Medical information processing device, medical information processing method, and program

Abstract

Provided is a medical information processing device that can stabilize the performance of a medical information analysis system. A medical information acquisition unit 21 acquires relevant medical information relevant to medical information. A determination unit comprising a conversion formula selection unit 22 and a conversion formula adjustment unit 23 determines a conversion formula for converting the medical information on the basis of the relevant medical information. An information conversion unit 24 generates and outputs converted medical information obtained by converting the medical information using the conversion formula.

Description

Medical information processing device, medical information processing method, and program

[0001] The present disclosure relates to a medical information processing device, a medical information processing method, and a program.

[0002] In recent years, medical information analysis systems that assist doctors in assessing medical conditions have been attracting attention. Medical information analysis systems are sometimes constructed by performing machine learning and other methods using medical information about patients. However, the nature of medical information often varies depending on the individual patient or the facility and equipment from which the medical information was acquired. For this reason, medical information analysis systems face problems such as a decrease in accuracy when a set of information with different characteristics from the set of information used to construct the system is input, resulting in unstable system performance. In particular, in regional medical collaborations and collaborations between multiple medical facilities, such as medical tourism, the instability of performance due to the variation in the nature of this information is considered to be a major issue.

[0003] In response to this, Patent Document 1 discloses a technology for converting a medical image acquired by another medical image display device into the same image as that of the other medical image display device by LUT (Look Up Table conversion) conversion using a lookup table.

[0004] JP 2013-009877 A

[0005] However, in the technology described in Patent Document 1, since the lookup table is predetermined, it is difficult to eliminate the instability of the performance of a medical information analysis system that uses medical information whose properties vary depending on the patient, etc.

[0006] An object of the present disclosure is to provide a medical information processing device, a medical information processing method, and a program that are capable of stabilizing the performance of a medical information analysis system.

[0007] A medical information processing device according to one aspect of the present disclosure is a medical information processing device that converts medical information of a patient, and includes an acquisition unit that acquires related medical information related to the medical information, a determination unit that determines a conversion formula for converting the medical information based on the related medical information, and a conversion unit that generates and outputs converted medical information by converting the medical information using the conversion formula.

[0008] According to the present invention, it is possible to stabilize the performance of a medical information analysis system.

[0009] Fig. 1 is a diagram illustrating a configuration of a medical information analysis system according to an embodiment of the present disclosure; Fig. 2 is a flowchart for explaining an example of medical information processing; Fig. 3 is a diagram illustrating an example of a transformation formula; Fig. 4 is a diagram illustrating an example of a transformed medical image; Fig. 5 is a diagram illustrating another example of a transformed medical image; Fig. 6 is a flowchart for explaining another example of medical information processing; Fig. 7 is a diagram illustrating another example of a transformed medical image.

[0010] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0011] In this embodiment, we will explain a medical information processing device, medical information processing method, and program that can stabilize system performance even when the input information input to a medical information analysis system that provides decision support to help doctors make decisions about medical conditions, etc. is different in nature from the construction information used to construct the system (specifically, to construct the medical information analysis device 13 described below), that is, even when the nature of the information varies depending on the individual patient or the facility and equipment that acquired the information.

[0012] 1 is a diagram showing the configuration of a medical information analysis system according to this embodiment. The medical information analysis system 1 shown in Fig. 1 includes an input device 10, a medical information processing device 11, a medical information storage device 12, and a medical information analysis device 13.

[0013] The input device 10 accepts input information from an operator or a higher-level system and transmits the input information to the medical information processing device 11. The medical information processing device 11 executes medical information processing to generate and output converted medical information by converting the patient's medical information. The medical information storage device 12 stores various information generated or used by the medical information processing device 11. The medical information analysis device 13 analyzes the converted medical information from the medical information processing device 11.

[0014] The medical information processing device 11 includes a medical information acquisition unit 21 , a conversion formula selection unit 22 , a conversion formula adjustment unit 23 , and an information conversion unit 24 .

[0015] The medical information acquisition unit 21 is an acquisition unit that acquires medical information to be processed and related medical information related to the medical information.

[0016] The conversion formula selection unit 22 and the conversion formula adjustment unit 23 constitute a determination unit that determines a conversion formula (conversion rule) for converting medical information acquired by the medical information acquisition unit 21 based on the associated medical information acquired by the medical information acquisition unit 21. The conversion formula selection unit 22 is a selection unit that selects (determines) one of a plurality of candidate conversion formulas as a conversion formula based on the associated medical information. The conversion formula adjustment unit 23 is an adjustment unit that determines a conversion formula by adjusting the coefficients of a predetermined reference conversion formula based on the associated medical information. Note that the determination unit may include at least one of the conversion formula selection unit 22 and the conversion formula adjustment unit 23.

