Subtype prediction program, prediction method, and prediction device for uterine fibroids

Abstract

To non-invasively predict a subtype and tissue constitution of hysteromyoma.SOLUTION: A hysteromyoma subtype prediction program (this program) according to the present invention predicts whether or not a subtype of hysteromyoma is a specific subtype having an MED 12 mutation on the basis of an MRI image obtained by imaging the hysteromyoma. The MRI image is captured under a specific imaging condition that can evaluate a degree of fibrosis in a human tissue. This program causes at least one processor to function as: an image acquisition unit 131 which acquires the MRI image; a region extraction unit 133 which extracts a region where the hysteromyoma is imaged as a region of interest in the acquired MRI image; a signal intensity acquisition unit 134 which acquires the signal intensity on the basis of pixels constituting the region of interest in the region of interest; and a prediction unit 135 which predicts the presence / absence of the specific subtype and the tissue constitution of the hysteromyoma imaged in the region of interest on the basis of the signal intensity.SELECTED DRAWING: Figure 2

Description

Technical Field

[0001] The present invention relates to a uterine fibroids subtype prediction program, a uterine fibroids subtype prediction method, and a uterine fibroids subtype prediction device.

Background Art

[0002] Uterine fibroids are estrogen-dependent benign tumors with a high incidence (about 30%) in mature women. Uterine fibroids cause severe dysmenorrhea, anemia, infertility, and miscarriage, and are diseases that cannot be ignored. Currently, the only radical treatment for uterine fibroids is surgical procedures (invasive methods) such as hysterectomy and myomectomy. However, in recent years, as the trend of late marriage and advanced maternal age progresses, a treatment method that can preserve the uterus while being able to conceive has been desired.

[0003] As a treatment method that can preserve the uterus, research on drug therapy has been conducted. Administration of gonadotropin-releasing hormone analog (GnRHa) is performed as an effective drug therapy (non-invasive method), but GnRHa is accompanied by side effects such as osteoporosis due to a decrease in blood estrogen concentration. Therefore, long-term administration of GnRHa is not possible. On the other hand, in recent years, selective progesterone receptor modulators (SPRM), which act on progesterone receptors (PGR) but are expected not to affect blood estrogen concentration, have attracted attention as candidates for drugs with few side effects. However, there are differences in the reduction rate of uterine fibroids in patients administered SPRM, and it is known that there are patients for whom SPRM is effective and patients for whom it is not effective. The inventors of the present invention predict that the difference in the effect of SPRM is related to the tissue composition in uterine fibroid tumors.

[0004] Uterine fibroids have several subtypes with different development pathways and causes (mutations). Among these subtypes, uterine fibroids with MED12 (Mediator complex subunit 12) mutations account for approximately 70% of uterine fibroids, and the MED12 mutation is considered one of the driver mutations in the development of uterine fibroids. In recent years, it has been reported that the tissue composition (such as cell composition and components such as the extracellular matrix (ECM)) in uterine fibroid tumors differs depending on the presence or absence of the MED12 mutation (see, for example, Non-Patent Document 1). Specifically, in uterine fibroids with MED12 mutations (hereinafter referred to as "MED12(+) fibroids"), approximately 50% of the cell composition is composed of smooth muscle cells, and the remaining approximately 50% is composed of fibroblasts. On the other hand, in uterine fibroids without MED12 mutations (hereinafter referred to as "MED12(-) fibroids"), most of the cell composition is composed of smooth muscle cells. Also, the ECM content is high in MED12(+) fibroids and low in MED12(-) fibroids. That is, the tissue composition of uterine fibroids varies greatly depending on the presence or absence of the MED12 mutation.

