Analysis device, analysis method, and medical image processing device

Abstract

The invention provides an analysis apparatus, an analysis method, and a medical image processing apparatus. The analysis device includes: a target organ analysis unit that performs analysis on the basis of a medical image of a target organ to obtain related organ information including at least a related organ name and a degree of confidence of a related organ related to the target organ; the associated organ determining part is used for determining an associated organ according to the confidence coefficient of the associated organ on the basis of the associated organ information; and a control unit that performs control so as to perform a process in which the related organ identified by the related organ identification unit is used as a new target organ and the target organ is analyzed by the target organ analysis unit.

Description

Technical Field

[0001] The present invention relates to an analysis device, an analysis method and a medical image processing device for analyzing lesions in medical images based on deep learning. Background Art

[0002] In clinical research and application of medicine, analysis based on medical images acquired by medical imaging devices can assist doctors in quickly locating the positions of organs, tissues, tumors, etc.

[0003] In addition, in the clinical research and application of medicine, the primary tumor of a certain organ may spread to other organs, which is called metastatic cancer. The detection and segmentation of malignant tumors, i.e., the metastasis of cancer, plays an important auxiliary role in the diagnosis, staging, and treatment evaluation of doctors, and is an important task in medical image processing.

[0004] In the prior art, in case of possible metastatic lesions in the subject, the doctor needs to check the medical images obtained by shooting the relevant organs in sequence based on experience, manually search and determine the possible metastatic lesions and perform segmentation measurements. In such a processing process, first, the relevant organs of the subject need to be examined in sequence based on medical imaging devices such as ultrasonic diagnostic devices, CT (computed tomography) imaging devices, MR (magnetic resonance) imaging devices, etc., to obtain medical images such as ultrasonic images, CT images, and MR images, and then the doctor reads the obtained medical images one by one, which is time-consuming for both the subject and the doctor. Moreover, when the doctor reads the images manually, it is easy to miss metastatic lesions and misdiagnose.

[0005] In order to reduce missed diagnoses, in the prior art, each manufacturer has developed a fully automatic lesion detection and segmentation algorithm for each organ. However, the existing fully automatic lesion detection and segmentation algorithms all have certain false positive results. In addition, if the corresponding fully automatic lesion detection and segmentation algorithm is run on all organs (including organs without lesions), the number of false positives will become greater due to the accumulation of errors, and doctors will need to manually reconfirm these false positive lesions, which will bring additional workload to doctors.

[0006] Therefore, regarding the detection and segmentation of metastatic cancer, a technology that can analyze, detect, and segment metastatic cancer quickly and with high accuracy is required. Summary of the invention

[0007] In view of such circumstances, an object of the present invention is to provide an analysis device, an analysis method, and a medical image processing device for improving the detection accuracy and diagnosis and treatment efficiency of metastatic cancer.

[0008] In order to achieve the above-mentioned purpose, the analysis device of the present invention is characterized in that it includes: a target organ analysis unit, which performs analysis based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ; and an associated organ determination unit, which determines the associated organ based on the associated organ information and the confidence of the associated organ.

[0009] The analysis device of the present invention is characterized in that it includes: a target organ analysis unit, which performs analysis based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ; an associated organ determination unit, which determines the associated organ based on the associated organ information and the confidence of the associated organ; and a control unit, which controls and repeatedly performs the following processing until it is determined by the associated organ determination unit that there is no associated organ, and the processing refers to taking the associated organ determined by the associated organ determination unit as a new target organ, analyzing it by the target organ analysis unit, and then determining the next new associated organ by the associated organ determination unit.

[0010] In addition, the analysis device of the present invention is characterized in that the target organ analysis unit includes: a target organ image range determination unit, which determines the target organ image range of the medical image of the target organ in an automatic manner or manually by the user; a target organ lesion detection unit, which detects all lesions of the target organ within the target organ image range, and obtains lesion location information of each lesion and the associated organ information corresponding to each lesion for each lesion.

[0011] In addition, the analysis device of the present invention is characterized in that the target organ analysis unit also includes a lesion analysis unit, which analyzes each lesion of all the lesions based on the lesion position information to obtain lesion information of each lesion.

[0012] In addition, the analysis device of the present invention is characterized in that the target organ lesion detection unit detects all lesions of the target organ based on the first deep learning model, and obtains the lesion location information of each lesion and the associated organ information corresponding to each lesion for each lesion.

[0013] In addition, the analysis device of the present invention is characterized in that the first deep learning model is trained using the first training data to: for each organ, detect all lesions of the target organ, and for each lesion of all the lesions, obtain the lesion location information of each lesion and the associated organ information corresponding to each lesion.

[0014] In addition, the analysis device of the present invention is characterized in that the first training data includes examination images of the subject and corresponding doctor's diagnosis information, and multiple follow-up examination images of the subject and corresponding doctor's diagnosis information.

[0015] In addition, the analysis device of the present invention is characterized in that the first training data includes examination images of the subject and corresponding doctor's diagnosis information, multiple follow-up examination images of the subject and corresponding doctor's diagnosis information, and additional information including biological information and multimodal information.

[0016] In addition, the analysis device of the present invention is characterized in that the associated organ determination unit includes: a candidate associated organ determination unit, which determines all candidate associated organs based on the associated organ information; a confidence determination unit, which determines the confidence of each candidate associated organ for all the candidate associated organs; and a determination unit, which determines the candidate associated organ with the highest confidence or the candidate associated organ with a confidence above a specified threshold as the associated organ.

[0017] In addition, the analysis device of the present invention is characterized in that the target organ analysis unit obtains associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ based on the second deep learning model.

[0018] In addition, the analysis device of the present invention is characterized in that the second deep learning model is trained to output, for an organ, the associated organs of the organ and the confidence level of the associated organs.

[0019] In addition, the analysis method of the present invention is characterized in that it includes: a target organ analysis step, in which analysis is performed based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ; and an associated organ determination step, in which the associated organ is determined based on the associated organ information and according to the confidence of the associated organ.

[0020] In addition, the analysis method of the present invention is characterized in that it includes: a target organ analysis step, in which analysis is performed based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ; an associated organ determination step, in which the associated organ is determined based on the associated organ information and according to the confidence of the associated organ; and a control step, in which control is performed to repeatedly perform the following processing until it is determined that there is no associated organ by the associated organ determination step, and the processing refers to taking the associated organ determined by the associated organ determination step as a new target organ, analyzing it by the target organ analysis step, and then determining the next new associated organ by the associated organ determination step.