[0017] The information conversion unit 24 is a conversion unit that converts the medical information acquired by the medical information acquisition unit 21 using the conversion formula determined by the determination unit, and generates and outputs converted medical information.

[0018] Each device, such as the medical information processing device 11 and medical information analysis device 13 shown in Fig. 1, is configured by, for example, a computer system including a processor (computer) and memory (neither of which is shown). In this case, each of the above-described components and functions is realized, for example, by the processor reading a program and executing the read program. The program can be recorded on a computer-readable recording medium 30, such as a memory. The recording medium 30 non-temporarily stores a computer program or data. At least a portion of the computer program and data stored in the memory or storage unit can be stored on the recording medium 30.

[0019] Fig. 2 is a flowchart for explaining an example of medical information processing by the medical information processing device 11. In the example of Fig. 2, the medical information is a medical image of a patient, such as a computed tomography (CT) image or an X-ray fluoroscopic image, and more specifically, a medical image of the patient's lungs.

[0020] First, the medical information acquisition unit 21 acquires an input medical image, which is a medical image to be processed, and associated medical information related to the input medical image (step S101). The input medical image may be input from the input device 10, or may be acquired from the medical information processing device 11 in accordance with an instruction from the input device 10. In the example of Fig. 2, the associated medical information is patient information about the patient appearing in the input medical image. Note that the input medical image and the patient information each have an ID number or a name, for example, assigned as identification information for identifying the patient, and the medical information acquisition unit 21 acquires patient information assigned with the same identification information as the identification information of the input medical image.

[0021] Next, the transformation formula selection unit 22 acquires predetermined target information from the acquired patient information, which is the related medical information (step S102). The target information is information determined according to the analysis content by the medical information analysis device 13. In this embodiment, the target information is the patient's medical history of lung cancer, more specifically, whether or not the patient has a medical history of lung cancer.

[0022] The transformation formula selection unit 22 selects a transformation formula for transforming the input medical image from among a plurality of candidate transformation formulas prepared in advance, based on the patient's medical history of pneumonia (step S103). Note that, if there are multiple types of transformation formulas, it is sufficient to select at least one of the multiple types of transformation formulas. In the example of FIG. 2 , the transformation formulas include a lung field transformation formula for transforming the medical image into a lung field observation image (an image for observing the lung field), and a vertical contour region transformation formula for transforming the medical image into a vertical contour region observation image (an image for observing the vertical contour region). Furthermore, the transformation formula selection unit 22 selects one of the plurality of candidate transformation formulas for the lung field based on the patient's medical history of pneumonia, and selects a predetermined reference transformation formula for the vertical contour region, regardless of the patient's medical history of pneumonia.

[0023] The information converter 24 then converts the input medical image using the conversion formula for the lung field area and the conversion formula for the longitudinal region selected by the conversion formula selector 22 to generate converted medical images, which are converted medical information (step S104).The information converter 24 outputs the converted medical images to the medical information analyzer 13 (step S105), and the process ends.

[0024] The medical information analysis device 13 performs medical information analysis using two types of images, i.e., lung field observation images and longitudinal region observation images. Examples of this type of medical information analysis include an analysis in which two types of images are displayed and compared visually by a user, or an analysis in which predetermined regions or features, such as lung field regions or lesions, are detected based on the lung field observation images and longitudinal region observation images. The medical information analysis device 13 may also be constructed by machine learning using medical images as training data.

[0025] The patient information acquired in step S101 may be free-form information, such as an electronic medical record or a referral letter between medical facilities, or may be in a prescribed format generated by a dedicated application or the like. When the patient information is in a prescribed format, the history of pneumonia may be, for example, binary information indicating the presence or absence of pneumonia. In this case, in step S102, the conversion formula selection unit 22 simply extracts the history of pneumonia from the patient information. When the patient information is in free-form information, the conversion formula selection unit 22 determines the history of pneumonia from the patient information using, for example, a text analysis technique such as NLP (Neuro Linguistic Programming). When the patient information is in free-form and handwritten, the conversion formula selection unit 22 recognizes the handwritten character information using, for example, a character recognition technique such as OCR (Optical Character Recognition / Reader), and then determines the history of pneumonia using the text analysis technique.