[0005] Under such circumstances, the inventors of the present invention have clarified through their own research that the expression of PGR, which is the target of SPRM, is limited to smooth muscle cells. From this, it is possible that the sensitivity to SPRM differs among fibroid subtypes with different cell compositions. Also, GnRHa and SPRM do not have the action of directly decomposing the ECM. Therefore, even if the smooth muscle cells and fibroblasts in the tumor are reduced, the collagen fibers remain, and the reduction of the tumor volume is limited. That is, depending on the amount of ECM content in the tumor, it is necessary to select and determine the treatment policy and the drugs to be used. Therefore, if the subtype and tissue composition of the uterine fibroids developing in a patient can be predicted in advance, the prediction results will be useful for determining the treatment policy and the drugs to be used.

Prior Art Documents

Non-Patent Documents

[0006]

Non-Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0007] An object of the present invention is to non-invasively predict the subtype and tissue composition of uterine leiomyoma.

Means for Solving the Problems

[0008] The subtype prediction program for uterine leiomyoma according to the present invention is executed by a subtype prediction device for uterine leiomyoma that predicts whether a uterine leiomyoma is a specific subtype having a MED12 (Mediator complex subunit 12) mutation and its tissue composition based on an MRI image of the uterine leiomyoma. The MRI image is taken under specific imaging conditions that enable evaluation of the fibrosis of the uterine leiomyoma. The subtype prediction device for uterine leiomyoma includes at least one processor. The at least one processor is functioned as an image acquisition unit that acquires the MRI image, a region extraction unit that extracts, as a region of interest, the region where the uterine leiomyoma is imaged in the acquired MRI image, a signal intensity acquisition unit that acquires a signal intensity corresponding to the specific imaging conditions based on the pixels constituting the region of interest in the region of interest, and a prediction unit that predicts the presence or absence of a specific subtype in the uterine leiomyoma imaged in the region of interest and the tissue composition of the uterine leiomyoma imaged in the region of interest based on the signal intensity.

Effects of the Invention

[0009] According to the present invention, the subtype and tissue composition of uterine leiomyoma can be non-invasively predicted.

Brief Description of the Drawings

[0010]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Embodiments for Carrying Out the Invention

[0011] Hereinafter, embodiments of a uterine myoma subtype prediction program (hereinafter referred to as "this program"), a prediction method (hereinafter referred to as "this method"), and a prediction device (hereinafter referred to as "this device") according to the present invention will be described with reference to the drawings.

[0012] The present invention predicts whether a uterine myoma being imaged is a uterine myoma with a MED12 (Mediator complex subunit 12) mutation (hereinafter referred to as "MED12(+) myoma") or a uterine myoma without it (hereinafter referred to as "MED12(-) myoma"), and its tissue composition, based on the signal intensity obtained from an MRI (Magnetic Resonance Imaging) image in which the uterine myoma is imaged. Here, MED12(+) myoma is an example of a specific subtype in the present invention.

[0013] "Signal intensity" is information assigned to each pixel of an MRI image, and is the intensity of a signal corresponding to the proton density, T1 value, T2 value, etc. of a biological tissue. Usually, in an MRI image, pixel information corresponding to the signal intensity (for example, color information such as luminance) is assigned to each pixel (corresponding relationship information described later). The signal intensity varies depending on the imaging conditions of the MRI image.

[0014] "Imaging conditions" mean the imaging conditions, techniques, and programs of MRI called sequences. That is, for example, the imaging conditions are MRI sequences such as T2WI (T2 Weighted Image), ADC (Apparent Diffusion Coefficient), T1map (T1 mapping), T2*BOLD (T2 Blood Oxygenation Level Dependent), MTC (Magnetization Transfer Contrast), MRE (Magnetic Resonance Elastography), ASL (Arterial Spin Labeling).

[0015] "MED12 mutation" is a somatic mutation in the gene encoding MED12, which is one of the proteins that make up the RNA polymerase II Mediator complex, and is a driver mutation that is frequently detected (about 70%) in uterine fibroids. As described above, in the MED12(+) mutation, the cell composition is composed of smooth muscle cells and fibroblasts at approximately the same ratio. On the other hand, in the MED12(-) mutation, most of the cell composition is composed of smooth muscle cells.