[0021] Furthermore, a medical image processing apparatus according to the present invention comprises: an image acquisition device for acquiring a medical image of a subject including at least a target organ; and any one of the above-described analysis devices.

[0022] [Effects of the invention]

[0023] As described above, according to the analysis device, analysis method, and medical image processing device of the present invention, it is possible to improve the detection accuracy and diagnosis and treatment efficiency of metastatic cancer. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 It is a schematic diagram showing the configuration of the analysis device according to the first embodiment of the present invention.

[0025] Figure 2 This is an operation flow chart of the analysis device according to the first embodiment of the present invention.

[0026] Figure 3 It is a schematic diagram of a specific example of the analysis device according to the first embodiment of the present invention.

[0027] Figure 4 It is a schematic diagram showing the configuration of an analysis device according to a second embodiment of the present invention.

[0028] Figure 5 This is an operation flow chart of the analysis device according to the second embodiment of the present invention.

[0029] Figure 6 It is a schematic diagram showing the configuration of an analysis device according to a third embodiment of the present invention.

[0030] Figure 7 This is an operation flow chart of the target organ analyzing unit of the analyzing device according to the third embodiment of the present invention.

[0031] Figure 8A to Figure 8B It is a schematic diagram of a specific example of the analysis device according to the third embodiment of the present invention.

[0032] Fig. 9A It is a schematic diagram showing the configuration of an analysis device according to a modified example of the third embodiment of the present invention.

[0033] Fig. 9B This is an operation flow chart of the target organ analyzing unit of the analyzing device according to the modified example of the third embodiment of the present invention.

[0034] Fig. 9C This is a schematic diagram of presentation information of an analysis device according to a modified example of the third embodiment of the present invention.

[0035] Fig.10 It is a schematic diagram showing the configuration of an analysis device according to a fourth embodiment of the present invention.

[0036] Fig.11 This is an operation flow chart of the analyzing device according to the fourth embodiment of the present invention.

[0037] Fig.12 It is a schematic diagram of an example of an analysis device according to a fourth embodiment of the present invention. DETAILED DESCRIPTION

[0038] Hereinafter, the analyzing device, the analyzing method, and the medical image processing device according to the present invention will be described with reference to the drawings.

[0039] (First Embodiment)

[0040] ※ Configuration of the analysis device 100

[0041] First, refer to Figure 1 , the structure of the analysis device 100 according to the first embodiment of the present invention is described. In addition, the analysis device includes various components, Figure 1 Only components related to the technical concept of the present invention are shown, while other components are omitted.

[0042] like Figure 1 As shown, the analysis device 100 of the present invention includes a target organ analysis unit 101 and a related organ identification unit 102 .

[0043] The target organ analysis unit 101, for example, can be composed of components or modules with information processing functions such as a CPU or MCU, and is used to analyze the medical image of the target organ to obtain associated organ information including at least the name of the associated organ associated with the target organ and the confidence of the associated organ.

[0044] Here, the target organ is an organ of the subject, which is the initial target organ for the analysis device to perform metastatic cancer analysis, such as the colon, liver, lung, lymph node, etc. In addition, the target organ may be an organ known to have a cancer lesion, and the associated organ refers to an organ where a cancer lesion may exist, and the cancer lesion of the associated organ is a lesion associated with the cancer lesion of the target organ, that is, a metastatic lesion. The initial target organ can be specified by the user, or it can be specified by performing known image processing such as image segmentation on a medical image containing the target organ.

[0045] The confidence level of the associated organ is a quantified value indicating the possibility of association with the target organ.

[0046] In addition, the medical image of the target organ only needs to be a medical image that includes the entire target organ. The medical image of the target organ can be an ultrasonic image, a CT image, an MR image or other medical image captured by a medical imaging device such as an ultrasonic diagnostic device, a CT (computed tomography) imaging device, or an MR (magnetic resonance) imaging device. That is, the medical image can be a three-dimensional image or a two-dimensional image. In addition, the medical image of the target organ can be obtained from a medical imaging device through communication, for example, by reading a storage device that pre-stores the medical image of the target organ, or by reading a medical image that has been processed by other components. In addition, in the present invention, it is provided that the medical image of the target organ can be obtained at any time as needed, and the analysis device can obtain any required medical image of the target organ.

[0047] In addition, the target organ analysis unit 101 may obtain the associated organ information by, for example, analyzing the examination images of the subject and the corresponding doctor's diagnosis information, or multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information.

[0048] As an example, when the examination image of the subject shows that in addition to the liver which is the target organ, there are cancer lesions in the lungs, based on the examination image of the subject and / or the corresponding doctor's diagnosis information, the target organ analysis unit 101 may regard the lungs as an associated organ of the target organ, i.e., the liver.

[0049] As another example, when the examination image of the subject shows cancer lesions in the lungs in addition to the liver as the target organ, and the examination image of the subsequent examination of the subject shows cancer lesions in the lymph nodes, the target organ analysis unit 101 can analyze and obtain analysis results such as "the associated organs of the target organ liver are the lungs and lymph nodes", "the associated organs of the target organ liver are the lungs", "the associated organs of the target organ liver are the lymph nodes", "the associated organs of the target organ are the lungs and lymph nodes, and the associated organs of the lungs are the lymph nodes".

[0050] In addition, the confidence of the associated organ included in the associated organ information can also be obtained by various well-known methods based on the examination image of the subject and the corresponding doctor's diagnosis information, the multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information. In this way, the target organ analysis unit 101 can analyze the medical image of the target organ and obtain the associated organ information including at least the name of the associated organ associated with the target organ and the confidence of the associated organ.

[0051] In addition, the method for the target organ analysis unit 101 to obtain related organ information can also be obtained by performing deep learning based on the examination images of the subject and the corresponding doctor's diagnosis information, and multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information, and then obtaining it based on a deep learning model obtained by deep learning. The method for obtaining related organ information based on deep learning will be described in detail later.

[0052] In addition, the form of representation of the associated organ information may be an image superimposed on a medical image of the target organ, or a table or database that at least associates the target organ, the associated organ, and the confidence level of the associated organ, or text data that at least includes the information on the target organ, the associated organ, and the confidence level of the associated organ.