[0026] 3 is a diagram showing an example of the conversion formula selected in step S103. In the example of FIG. 3, the conversion formula is a luminance conversion formula for converting the luminance value (pixel value) of the image. The conversion formula for the lung field region differs depending on whether or not there is a history of pneumonia, while the conversion formula for the longitudinal region is fixed regardless of whether or not there is a history of pneumonia.

[0027] 4A and 4B are diagrams for comparing a transformed medical image according to this embodiment with a transformed medical image according to a reference example. The reference example is an example in which transformed medical images (an image for observing the lung field area and an image for observing the longitudinal region) are generated using a predetermined transformation formula for the lung field area and a transformation formula for the longitudinal region. FIG. 4A shows a transformed medical image according to the reference example, and FIG. 4B shows a transformed medical image according to this embodiment. The input medical image is common to both the reference example and this embodiment.

[0028] In the reference example shown in Figure 4(a), for example, if the brightness values ​​of the lung field area in the input medical image are generally high, the image for the lung field area will be brighter than if the brightness values ​​are not high, making it difficult to distinguish between diseased areas, which often appear as high-brightness areas within the lung field, and normal lung field areas.

[0029] In contrast, in the case of patients with a history of pneumonia, the brightness of the lung field region of the input medical image is expected to be high due to the influence of inflammatory scars in the lung field, etc. Taking advantage of this, in this embodiment, a conversion formula is selected according to the patient's history of pneumonia, as shown in FIG. 4B. This makes it possible to select a conversion formula for the lung field region that provides higher resolution in high-brightness regions for patients with a history of pneumonia compared to patients without a history of pneumonia. Furthermore, a conversion formula similar to that of the reference example can be selected for the longitudinal region, regardless of whether or not the patient has a history of pneumonia. Therefore, the input medical image can be converted by increasing the resolution of high-brightness regions in the lung field observation image and by not unnecessarily emphasizing high-brightness regions in the longitudinal region observation image.

[0030] Note that the patient information used to select the transformation formula is not limited to a history of pneumonia. For example, if the input medical image is a medical image captured using a different imaging modality than the system configuration information, the characteristics of the image may differ depending on the patient's physique. For example, if the input medical image is an ultrasound image, the brightness value of the entire image is expected to vary depending on the distance between the imaging site and the probe or the body tissue between them. In this case, the determination unit, for example, obtains physique information regarding the patient's physique from the patient information, estimates the amount of fat or muscle between the probe and the imaging site from the physique information, and determines the transformation formula based on the estimated value.

[0031] Furthermore, the related medical information is not limited to patient information, but may also be the imaging conditions under which the input medical image was captured. Figure 5 is a diagram for comparing the converted medical image of this embodiment with the converted medical image of a reference example when a transformation formula is determined using the imaging conditions as related medical information. Figure 4(a) shows the converted medical image of the reference example, and Figure 4(b) shows the converted medical image of this embodiment. The example of Figure 5(a) is similar to the example of Figure 4(a).

[0032] 5A, in the reference example, for example, if the brightness values ​​of the input medical image are generally low, the brightness values ​​of the converted medical image will also be lower than in other cases, making it difficult to observe or identify lesions and the like that appear in the image, which are recognized mainly by the brightness value difference from the background. For example, if the image is captured at a lower-than-expected imaging dose, the brightness values ​​of the input medical image will be lower overall.

[0033] In contrast, in this embodiment, as shown in Fig. 5(b), the conversion formula selection unit 22 selects a conversion formula depending on whether the imaging dose included in the imaging conditions exceeds a threshold. Therefore, for both the lung field region conversion formula and the longitudinal region conversion formula, when the imaging dose is equal to or less than the threshold, the conversion formula selection unit 22 can select a conversion formula that increases the resolution of low-brightness regions compared to when the imaging dose exceeds the threshold. Therefore, it is possible to make both the lung field region observation image and the longitudinal region observation image easier to analyze.

[0034] If the input medical image is a DICOM (Digital Imaging and Communication in Medicine) image, the imaging conditions used as patient information can be acquired from its header. In this case, the imaging conditions are recorded as numerical values, and the conversion formula selection unit 22 can select a conversion formula by comparing the numerical values ​​with a threshold value. The imaging conditions may also be acquired from information recorded separately from the input medical image. In this case, if the imaging conditions are recorded as numerical values, the conversion formula selection unit 22 can select a conversion formula by comparing the numerical values ​​with a threshold value. In addition, if the imaging conditions are free-form or handwritten information, the conversion formula selection unit 22 quantifies the imaging conditions using text analysis technology, character recognition technology, or the like.