[0016] Figure 1 is a network configuration diagram showing an embodiment of the present device. This figure shows that the present device 1 is connected to an external device 2 via a network N.

[0017] Based on the MRI image of the uterine fibroid taken, the present device 1 predicts whether the uterine fibroid imaged in the MRI image is a MED12(+) fibroid, a MED12(-) fibroid, and its tissue composition.

[0018] The external device 2 is, for example, an information storage device such as a server that stores information (such as MRI images, etc.) necessary for the operation of the present device 1.

[0019] In the present invention, the external device may also be an MRI device that takes MRI images.

[0020] The network N is, for example, a communication network such as the Internet, a mobile communication network, a LAN (Local Area Network), a WAN (Wide Area Network), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0021] ●Uterine Fibroid Subtype Prediction Device● ●Configuration of Uterine Fibroid Subtype Prediction Device First, the present device 1 in which this program is executed will be described. Figure 2 is a functional block diagram of the present device 1.

[0022] The present device 1 is implemented by, for example, a personal computer. In the present device 1, the present program operates, and the present program cooperates with the hardware resources of the present device 1 to implement the present method.

[0023] Here, by causing a computer (not shown) to execute the present program, the present program can cause the same computer to function in the same manner as the present device 1 and cause the same computer to execute the present method.

[0024] The present device 1 includes a communication unit 11, a storage unit 12, a control unit 13, an operation unit 14, and a display unit 15.

[0025] The communication unit 11 is connected to an external device 2 via a network N. The communication unit 11 is constituted by, for example, a communication module or a communication interface.

[0026] The storage unit 12 stores information (for example, MRI images, correspondence information described later, correlation information described later, etc.) necessary for the present device 1 to execute the present method described later. The storage unit 12 is constituted by, for example, a storage medium such as an HDD (Hard Disk Drive), an SSD (Solid State Drive), or a flash memory provided in the present device 1.

[0027] The control unit 13 controls the overall operation of the present device 1 and executes the present method described later. The control unit 13 is constituted by, for example, a CPU (Central Processing Unit) provided in the present device 1, a RAM (Random Access Memory) that functions as a working area of the CPU, and a ROM (Read Only Memory) that stores various information such as the present program. The control unit 13 is an example of a processor in the present invention. The control unit 13 includes an image acquisition unit 131, a condition specification unit 132, a region extraction unit 133, a signal intensity acquisition unit 134, a prediction unit 135, and an output unit 136.

[0028] The image acquisition unit 131 acquires an MRI image in which a uterine myoma of a patient is photographed. The specific operation of the image acquisition unit 131 will be described later.

[0029] The condition specifying unit 132 specifies the imaging conditions of the MRI image based on the MRI image. The specific operation of the condition specifying unit 132 will be described later.

[0030] The region extraction unit 133 extracts, as a region of interest, the region in the MRI image acquired by the image acquisition unit 131 where the uterine myoma is photographed. The specific operation of the region extraction unit 133 will be described later.

[0031] The "region of interest" is the region in the MRI image occupied by the uterine myoma that is the prediction target of the subtype.

[0032] The signal intensity acquisition unit 134 acquires, in the region of interest, the signal intensity corresponding to the specific imaging conditions described later based on the pixels constituting the region of interest. The specific operation of the signal intensity acquisition unit 134 will be described later.

[0033] The prediction unit 135 predicts, based on the signal intensity, whether the uterine myoma photographed in the region of interest is a MED12(+) myoma or a MED12(-) myoma, and the tissue composition of the uterine myoma photographed in the region of interest. The specific operation of the prediction unit 135 will be described later.

[0034] The output unit 136 outputs various information (for example, information indicating the prediction result, etc.) obtained by executing the method described later. The specific operation of the output unit 136 will be described later.

[0035] The operation unit 14 is a device that is operated (for example, information input / selection operation, etc.) by the user of the present apparatus 1 (for example, medical staff such as doctors and nurses). The operation unit 14 is, for example, a keyboard, a mouse, or a touch panel.