[0053] In addition, the associated organ determination unit 102 may be composed of a component or module having an information processing function, such as a CPU or an MCU, and is used to determine the associated organ based on the associated organ information and the confidence level of the associated organ.

[0054] The method by which the associated organ determination unit 102 determines the associated organs may be, for example, based on the associated organ information, taking all the associated organs whose names are included in the associated organ information, or the associated organs whose confidences are included in the associated organ information and are above a specified value as candidate associated organs, and taking the unique value or the maximum value of the confidences of the candidate associated organs as the confidences of the candidate associated organs according to the confidences of the associated organs included in the associated organ information, and determining the candidate associated organ with the highest confidence or the candidate associated organ with a confidence above a specified threshold as the associated organ. In addition, when a candidate associated organ has been subjected to lesion detection, the confidence of the candidate associated organ is set to 0.

[0055] In addition, the associated organ determined by the associated organ determination unit 102 may be one, a specified number, or multiple numbers with confidences above a specified threshold. In addition, when the confidences included in the associated organ information are all below a specified lower limit, the associated organ determination unit 102 may determine that there is no associated organ.

[0056] In addition, the above-mentioned specified values, specified thresholds, and specified lower limits can be set or appropriately changed in a variety of ways. For example, they can be set by user input, or automatically set based on medical images, organs, user usage preferences, confidence calculation methods, etc.

[0057] ※Operation of the analysis device 100

[0058] Next, an analysis method applied by the analysis device 100 according to the first embodiment of the present invention will be described. Figure 2 This is an operation flow chart of the analysis device according to the first embodiment of the present invention. Figure 3It is a schematic diagram of a specific example of the analysis device according to the first embodiment of the present invention.

[0059] like Figure 2 As shown, after the analysis device 100 starts processing, in step S100, the target organ analysis unit 101 analyzes the medical image of the target organ to obtain the associated organ information including at least the name of the associated organ and the confidence of the associated organ associated with the target organ, and then enters step S200. Figure 3 The abdominal medical image including the target organ colon as shown in the upper paragraph is obtained Figure 3 The related organ information pointed by the arrows in the middle are “related organs: liver (0.85)” and “related organs: liver (0.78); lung (0.66)”.

[0060] In step S200, the associated organ determination unit 102 determines the associated organ according to the confidence of the associated organ based on the associated organ information obtained by the target organ analysis unit 101 in step S100, and then ends the processing. Here, the associated organ information obtained by the target organ analysis unit 101 can use "liver" and "lung" as candidate associated organs. For the candidate associated organ "liver", there are two confidence values ​​of 0.85 and 0.78, and 0.85 can be set as the confidence of the candidate associated organ "liver". In addition, for the candidate associated organ "lung", its confidence is a unique value, namely 0.66. In this way, when the condition of determining the candidate associated organ with the highest confidence value as the associated organ is adopted, the associated organ determination unit 102 can determine the candidate associated organ "liver" with a confidence of 0.85 as the associated organ. In addition, when the condition that the confidence of the candidate related organs is equal to or greater than a predetermined value (for example, “0.6”) is adopted, the related organ identification unit 102 may identify the two organs of the candidate related organs “liver” and “lung” as related organs.

[0061] According to the analysis device of the first embodiment, based on the medical image of the target organ, the associated organ information including the confidence of the associated organ is obtained, and based on the confidence of the associated organ, the associated organ is automatically detected and determined. Therefore, compared with the prior art, it is not necessary to run the corresponding fully automatic lesion detection and segmentation algorithm for all organs (including organs without lesions), which avoids additional useless calculations, avoids the accumulation of errors, and reduces false positives; in addition, it reduces the doctor's manual reconfirmation of these false positives and reduces the doctor's workload; in addition, it is not necessary to manually confirm the associated organs as in the prior art, avoiding dependence on the doctor's experience and reducing misdiagnosis and missed diagnosis caused by manual confirmation.

[0062] (Second Embodiment)

[0063] ※ Configuration of the analysis device 100A

[0064] Reference Figure 4 , the structure of the analysis device 100A according to the second embodiment of the present invention is described. Figure 4 It is a schematic diagram showing the configuration of an analysis device 100A according to a second embodiment of the present invention.

[0065] like Figure 4 As shown, the analyzing device 100A of this embodiment includes a control unit 103 in addition to the target organ analyzing unit 101 and the related organ identifying unit 102 .

[0066] Here, the target organ analyzing unit 101 and the related organ identifying unit 102 are respectively the same as the target organ analyzing unit 101 and the related organ identifying unit 102 in the first embodiment, and will not be described in detail here.

[0067] The control unit 103 of the analysis device 100A of this embodiment can be composed of components or modules with information processing functions, such as a CPU or MCU. The control unit 103 controls and repeatedly performs the following processing until it is determined as no associated organ by the associated organ determination unit 102. The processing refers to taking the associated organ determined by the associated organ determination unit 102 as a new target organ, analyzing it through the target organ analysis unit 101, and then determining the next new associated organ through the associated organ determination unit.

[0068] Furthermore, the control unit 103 may serve as a control unit for the entire analysis device, be connected to each component included in the analysis device, and coordinate and control the operation of each component.

[0069] ※Operation of the analysis device 100A

[0070] Next, an analysis method applied to the analysis device 100A according to the second embodiment of the present invention will be described. Figure 5 This is an operation flow chart of the analysis device 100A according to the second embodiment of the present invention.

[0071] like Figure 5 As shown, after the analysis device 100A starts processing, it executes the processing of steps S100 to S200, and then proceeds to step S300. Here, the processing of steps S100 to S200 is the same as Figure 2 The processing of steps S100 to S200 is the same, and the repeated detailed description is omitted here.

[0072] In step S300, the analysis device 100A determines through the control unit 103 whether it is determined as an unrelated organ by the associated organ determination unit 102. When it is determined that it is an unrelated organ by the associated organ determination unit 102 (step S300 is Y), the processing is terminated. On the other hand, when it is determined that it is an associated organ by the associated organ determination unit 102 (step S300 is N), it enters step S301.

[0073] In step S301, the control unit 103 controls to use the associated organ determined by the associated organ determination unit 102 as a new target organ, and then the processing returns to step S100, and steps S100 to 200 are repeatedly executed, that is, the new target organ is analyzed by the target organ analysis unit 101, and then the next new associated organ is determined by the associated organ determination unit 102.