[0035] In the example described above, the transformation formula is determined by the transformation formula selection unit 22. However, hereinafter, an example will be described in which the transformation formula is determined by both the transformation formula selection unit 22 and the transformation formula adjustment unit 23. Fig. 6 is a flowchart for explaining an example of medical information processing in this case. Note that, as in the example of Fig. 2, the input information is a medical image of a patient's lungs. Furthermore, explanations of processes similar to those described in Fig. 2 will be omitted as appropriate.

[0036] First, the medical information acquisition unit 21 acquires an input medical image, which is a medical image to be processed, and related medical information related to the input medical image (step S201). In the example of Fig. 6, the related medical information is the imaging conditions under which the patient was imaged when the input medical image was acquired. The conversion formula selection unit 22 acquires the imaging dose as predetermined target information from the imaging conditions (step S202).

[0037] The conversion formula selection unit 22 selects a conversion formula for converting the input medical image (a conversion formula for the lung field region and a conversion formula for the longitudinal region) from among a plurality of candidate conversion formulas prepared in advance based on the imaging dose (step S203). As a method for selecting the conversion formula, for example, the method described with reference to FIG. 2 can be adopted.

[0038] Then, the conversion formula adjustment unit 23 adjusts the coefficients of the conversion formula selected by the conversion formula selection unit 22 based on the imaging dose (step S204). For example, the conversion formula adjustment unit 23 prepares in advance a calculation formula for calculating coefficients from the imaging dose, and adjusts the coefficients of the conversion formula to coefficients obtained by substituting the imaging dose into the calculation formula. The calculation formula is, for example, a formula for calculating coefficients such that the resolution of low-luminance value regions of the input medical image increases as the imaging dose increases.

[0039] Then, the information conversion unit 24 generates a converted medical image by converting the input medical image using the conversion formula whose coefficients have been adjusted by the conversion formula adjustment unit 23 (step S205), and the information conversion unit 24 outputs the converted medical image to the medical information analysis device 13 (step S206), thereby completing the processing.

[0040] Alternatively, instead of selecting a transformation formula by the transformation formula selector 22, the transformation formula adjuster 23 may adjust the coefficients of a predetermined reference transformation formula to determine the transformation formula for transforming the input medical image.

[0041] Next, an example will be described in which a three-dimensional medical image, more specifically a three-dimensional CT image, is used as the input medical image. The relevant medical information is the imaging conditions and reconstruction conditions of the three-dimensional CT image, and the conversion formula is a resampling formula for resampling the three-dimensional CT image. The reconstruction conditions are conditions for obtaining a three-dimensional CT image, which is a cross-sectional image, by analyzing X-ray transmission images, which are the original data from all directions obtained by CT scanning.

[0042] FIG. 7 is a diagram for explaining in more detail an example in which a three-dimensional CT image is used as an input medical image.

[0043] 7A shows a 3D CT image generated under the specified conditions, which are the imaging and reconstruction conditions for the 3D CT image used in building the system. Here, the imaging condition is the imaging dose, and the reconstruction conditions are the slice thickness and slice spacing.

[0044] 7B shows an example in which the slice thickness and imaging dose are equivalent to the prescribed conditions, but the slice spacing is narrower than the prescribed conditions. In this example, the image quality of each slice of the 3D CT image is assumed to be comparable to the prescribed image quality, but the resolution in the slice direction is higher than the prescribed resolution. Therefore, a resampling formula that thins out the slices is selected as the conversion formula.

[0045] FIG. 7( c) shows an example in which the slice spacing is equal to the specified condition, the slice thickness is narrower than the specified condition, and the imaging dose is higher than the specified condition. In this example, it is assumed that the image quality of each slice of the 3D CT image is higher than the specified image quality. Therefore, a resampling formula, which is a conversion formula, is selected that integrates the slices. Note that in the slice integration, the average value of the luminance values ​​of each integrated slice is used as the luminance value of the integrated slice. In other words, the input medical image is blurred.

[0046] Note that if the image quality of the input medical image is higher than the specified image quality, there is a risk that areas other than those being observed or analyzed (e.g., noise) will also be sharpened. For example, when the medical information analysis device 13 performs an analysis to automatically detect a lesion, if the image quality of the input medical image is higher than the specified image quality, there is a risk of false detection, such as recognizing a non-lesion area as a lesion. Therefore, if the image quality of the input medical image is higher than the specified image quality, the performance of the medical information analysis device 13 can be stabilized by blurring the input medical image.