[0036] The display unit 15 is a device that displays the information output by the control unit 13. The display unit 15 is, for example, a monitor or a display.

[0037] Note that the operation unit and the display unit in this device may be configured by, for example, a touch panel type display.

[0038] ● Subtype prediction method for uterine fibroids Next, the method executed by this device 1 will be described. In the following description of this method, FIG. 2 is also referred to.

[0039] FIG. 3 is a flowchart showing an embodiment of this method.

[0040] First, the image acquisition unit 131 acquires an MRI image from the external device 2 via the communication unit 11, for example (S1: Image acquisition step). Here, the MRI image acquired by the image acquisition unit 131 is taken under a specific imaging condition (hereinafter referred to as "specific imaging condition") selected from a plurality of imaging conditions capable of evaluating the degree of fibrosis in human tissues. In this embodiment, the specific imaging condition is, for example, T2WI.

[0041] "Imaging conditions capable of evaluating the degree of fibrosis in human tissues" include not only imaging conditions that are documented to be capable of evaluating the degree of fibrosis in human tissues (for example, the amount of collagen fibers), but also imaging conditions that may be useful for such evaluation (for example, imaging conditions indirectly capable of evaluating the degree of fibrosis), and imaging conditions already clinically used for fibrosis evaluation. That is, for example, the specific imaging condition can be selected from among imaging conditions that include not only the imaging conditions mainly selected when taking MRI images in obstetrics and gynecology, but also the imaging conditions mainly selected when taking MRI images in other medical departments (for example, neurosurgery, etc.). In the present invention, the specific imaging condition is preferably any one of, for example, T2WI, ADC, T1map, T2*BOLD, MTC, MRE, ASL, more preferably any one of T2WI, ADC, T1map, T2*BOLD, and even more preferably any one of T2WI, T1map, T2*BOLD.

[0042] In the present invention, when an MRI image is stored in the storage unit in advance, the image acquisition unit may acquire the MRI image from the storage unit. Further, the image acquisition unit may acquire an MRI image selected by the user of the present apparatus via the operation unit. In this case, for example, a selection screen for the MRI image is output to the display unit by the output unit.

[0043] Next, the condition specifying unit 132 specifies specific imaging conditions under which the MRI image was taken based on the MRI image (S2). That is, for example, the condition specifying unit 132 specifies the specific imaging conditions based on information indicating the imaging conditions (imaging condition information) included in the pixel information of the MRI image or information associated with the MRI image.

[0044] In the present invention, the condition specifying unit may acquire the specific imaging conditions based on information input by the user of the present apparatus via the operation unit, for example, or may acquire the specific imaging conditions set in advance in the present program. Further, when the imaging condition information of the specific imaging conditions is associated with the MRI image in advance, the condition specifying unit may specify the specific imaging conditions based on the imaging condition information.

[0045] Next, the region extraction unit 133 extracts, as a region of interest, one closed region in the MRI image in which the uterine myoma is imaged (S3: region extraction step). Specifically, for example, the region extraction unit 133 extracts, as the region of interest, the region selected by the user of the present apparatus via the operation unit with respect to the MRI image displayed on the display unit 15. In this case, for example, the MRI image is output to the display unit 15 by the output unit 136.

[0046] FIG. 4 is an image diagram showing an example of extraction of a region of interest, where (a) shows the MRI image before extraction of the region of interest and (b) shows the MRI image after extraction of the region of interest. In the figure, the region surrounded by the dashed line is the extracted region of interest.

[0047] In the present invention, the method for extracting the region of interest by the region extraction unit is not limited to this embodiment. That is, for example, the region extraction unit may extract the region of interest by image recognition using a learned machine learning model.

[0048] Return to FIGS. 2 and 3. Next, the signal intensity acquisition unit 134 reads out the correspondence information corresponding to the specific imaging condition from the storage unit 12 based on the imaging condition information corresponding to the specific imaging condition (S4).