[0074] The analysis device according to this embodiment has the effect of the first embodiment. In addition, since it is provided with a control unit 103, it controls and repeatedly performs the process of "taking the associated organ determined by the associated organ determination unit 102 as a new target organ, analyzing it through the target organ analysis unit 101, and then determining the next new associated organ through the associated organ determination unit" until it is determined as no associated organ by the associated organ determination unit 102. As a result, all associated organs can be automatically acquired based on the medical image of the target organ without the user performing any operation. The efficiency and accuracy of diagnosis and treatment are further improved.

[0075] (Third Embodiment)

[0076] ※ Configuration of the analysis device 100B

[0077] First, refer to Figure 6 , the structure of the analysis device 100B according to the third embodiment of the present invention is described. Figure 6 It is a schematic diagram showing the configuration of an analysis device 100B according to a third embodiment of the present invention.

[0078] like Figure 6 As shown, the analysis device 100B of this embodiment is provided with a target organ analysis unit 101 and an associated organ determination unit 102 as in the first embodiment, and is different from the first embodiment in that the target organ analysis unit 101 includes a target organ image range determination unit 1011 and a target organ lesion detection unit 1012. In addition, the associated organ determination unit 102 is the same as the associated organ determination unit 102 in the first embodiment, and thus will not be described in detail here.

[0079] In this embodiment, the target organ image range determination unit 1011, for example, can be composed of components or modules with information processing functions such as CPU or MCU, and is used to determine the target organ image range for the medical image of the target organ automatically or manually by the user.

[0080] Here, the “target organ image range” is, for example, an image range including the entire target organ.

[0081] The method for determining the range of the target organ may be, for example, that the user specifies (for example, performs a selection operation) the position of a certain part (for example, a lesion) on the target organ with respect to the medical image of the subject, and then automatically uses a known organ segmentation algorithm or an organ detection algorithm to automatically determine the target organ image range. Alternatively, the user specifies the target organ with respect to the medical image of the subject, and then automatically uses a known organ segmentation algorithm or an organ detection algorithm to automatically determine the target organ image range. Alternatively, the user specifies the image range that includes the target organ with respect to the medical image of the subject, and then automatically uses a known organ segmentation algorithm or an organ detection algorithm to optimize the image range that includes the target organ (for example, expand, translate, etc.) so that the target organ image range area includes the complete target organ.

[0082] In addition, the target organ lesion detection unit 1012, for example, can be composed of components or modules with information processing functions such as a CPU or MCU, and is used to detect all lesions of the target organ within the target organ image range determined by the target organ image range determination unit 101, and for each lesion of all detected lesions, obtain the lesion location information of each lesion and the associated organ information corresponding to each lesion.

[0083] Here, the lesion location information represents the relationship between the lesion and the medical image (or the target organ image range), and other similar information such as mask and radiomics information may be used.

[0084] In addition, the associated organ here is similar to the above information, and is also an organ where cancer lesions may exist, and the cancer lesions of the associated organ are lesions associated with the cancer lesions of the target organ, i.e., metastatic lesions. In addition, the associated organ information here, like the above, also includes at least the name of the associated organ and the confidence. In addition, the associated organ information of this embodiment corresponds to the target organ as described above, and further corresponds to the lesions of the target organ. That is to say, when there are multiple lesions on the same organ, the multiple lesions may be associated with the same or different other organs.

[0085] In addition, the method by which the target organ lesion detection unit 1012 detects all lesions of the target organ based on the target organ image range and obtains the lesion location information of each lesion and the associated organ information corresponding to each lesion may be, for example, a method based on deep learning.

[0086] For example, a deep learning model can be trained for each organ in advance or in real time. The deep learning model is trained to detect lesions for each organ and output associated organ information. In addition, the training data includes, for example, the examination image of the subject and the corresponding doctor's diagnosis information, and multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information.

[0087] In addition, a deep learning model may be trained for each organ in advance or in real time. The deep learning model is trained to detect only lesions for each organ.

[0088] In addition, in the case of obtaining organ lesions based on a deep learning model, the confidence level can be defined as follows: for a current lesion and a specified organ, the value indicating the possibility that the specified organ is an associated organ of the current lesion is defined as follows:

[0089]

[0090] Among them, A i is the similarity between the lesion location information and the i-th annotation (GT, Ground Truth) learned based on the deep learning model (the value range is 0 to 1, 1 is exactly the same, and 0 is completely different);

[0091] B i Is the specified organ an associated organ as noted above (a value of 1 indicates yes, a value of 0 indicates no);

[0092] n is the total number of annotated lesions used in deep learning model training.

[0093] The above definition of the confidence level of the related organ is merely an example and is not limited thereto, and various existing definitions may be adopted.

[0094] ※Operation of the analysis device 100B

[0095] Next, an analysis method applied to the analysis device 100B according to the third embodiment of the present invention will be described. Figure 7 This is a flowchart of the operation of the target organ analyzing unit 101 of the analyzing device 100B according to the third embodiment of the present invention. Figure 8A to Figure 8B It is a schematic diagram of a specific example of the analysis device according to the third embodiment of the present invention.

[0096] Figure 7 The action flow chart shown is an action flow chart of the target organ analysis unit 101, and thus can be substituted for Figure 2 and Figure 5 Step S100 in .

[0097] like Figure 7As shown, in the target organ analysis unit 101 of the analysis device 100B, first in step S101, the target organ image range determination unit 1011 determines the target organ image range for the medical image of the target organ automatically or manually by the user, and then enters step S102.

[0098] Here, for example Fig. 8A As shown, when the user uses an indication means not shown in the figure to select a medical image ( Fig. 8A When the position indicated by the dot (a certain position on the colon) is clicked in the image on the left, the target organ image range determination unit 1011 uses a known organ segmentation algorithm or an organ detection algorithm to segment or detect the organ at the position indicated by the dot, i.e., the target organ, and uses a known image cropping algorithm to crop the target organ from the medical image. Fig. 8A The image range including the target organ shown on the right is the target organ image range. By cropping the target organ image range from the medical image of the target organ, on the one hand, it does not affect the segmentation or detection of the target organ, and does not affect the analysis accuracy. On the other hand, it can reduce the image size, avoid useless calculations, and improve the analysis, detection and diagnosis and treatment efficiency.