[0047] While the above examples use medical images as medical information, medical information is not limited to medical images. For example, medical information may be the results of a patient's blood test. In this case, the transformation formula selection unit 22 selects a transformation formula that extracts necessary information from the blood test results or deletes unnecessary information based on personal information (e.g., whether or not the patient is pregnant). This allows for selection of not using test results when it is medically known that a specific physical condition, such as pregnancy, causes a change in the reference value of the test results (e.g., the reference range for white blood cell count) compared to other conditions (prescribed conditions). In this case, the medical information can be brought closer to the prescribed condition, thereby stabilizing the performance of the medical information analysis device 13.

[0048] The above-described embodiments of the present disclosure are merely illustrative examples of the present disclosure, and are not intended to limit the scope of the present disclosure to these embodiments alone. Those skilled in the art may implement the present disclosure in various other forms without departing from the scope of the present disclosure.

[0049] For example, the above-described embodiments have been described in detail to clearly explain the present invention, and are not necessarily limited to those including all of the described configurations. Furthermore, it is possible to replace part of the configuration of one embodiment with the configuration of another embodiment, and it is also possible to add the configuration of another embodiment to the configuration of one embodiment. Furthermore, it is also possible to add, delete, or replace part of the configuration of each embodiment with other configurations.

[0050] Furthermore, the above-described configurations, functions, processing units, processing means, etc. may be partially or entirely implemented in hardware, for example, by designing them as integrated circuits. The above-described configurations, functions, etc. may also be implemented in software, with a processor interpreting and executing a program that implements each function. Information such as the programs, tables, and files that implement each function can be stored in a memory, a recording device such as a hard disk or SSD (Solid State Drive), or a recording medium such as an IC card, SD card, or DVD.

[0051] In addition, the control lines and information lines shown are those that are considered necessary for the explanation, and do not necessarily show all the control lines and information lines in the product. In reality, it can be assumed that almost all components are interconnected.

[0052] 1: Medical information analysis system 10: Input device 11: Medical information processing device 12: Medical information storage device 13: Medical information analysis device 21: Medical information acquisition unit 22: Conversion formula selection unit 23: Conversion formula adjustment unit 24: Information conversion unit

Claims

1. A medical information processing device that converts a patient's medical information, comprising: an acquisition unit that acquires related medical information related to the medical information; a determination unit that determines a conversion formula for converting the medical information based on the related medical information; and a conversion unit that generates and outputs converted medical information by converting the medical information using the conversion formula.

2. The medical information processing device according to claim 1, wherein the determination unit includes a selection unit that determines, based on the related medical information, one of a plurality of candidate transformation formulas as the transformation formula.

3. The medical information processing device according to claim 1, wherein the determination unit includes an adjustment unit that determines the conversion formula by adjusting coefficients of a predetermined reference conversion formula based on the related medical information.

4. The medical information processing device according to claim 1, wherein the relevant medical information includes patient information relating to the patient.

5. The medical information processing device according to claim 4, wherein the medical information is a medical image of the patient's lungs, and the patient information includes the patient's medical history of pneumonia.

6. The medical information processing device according to claim 5, wherein the conversion formula includes a conversion formula for a lung field area that converts the medical image into a lung field area observation image for observing the lung field area, and a conversion formula for a vertical area that converts the medical image into a vertical area observation image for observing the vertical area.

7. The medical information processing device according to claim 1, wherein the medical information is a medical image of the patient.

8. The medical information processing device according to claim 7, wherein the related medical information includes imaging conditions for capturing the medical image.

9. The medical image processing device according to claim 7, wherein the conversion formula is a luminance conversion formula for converting a luminance value of the medical image.

10. The medical information processing device of claim 7, wherein the medical image is a three-dimensional medical image, the related medical information includes reconstruction conditions for reconstructing the three-dimensional medical image, and the transformation formula is a resampling formula for resampling the three-dimensional medical image.

11. A medical information processing method by a medical information processing device that converts a patient's medical information, comprising: acquiring related medical information related to the medical information; determining a conversion formula for converting the medical information based on the related medical information; and generating and outputting converted medical information by converting the medical information using the conversion formula.

12. A program for causing a computer that converts a patient's medical information to execute the steps of: acquiring associated medical information related to the medical information; determining a conversion formula for converting the medical information based on the associated medical information; and generating and outputting converted medical information by converting the medical information using the conversion formula.

Images