. The correlation array frequencies were calculated in the range of 0° to 355° while rotating the rectangles in 5° increments, and the Mann-Whitney U test was used to compare the results for each angle. In this radar chart, the numbers on the outer edge represent the angles of the rectangular area (0° to 355° in 5° increments). The solid lines in the radar chart connect the average correlation array frequencies of the corresponding angles for subjects with benign nodules, while the dotted lines connect the average correlation array frequencies of the corresponding angles for subjects with malignant nodules. In this experiment, a statistically significant difference was determined for each angle based on the p-value, and angles showing a significant difference are marked with a black dot on the outer edge.

[0037] These results show a significant difference in the calculated correlation array frequencies between subjects with benign nodules and those with malignant nodules. Furthermore, statistically significant differences are observed in most angles. This indicates that the benign / malignant nature of nodules can be distinguished by the value of the correlation array frequency.

[0038] Note that no significant difference was found for some angles, such as "0°". This is presumably because, although there was a difference in the mean value of the correlation array frequency between benign and malignant lesions, the variance (dispersion) was large, preventing the p-value from decreasing sufficiently, resulting in a determination of no significant difference.

[0039] <Effects of the Embodiment> In this embodiment, in a medical image, a plurality of regions (for example, rectangular regions with different angles at 5° intervals) including a region of interest (for example, a nodule) and having different angles from each other are set, and based on these plurality of regions, an evaluation value for evaluating the region of interest is calculated (for example, for an array of correlation values corresponding to each of the plurality of rectangular regions, a predetermined frequency analysis (for example, Fourier analysis) is applied to calculate a correlation array frequency corresponding to each of the plurality of rectangular regions with different angles), and based on the evaluation value, an evaluation of the region of interest is performed (for example, the benignancy / malignancy of the nodule is distinguished by the value of the correlation array frequency). In the calculation procedure of this example, the pulmonary nodule is surrounded by a rectangle (either manual operation on the image or automatic setting based on image processing is acceptable), and the correlation values (r i,i+1 ; 1≤i≤255) between adjacent rows are calculated, Fourier analysis is performed on the correlation values, and the correlation array frequency is obtained by calculating the frequency included in the transition of the correlation values.

[0040] In this way, since the correlation array frequency is a feature amount that can be obtained without requiring a mask image, which was necessary for the calculation of conventional radiomics, it is possible to apply a simple evaluation method to clinical practice. Note that since the correlation array frequency is a feature amount calculated using information on both the nodule part (for example, the tumor part) and the surrounding tissue, there is a possibility of quantifying the interaction between the tumor and the surrounding tissue, which was difficult to evaluate with conventional radiomics.

[0041] <Example> Next, an example of an evaluation device that applies the above information processing device 1 or includes the above information processing device 1 will be described. This device may be configured as a terminal, an electronic device (electronic apparatus), or the like.

[0042] FIG. 6 is a diagram showing an example of the functional configuration of the evaluation device 10. The evaluation device 10 includes, for example, a processing unit 100, a storage unit 200, a data acquisition unit 5 that acquires data such as medical images, an operation unit 310, a display unit 320, and a communication unit 330.

[0043] The processing unit 100 is a processing unit (control device) that comprehensively controls each part of the evaluation device 10 and performs various processing operations according to various programs such as system programs stored in the memory unit 200, and is configured to have processing circuits such as a CPU, GPU, DSP, ASIC, FPGA, etc. The processing unit 100 may have, for example, the functional units shown in Figure 1 as its main functional units.

[0044] The storage unit 200 is a storage device that includes memory circuits such as ROM (Read Only Memory), RAM, and flash memory, as well as hard disk drives and magneto-optical disk drives.

[0045] The memory unit 200 stores, for example, an evaluation program 210 and various data 220.

[0046] The evaluation program 210 is, for example, a program that is read by the processing unit 100 and executed as an evaluation process.