[0099] In step S102, the target organ lesion detection unit 1012 detects all lesions of the target organ within the target organ image range, and obtains the lesion position information of each lesion and the associated organ information corresponding to each lesion for each lesion, and then ends the processing.

[0100] Here, for example Figure 8B As shown in the figure, the target organ image range (such as Figure 8B The left figure) is input into the pre-trained deep learning network, and the following is obtained: Figure 8B Specifically, two lesions "lesion 1" and "lesion 2" indicated by dotted lines and solid lines, lesion position information of each of "lesion 1" and "lesion 2", and associated organ information corresponding to "lesion 1" and "lesion 2" are obtained.

[0101] (Variation Example)

[0102] This embodiment is not limited to the above structure, and may have various modifications. For example, Fig. 9A A schematic diagram showing the configuration of an analysis device according to a modified example of the present embodiment.

[0103] like Fig. 9AAs shown, the analysis device 100BB of the modified example of this embodiment is provided with a target organ analysis unit 101 and an associated organ determination unit 102 as in the third embodiment, and the target organ analysis unit 101 includes a target organ image range determination unit 1011 and a target organ lesion detection unit 1012. Among them, the associated organ determination unit 102, the target organ image range determination unit 1011, and the target organ lesion detection unit 1012 of this embodiment are the same as the associated organ determination unit 102, the target organ image range determination unit 1011, and the target organ lesion detection unit 1012 of the third embodiment, and therefore will not be described in detail here.

[0104] In addition, the analyzing device 100BB according to the modified example of the present embodiment is different from the third embodiment in that the target organ analyzing unit 101 further includes a lesion analyzing unit 1013 .

[0105] The lesion analysis unit 1013 may be composed of components or modules with information processing functions, such as a CPU or MCU, and is used to analyze each lesion of all lesions based on the lesion location information obtained by the target organ lesion detection unit 1012 to obtain lesion information of each lesion.

[0106] Here, the lesion information includes but is not limited to lesion contour segmentation, measurement value calculation related to the length of the major and minor axes and volume of the tumor, histogram calculation, texture analysis, etc.

[0107] The lesion information can be presented to the user together with the lesion location information and / or associated organ information (e.g. Fig. 9C ), which helps users judge the reliability of related organ information and helps improve the accuracy of users' judgment on cancer metastasis.

[0108] In addition, although not shown in the figure, the analysis device may also include a prompting unit for prompting the user various information such as associated organ information, lesion location information, lesion information, etc. As a method of prompting various messages, prompting can be performed by light, warning sound, etc., or by sound and / or screen as needed.

[0109] in addition, Fig. 9B This is a flowchart of the operation of the target organ analyzing unit 101 of the analyzing device 100BB according to the modified example of the third embodiment of the present invention. Fig. 9B The action flow chart shown is an action flow chart of the target organ analysis unit 101, and thus can be substituted for Figure 2 and Figure 5 Step S100 in .

[0110] like Fig. 9BAs shown, in the target organ analysis unit 101 of the analysis device 100BB, the operations of steps S101 to S102 are first performed, and then the process proceeds to step S103. Here, the operations of steps S101 to S102 are similar to those of steps S103. Figure 7 The actions of steps S101 to S102 are the same as those in the above description and will not be repeated here.

[0111] In step S103 , the lesion analysis unit 1013 analyzes each lesion of all lesions based on the lesion position information acquired by the target organ lesion detection unit 1012 in step S102 to obtain lesion information of each lesion.

[0112] The analysis device according to this embodiment and its variants has the technical effect of the first embodiment. In addition, since it is equipped with a target organ image range determination unit 1011 and a target organ lesion detection unit 1012, it can detect all lesions of the target organ, and for each lesion of all the detected lesions, obtain the lesion location information of each lesion and the associated organ information corresponding to each lesion. In addition, it can also be equipped with a lesion analysis unit 1013 to obtain the lesion information of each lesion. In this way, the analysis device of this embodiment and its variants can obtain more detailed information about organs and lesions, which helps users to quickly, accurately and reliably make judgments and diagnose and treat metastatic cancer.

[0113] (Fourth Embodiment)

[0114] ※ Configuration of the analysis device 100C

[0115] Reference Fig.10 , the structure of the analysis device 100C according to the fourth embodiment of the present invention is described. Fig.10 It is a schematic diagram showing the configuration of an analysis device according to a fourth embodiment of the present invention.

[0116] like Fig.10As shown, the analysis device 100C of this embodiment is provided with a target organ analysis unit 101, an associated organ determination unit 102 and a control unit 103 in the same manner as the modified example of the third embodiment, and the target organ analysis unit 101 includes a target organ image range determination unit 1011, a target organ lesion detection unit 1012 and a lesion analysis unit 1013. The difference from the modified example of the third embodiment is that the associated organ determination unit 102 includes a candidate associated organ determination unit 1021, a confidence determination unit 1022 and a determination unit 1023. In addition, the target organ analysis unit 101, the control unit 103, the target organ image range determination unit 1011, the target organ lesion detection unit 1012 and the lesion analysis unit 1013 are the same as the target organ analysis unit 101, the control unit 103, the target organ image range determination unit 1011, the target organ lesion detection unit 1012 and the lesion analysis unit 1013 in the modified example of the third embodiment, and therefore will not be described in detail here.

[0117] The candidate associated organ determination unit 1021 , for example, may be composed of a component or module having information processing functions such as a CPU or an MCU, and is used to determine all candidate associated organs based on the associated organ information obtained by the target organ analysis unit 101 .

[0118] The confidence determination unit 1022 , for example, may be composed of a component or module having information processing functions such as a CPU or an MCU, and is used to determine the confidence of each candidate associated organ for all the candidate associated organs determined by the candidate associated organ determination unit 1021 .

[0119] The determination unit 1023, for example, can be composed of a component or module with information processing functions such as a CPU or MCU, and is used to determine the candidate associated organ with the highest confidence determined by the confidence determination unit 1022 or the candidate associated organ with a confidence above a specified threshold as an associated organ.

[0120] ※Operation of the analysis device 100C

[0121] Next, an analysis method applied to the analysis device 100C according to the fourth embodiment of the present invention will be described. Fig.11 This is an operation flow chart of the analyzing device 100C according to the fourth embodiment of the present invention.

[0122] like Fig.11 As shown, after the analysis device 100C starts processing, it executes the processing of steps S101 to S103, then enters the processing of steps S201 to 203, and then enters the processing of steps S300 to S301. Here, the processing of steps S101 to S103 is equivalent to Figure 5 The processing of step S100 and steps S201 to S203 are equivalent to Figure 5The processing of step S200 in step S101 to S103 is the same as Fig. 9B The processing of steps S101 to S103 is the same as that of steps S300 to S301. Figure 5 The processing of step S300 to step S301 is the same, so the repeated detailed description is omitted here, and only the processing of steps S201 to S203 is explained.

[0123] In step S201 , the candidate related organ specifying unit 1021 of the related organ specifying unit 102 specifies all candidate related organs based on the related organ information obtained in step S102 , and then proceeds to step S202 .

[0124] In step S202 , the confidence determination unit 1022 of the related organ determination unit 102 determines the confidence of each candidate related organ for all candidate related organs determined by the candidate related organ determination unit 1021 in step S201 , and then proceeds to step S203 .

[0125] In step S203, the determination unit 1023 of the related organ determination unit 102 determines the candidate related organ with the highest confidence determined by the confidence determination unit 1022 in step S202 or the candidate related organ with a confidence greater than a predetermined threshold as a related organ, and then proceeds to step S300.

[0126] (Example)

[0127] Next, embodiments of the fourth embodiment will be described with reference to specific examples.

[0128] Fig.12 It is a schematic diagram of a specific example of the operation of the analyzing device 100C according to the fourth embodiment of the present invention.

[0129] exist Fig.12 The medical image shown in is a medical image of a subject. In this embodiment, it is assumed that the analysis device 100C according to the fourth embodiment performs analysis on the medical image of the subject regarding metastatic cancer.

[0130] First, the analysis device 100C determines the target organ image range based on the medical image of the subject through the target organ image range determination unit 1011. In this embodiment, it is assumed that the user clicks on the colon part of the medical image of the subject, and the determined target organ image range is as shown in the reference Fig. 8A As shown, the target organ is the colon ( Fig.11 Step S101).

[0131] Next, the target organ lesion detection unit 1012 detects all lesions of the target organ, i.e., the colon, and obtains the lesion location information of each lesion and the associated organ information corresponding to each lesion ( Fig.11 For example, a lesion on the colon is detected ( Fig.12 The dots shown in the lower colon area are obtained, and the related organ information such as “lesion 1 (lesion location: ####); related organs: liver (0.8), lung (0.6), stomach (0.1)” is obtained.

[0132] Next, the lesion analysis unit 1013 performs lesion analysis on all lesions of the colon to obtain lesion analysis information ( Fig.11 Step S103).

[0133] Next, the candidate related organ determination unit 1021 determines all candidate related organs ( Fig.11 In step S201), in this embodiment, all candidate associated organs are liver, lung, and stomach.

[0134] Next, the confidence determination unit 1022 determines the confidence of each candidate associated organ for all candidate associated organs ( Fig.11 In this embodiment, the confidences of the candidate associated organs are (0.8), (0.6), and (0.1), respectively.

[0135] Next, the determination unit 1023 determines the candidate related organ with the highest confidence or the candidate related organ with a confidence greater than a predetermined threshold as the related organ ( Fig.11 In step S203). In this embodiment, the candidate associated organ with the highest confidence is determined as the associated organ. Therefore, the liver among the candidate associated organs is determined as the associated organ because of the highest confidence (0.8).

[0136] Next, the control unit 103 determines whether it is determined to be an unrelated organ ( Fig.11 In this embodiment, since the liver is determined as an associated organ of the target organ colon, the judgment in step S300 is no.

[0137] Next, the control unit 103 determines the associated organ (liver) as a new target organ ( Fig.11 ) and returns to step S101.

[0138] At this point, the analysis device 100C has completed Fig.12 The processing of this cycle of "1" shown.

[0139] Next, start Fig.12 The process of the cycle "2" shown in the figure. Fig.12 The processing of each cycle of "2", "3", and "4" shown is similar to the specific process of the cycle of "1" above. Therefore, only a brief description is given below.

[0140] exist Fig.12 In the cycle "2" shown, the following processing is performed.

[0141] First, the target organ image range determining unit 1011 of the analyzing device 100C determines the target organ image range (i.e., the liver) of a new target organ based on the medical image of the subject. Fig.11 Step S101).

[0142] Next, the target organ lesion detection unit 1012 detects two lesions in the target organ, namely the liver ( Fig.12 ), and obtained the related organ information such as “lesion 1 (lesion location: ######; related organs: liver (0.8), lung (0.5)”, “lesion 2 (lesion location: %%%%%%); related organs: lymph (0.4), lung (0.6), stomach (0.1)” ( Fig.11 Step S102).

[0143] Next, the lesion analysis unit 1013 performs lesion analysis on all lesions of the liver to obtain lesion analysis information ( Fig.11 Step S103).

[0144] Next, the candidate related organ determination unit 1021 determines that all candidate related organs are liver, lung, lymph node, stomach ( Fig.11 Step S201).

[0145] Next, the confidence determination unit 1022 determines the confidence of each of the candidate associated organs, liver, lung, lymph, and stomach, for all candidate associated organs. The confidence of the liver among the candidate associated organs is the highest, which is (0.8), but since it has been set as a target organ for detection, its confidence as a candidate associated organ is 0. Therefore, the confidence of the candidate associated organs, liver, lung, lymph, and stomach, are respectively (0), (0.6) (here, the maximum value between 0.5 and 0.6 is taken), (0.4), (0.1) ( Fig.11 Step S202).

[0146] Next, the determination unit 1023 determines that the associated organ is the lung ( Fig.11 Step S203).

[0147] Next, the control unit 103 determines that it is not an unrelated organ ( Fig.11 (the answer of step S300 is No).

[0148] Next, the control unit 103 determines the associated organ (lung) as a new target organ ( Fig.11 ) and returns to step S101.

[0149] At this point, the analysis device 100C has completed Fig.12 The process of the cycle "2" shown.

[0150] Next, start Fig.12 The processing of the cycle "3" shown in Fig.12 In the loop "3" shown, the same processing as the loop "2" is performed, "lymph" is determined as the associated organ and "lymph" is set as the new target organ, and the process returns to step S101.

[0151] Next, start Fig.12 The processing of the cycle "4" shown in Fig.12 In the cycle "4" shown, the following processing is performed.