[0047] Various data 220 stores various types of data related to the evaluation process. For example, medical image data, region of interest data, rectangular region data, correlated array frequency data, evaluation result data for the region of interest, etc., may be stored.

[0048] The operation unit 310 is configured to have input devices for the user to perform various operations on the evaluation device 10, such as operation buttons and operation switches. The operation unit 310 may also have a touch panel (not shown) that is integrally configured with the display unit 320, and this touch panel may function as an input interface between the user and the evaluation device 10. The operation unit 310 may output operation signals to the processing unit 100 according to user operations, for example. An input device that accepts sound (including voice) input as user input may also be configured.

[0049] The display unit 320 is a display device configured with, for example, an LCD (Liquid Crystal Display) or an OLED (Organic Electro-luminescence Display), and performs various displays based on the display signals output from the processing unit 100.

[0050] Alternatively, the evaluation device 10 may transmit various types of information to a display device (display unit) provided separately from the evaluation device 10, and the display device may display the information received from the evaluation device 10. Alternatively, the information processing device 1 may transmit various types of information to the display device, and the display device may display the information received from the information processing device 1.

[0051] The communication unit 330 is a communication device for sending and receiving information used within the device to and from an external information processing device. Various communication methods can be applied to the communication unit 330, including wired connection via a cable compliant with a predetermined communication standard such as Ethernet or USB (Universal Serial Bus), wireless connection using wireless communication technology compliant with a predetermined communication standard such as Wi-Fi (registered trademark) or 5G (fifth-generation mobile communication system), and connection using short-range wireless communication such as Bluetooth (registered trademark). The communication unit 330 may transmit or receive various types of information from an external device in accordance with the control of the processing unit 100.

[0052] The processing unit 100 of the evaluation device 10 performs evaluation processing according to, for example, the evaluation program 210 stored in the storage unit 200. The evaluation processing may be performed according to, for example, the flowchart in Figure 4.

[0053] The evaluation device 10 may be an example of a user's terminal, and an application program for evaluation may be stored on this terminal. Then, this application program may be executed on the terminal, and the evaluation process may be performed by a server that can communicate with the terminal, and the results may be displayed on the terminal's display unit. The application can be a native application, a web application, or a hybrid application.

[0054] <Display screen> Figure 7 shows an example of a screen displayed on the display unit 320 of the evaluation device 10 in this embodiment. Here, an example of a screen displayed in the evaluation application is shown. A region of interest is set, which in this example includes the nodule (a region that can be visually identified as a nodule), and a region of interest including the pulmonary nodule region is set (S110). A rectangular region setting process (S120), a correlation value calculation process (S130), a frequency calculation process (S140), and a region of interest evaluation process (S150) are performed on this region of interest, and the evaluation results are output (S160). This screen is from the evaluation application, and for a particular patient, it displays the evaluation results from S160 (in this example, a radar chart and the probability that the nodule being evaluated is malignant) along with a medical image including the region of interest. In this example, information indicating that the probability that the nodule being evaluated is malignant is determined to be "12%" is displayed.

[0055] <Effects of the Example> According to the evaluation device 10 of this embodiment, the same functions and effects as those of the previously described embodiment can be obtained. Furthermore, the evaluation device 10 may be equipped with a display unit 320 that displays information related to the determination of knots, thereby allowing the user to confirm the information related to the determination of knots.

[0056] <Variation> In the above embodiment, the region of interest was evaluated based on the correlation array frequencies corresponding to all angles, but the region of interest may also be evaluated based on the correlation array frequencies corresponding to some angles. Since it is only some angles, this may include cases where the region of interest is evaluated based on the correlation array frequency corresponding to one angle, or it may be evaluated based on the correlation array frequencies corresponding to two or more angles, although not all of them. Furthermore, in either all angles or some angles, if there are angles where the correlation array frequency deviates by a certain amount or more from the correlation array frequency corresponding to other angles, the correlation array frequency corresponding to that deviating angle may be excluded, and the evaluation of the region of interest may be performed.