[0152] First, the analyzing device 100C determines the target organ image range (i.e., the lymph node) of a new target organ based on the medical image of the subject by the organ image range determining unit 1011. Fig.11 Step S101).

[0153] Next, the target organ lesion detection unit 1012 detects three lesions of the target organ, i.e., lymph nodes ( Fig.12 The lymphatic part is indicated by the dotted line), and the related organ information such as "Focus 1 (lesion location: #####33#); related organs: liver (0.8), lung (0.4)", "Focus 2 (lesion location: %%%%%%%%); related organs: lung (0.6), stomach (0.05)", "Focus 3 (lesion location: %%%%%%%%%%); related organs: lung (0.5), lymph (0.5), stomach (0.1)" was obtained ( Fig.11 Step S102).

[0154] Next, the lesion analysis unit 1013 performs lesion analysis on all lymphatic lesions to obtain lesion analysis information ( Fig.11 Step S103).

[0155] Next, the candidate related organ determination unit 1021 determines that all candidate related organs are liver, lung, lymph node, stomach ( Fig.11 Step S201).

[0156] Next, the confidence determination unit 1022 determines the confidence of each of the candidate associated organs, liver, lung, lymph, and stomach, for all candidate associated organs. Since the liver, lung, and lymph among the candidate associated organs have been set as target organs for detection, their confidence as candidate associated organs is 0. Therefore, the confidence of the candidate associated organs, liver, lung, lymph, and stomach, are (0), (0), (0.1), (0.2), (0.3), (0.4), (0.5), (0.6), (0.7), (0.8), (0.9), (0.10), (0.11), (0.12), (0.13), (0.14), (0.15), (0.16), (0.17), (0.18), (0.19), (0.11), (0.17), (0.19), (0.11), (0.12), (0.13), (0.14), (0.15), (0.16), (0.17), (0.18), (0.19), (0.11), (0.12), (0.13), (0.14), (0.15), (0.16), (0.16), (0.17), (0.13), (0.1 Fig.11 Step S202).

[0157] Next, the determination unit 1023 determines that the stomach is not an associated organ because the confidence level of the stomach as an associated organ is only 0.1, which is lower than the predetermined lower limit value (0.5) of the confidence level of the associated organ set in advance. Fig.11 Step S203).

[0158] Next, the control unit 103 determines that there is no related organ ( Fig.11 If step S300 is yes), the processing ends.

[0159] According to the description of this embodiment, in the fourth embodiment of the present invention, the user only specifies the colon as the target organ, and the analysis device 100C automatically performs metastatic cancer analysis on the colon, the liver, the liver, the lung, and the lymph nodes of the lung. Therefore, instead of running the corresponding fully automatic lesion detection and segmentation algorithm on all organs (including organs without lesions), the fully automatic lesion detection and segmentation algorithm is only run on the associated organs, which avoids extra useless calculations, avoids the accumulation of errors, and reduces false positives; in addition, it reduces the doctor's manual reconfirmation of these false positives and reduces the doctor's workload; in addition, it does not need to manually confirm the associated organs as in the prior art, avoiding dependence on the doctor's experience and reducing misdiagnosis and missed diagnosis caused by manual confirmation.

[0160] (Variation Example)

[0161] The above embodiments are described, but the present invention is not limited to the above embodiments and variations. For example, those skilled in the art may appropriately add, delete, or change the design of the structural elements of the above embodiments and variations, or appropriately combine the features of the embodiments, and all of these are within the scope of the present invention as long as they conform to the technical concept of the present invention.

[0162] (Variation 1)

[0163] As described above, in the present invention, an example of performing lesion detection on a certain organ based on a trained deep learning model and obtaining the associated organ information corresponding to the lesion is illustrated, but it is not limited to this. For example, it is also possible not to perform lesion detection on an organ, and train the deep model for a certain organ to obtain the associated organs and the associated organ confidence of the organ.

[0164] Furthermore, the above two deep learning models can also be comprehensively applied in the analysis device to determine the associated organs and the confidence of the associated organs, thereby further improving the accuracy of metastatic cancer analysis.

[0165] (Variation 2)

[0166] As described above, in the present invention, an example is exemplified in which the analysis device determines the target organ image range based on the user's input, but the present invention is not limited thereto.

[0167] For example, the target organ may be determined by an automatic method in which the analysis device automatically selects the range of the initial target organ based on the basic information of the medical image without user input.

[0168] Here, the basic information of medical images includes Study Description, Series Description, Convolution Kernel, etc.

[0169] For example, when the basic information of a medical image is as follows,

[0170] Study Description:CT CHEST ABDOMEN PELVI

[0171] Series Description: Lung + N14 A30

[0172] Convolution Kernel: Lung

[0173] Others

[0174] The analyzing device can select the lung as the initial target organ based on the string information "lung" extracted from the Series Description and the Convolution Kernel.

[0175] In addition, when the basic information of the medical image is as follows,

[0176] Study Description:Abdomen^2_1__AbdMultiPhaseCE(Adult)

[0177] Series Description:Venous Phase 1.0B30f

[0178] Body Part Examined:ABDOMEN

[0179] Others

[0180] As an analysis device, it is not possible to obtain the initial target organ from the character string information alone. In this case, the liver can be automatically selected as the initial target organ based on the extracted character string "ABDOMEN" and the user configuration / software configuration.

[0181] In this way, the analysis device can automatically perform rapid, accurate, and reliable metastatic cancer analysis on the medical image of the subject without requiring user input.

[0182] (Variant 3)

[0183] As described above, in the fourth embodiment of the present invention, an example is shown in which the analysis device automatically performs metastatic cancer analysis on a plurality of organs and their related organs through a plurality of cycles, but the present invention is not limited to this.

[0184] For example, based on the multiple associated organs of the target organ and their respective confidence levels, instead of processing one associated organ as a new target organ each time, the multiple associated organs of the target organ can be directly analyzed using existing analysis modules corresponding to the multiple associated organs.

[0185] In this way, instead of analyzing metastatic cancer for all organs as in the prior art, irrelevant organs are excluded based on target organs, associated organs and their confidence, and only the selected associated organs are analyzed based on the existing analysis module. Therefore, instead of running the corresponding fully automatic lesion detection and segmentation algorithm on all organs (including organs without lesions), the accumulation of errors can be reduced, the number of false positives can be reduced, and the need for doctors to manually reconfirm these false positive lesions is reduced, thus reducing the workload of doctors and other users.

[0186] (Variant 4)

[0187] As described above, an example is illustrated in which the analysis device uses the examination images of the subject and the corresponding doctor's diagnosis information, multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information, etc. as training data (Table 1) to train a deep learning model, but it is not limited to this.

[0188] (Table 1)

[0189] Lesion ID Location Information Related organ information Lesion 1 (100,84,411)–(123,122,434) Liver, lung Lesion 2 (40,45,450)–(55,72,465) Liver, lung Lesion 3 … … Lesion 4 … …

[0190] For example, the examination images of the subject and the corresponding doctor's diagnosis information, multiple follow-up examination images of the subject and the corresponding doctor's diagnosis information, and additional information including biological information and multimodal information can also be used as training data (Table 2) to train the deep learning model.

[0191] (Table 2)

[0192]

[0193] Therefore, by introducing additional information, a deep learning model that is closer to reality can be trained, which helps to improve the analysis accuracy of the analysis device.

[0194] In addition, as described above, the example of determining the associated organs by considering the lesions etc. is mainly for organs, but it is not limited to this. For example, the technical solution of the present invention can determine the associated organs based on the associated information of the anatomical structure (lymph nodes etc.). Thus, it can help improve the analysis accuracy of the analysis device.

[0195] In addition, the numerical range involved in the present invention is only an example and can be appropriately set or changed according to actual needs.

[0196] In addition, the present invention describes in detail the specific manifestations of the analysis device and the analysis method it performs, but the specific manifestations of the present invention are not limited to this, and can also be expressed in various forms including a medical image processing device, an integrated circuit, a program, and a medium recording the program.

[0197] For example, the medical image processing apparatus according to the present invention includes an image acquisition device for acquiring a medical image of a subject including at least a target organ, and the above-mentioned analysis device.

[0198] The image acquisition device here can be a medical imaging device such as an ultrasonic diagnostic device, a CT (computed tomography) imaging device, or an MR (magnetic resonance) imaging device. It can also be a device with a communication function that can communicate with medical imaging devices such as an ultrasonic diagnostic device, a CT (computed tomography) imaging device, or an MR (magnetic resonance) imaging device to acquire medical images. It can also be a device that acquires medical images by reading a storage device that pre-stores medical images of target organs.

Claims

1. An analysis device, characterized in that: include: A target organ analysis unit, which performs analysis based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence level of the associated organ associated with the target organ; as well as The associated organ identifying unit identifies the associated organ based on the associated organ information and according to the confidence level of the associated organ.

2. An analysis device, characterized in that: include: A target organ analysis unit, which performs analysis based on the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence level of the associated organ associated with the target organ; an associated organ determining unit that determines an associated organ based on the associated organ information and according to the confidence level of the associated organ; and The control unit controls and repeatedly performs the following processing until it is determined by the associated organ determination unit that there is no associated organ, and the processing refers to taking the associated organ determined by the associated organ determination unit as a new target organ, analyzing it by the target organ analysis unit, and then determining the next new associated organ by the associated organ determination unit.

3. The analysis device according to claim 1 or 2, characterized in that The target organ analysis unit comprises: A target organ image range determination unit, for determining the target organ image range of the medical image of the target organ in an automatic manner or in a manual manner by a user; The target organ lesion detection unit detects all lesions of the target organ within the target organ image range, and acquires lesion position information of each lesion and the associated organ information corresponding to each lesion for each lesion.

4. The analysis device according to claim 3, characterized in that The target organ analysis unit further includes a lesion analysis unit, which analyzes each lesion of all the lesions based on the lesion position information to obtain lesion information of each lesion.

5. The analysis device according to claim 3, characterized in that The target organ lesion detection unit detects all lesions of the target organ based on the first deep learning model, and obtains the lesion location information of each lesion and the associated organ information corresponding to each lesion for each lesion.

6. The analysis device according to claim 5, characterized in that The first deep learning model is trained using the first training data to: for each organ, detect all lesions of the target organ, and for each lesion of all the lesions, obtain the lesion location information of each lesion and the associated organ information corresponding to each lesion.

7. The analysis device according to claim 6, characterized in that The first training data includes an examination image of the subject and corresponding doctor's diagnosis information, and multiple follow-up examination images of the subject and corresponding doctor's diagnosis information.

8. The analysis device according to claim 6, characterized in that The first training data includes examination images of the subject and corresponding doctor's diagnosis information, multiple follow-up examination images of the subject and corresponding doctor's diagnosis information, and additional information including biological information and multimodal information.

9. The analysis device according to claim 1 or 2, characterized in that: The associated organ determination unit includes: a candidate associated organ determination unit, which determines all candidate associated organs based on the associated organ information; A confidence determination unit, for all the candidate associated organs, determines the confidence of each of the candidate associated organs; The determination unit determines, as a related organ, the candidate related organ having the highest confidence level or the candidate related organ having the confidence level being equal to or greater than a predetermined threshold.

10. The analysis device according to claim 1 or 2, characterized in that The target organ analysis unit obtains associated organ information including at least an associated organ name and a confidence level of the associated organ associated with the target organ based on the second deep learning model.

11. The analysis device according to claim 10, characterized in that The second deep learning model is trained to output, for an organ, an associated organ of the organ and a confidence level of the associated organ.

12. An analysis method, characterized in that include: a target organ analysis step, analyzing the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence level of the associated organ associated with the target organ; as well as The associated organ determination step is to determine the associated organ based on the associated organ information and according to the confidence level of the associated organ.

13. An analysis method, characterized in that include: a target organ analysis step, analyzing the medical image of the target organ to obtain associated organ information including at least the name of the associated organ and the confidence level of the associated organ associated with the target organ; an associated organ determining step, determining an associated organ based on the associated organ information and according to the confidence level of the associated organ; and The control step controls and repeatedly performs the following processing until it is determined that there is no associated organ through the associated organ determination step, wherein the processing refers to taking the associated organ determined by the associated organ determination step as a new target organ, analyzing it through the target organ analysis step, and then determining the next new associated organ through the associated organ determination step.

14. A medical image processing device comprising: An image acquisition device for acquiring a medical image of the subject including at least a target organ; and The analyzing device according to any one of claims 1 to 11.

Images