[0057] Furthermore, the correlation array frequencies corresponding to each angle may be displayed on the display device so that the user can confirm them.

[0058] Furthermore, if there is a certain level of variation in the correlation array frequencies corresponding to each angle, for example, (i) Redefine the region of interest (for example, change the size of the region of interest) (ii) Reconfigure multiple rectangular regions with different angles (for example, change the size of the rectangular regions) You may perform one or both of the following to recalculate the correlation array frequencies, and then make an assessment of the region of interest based on the recalculated correlation array frequencies.

[0059] Furthermore, for example, angles with significant differences can be identified from the radar chart shown in Figure 5. Therefore, calculations may be performed only for angles with significant differences to evaluate the region of interest.

[0060] Furthermore, in the above embodiment, an example was described in which a rectangular region is set as a region containing the region of interest and having different angles from each other. However, regions other than rectangles (for example, a region containing the region of interest and having an outline that expands the region of interest) may also be set as a region with different angles from each other, and these regions may be rotated by n degrees to form 360 / n regions. In this case, the largest region within the set region that can be divided into multiple rows by dividing the space between two opposing sides at equal intervals and multiple columns by dividing the space between the other two opposing sides at equal intervals may be automatically set, and the correlation value may be calculated for the automatically set region.

[0061] Furthermore, although the above embodiment exemplifies the case where the lungs are the target organ, it is not limited to this, and other organs may also be targeted. Furthermore, it is not limited to cases involving humans; it may also involve animals. [Explanation of symbols]

[0062] 1. Information Processing Device 10 Evaluation device

Claims

1. On the computer, In medical imaging, setting multiple regions that include the region of interest but have different angles from each other, Based on the aforementioned multiple domains, an evaluation value is calculated for evaluating the domain of interest, Based on the aforementioned evaluation values, an evaluation of the area of ​​interest is performed, A program to execute.

2. Calculating the aforementioned evaluation value includes, for each of the multiple regions, calculating the correlation value between adjacent rows or columns within the region and obtaining an array of the aforementioned correlation values. The program according to claim 1.

3. Calculating the aforementioned evaluation value includes performing frequency analysis on the sequence for each of the plurality of regions to calculate the frequency corresponding to the sequence. The program according to claim 2.

4. The evaluation value is two or more of the frequencies calculated for each of the multiple regions. The program according to claim 3.

5. The region of interest includes a nodule, The aforementioned evaluation includes evaluating whether the nodule is benign or malignant. The program according to any one of claims 1 to 4.

6. Performing the aforementioned evaluation includes evaluating the benign or malignant nature of the nodule based on a threshold value set as a threshold for the evaluation value used to evaluate the benign or malignant nature of the nodule. The program according to claim 5.

7. The aforementioned multiple regions are 360 / n regions obtained by rotating one region by n degrees. The program according to claim 1.

8. The aforementioned region is rectangular, Calculating the aforementioned evaluation value includes, for each of the multiple regions, dividing the space between two opposing sides at equal intervals to form multiple rows, dividing the space between two other opposing sides at equal intervals to form multiple columns, calculating the correlation value between adjacent rows or columns within the region, and obtaining the aforementioned array. The program according to claim 2.

9. An information processing device, An information processing device comprising a processing unit that, in a medical image, sets multiple regions including a region of interest and having different angles from each other, calculates evaluation values ​​for evaluating the region of interest based on the multiple regions, and performs an evaluation of the region of interest based on the evaluation values.

10. A method of information processing performed by a computer, In medical imaging, setting multiple regions that include the region of interest but have different angles from each other, Based on the aforementioned multiple domains, an evaluation value is calculated for evaluating the domain of interest, Based on the aforementioned evaluation values, an evaluation of the area of ​​interest is performed, Information processing methods including

Images