[0049] The "correspondence information" is information indicating the correspondence between the pixel information of the MRI image and the signal intensity for each specific imaging condition. That is, for example, in the correspondence information, a signal intensity value is assigned to each pixel information. The correspondence information is, for example, set in advance for each specific imaging condition, associated with the imaging condition information of the specific imaging condition, and stored in the storage unit 12.

[0050] Next, the signal intensity acquisition unit 134 acquires the signal intensity for each pixel constituting the region of interest based on the correspondence information (S5: signal intensity acquisition step). Specifically, the signal intensity acquisition unit 134 acquires pixel information for each pixel constituting the region of interest, and acquires the signal intensity assigned to each pixel information by referring to the correspondence information.

[0051] Next, the signal intensity acquisition unit 134 calculates the average value "Va" of the acquired signal intensities (S6).

[0052] Next, the prediction unit 135 compares the calculated average value "Va" with the threshold value "V1" (S7).

[0053] The "threshold value 'V1'" is a threshold value that indicates the boundary for distinguishing whether the average value "Va" of the signal intensity in the region of interest belongs to the value indicating MED12(+) myoma or the value indicating MED12(-) myoma. That is, the threshold value "V1" is the boundary value for determining whether the uterine myoma imaged in the MRI image is MED12(+) myoma or MED12(-) myoma. In the present embodiment, the threshold value "V1" is calculated based on the signal intensities of a plurality of MRI images in which uterine myomas with the presence or absence of MED12 mutation determined in advance by an examination using a specimen are imaged for each specific imaging condition, for example, and is stored in the storage unit 12 in association with the corresponding imaging condition information.

[0054] In the present invention, the threshold value "V1" may be set by machine learning based on the signal intensity or its average value in the region of interest and the analysis result of MED12 mutation using a specimen.

[0055] When the average value "Va" belongs to the value indicating MED12(+) myoma ("MED12(+)" in S7), the prediction unit 135 predicts that the uterine myoma imaged in the region of interest is MED12(+) myoma (S8: prediction step). That is, for example, in the present embodiment (the specific imaging condition is T2WI), when the average value "Va" is less than the threshold value "V1", the prediction unit 135 predicts that the uterine myoma imaged in the region of interest is MED12(+) myoma.

[0056] On the other hand, when the average value "Va" belongs to the value indicating MED12(-) myoma ("MED12(-)" in S7), the prediction unit 135 predicts that the uterine myoma imaged in the region of interest is MED12(-) myoma (S9: prediction step). That is, for example, in the present embodiment (the specific imaging condition is T2WI), when the average value "Va" is equal to or greater than the threshold value "V1", the prediction unit 135 predicts that the uterine myoma imaged in the region of interest is MED12(-) myoma.

[0057] In the present invention, the prediction unit may predict the presence or absence of the MED12 mutation, for example, by machine learning based on the signal intensity of the region of interest or its average value and the analysis result of the MED12 mutation using the specimen.

[0058] Next, the prediction unit 135 reads out the correlation information corresponding to the specific imaging conditions from the storage unit 12 based on the imaging condition information corresponding to the specific imaging conditions (S10).

[0059] The "correlation information" is, for example, information indicating a correlation function (e.g., a linear function) between the average value "Va" of the signal intensity and the ratio of collagen fibers for each specific imaging condition. The correlation information is determined in advance for each specific imaging condition and is stored in the storage unit 12 in association with the corresponding imaging condition information.

[0060] In the present invention, the correlation information may be constructed, for example, by machine learning based on the signal intensity of the region of interest and the amount of collagen fibers actually measured using the specimen.

[0061] Next, the prediction unit 135 predicts the tissue composition (ratio of collagen fibers) of the uterine myoma imaged in the region of interest based on the correlation information and the average value "Va" of the signal intensity (S11). The prediction results (presence or absence of the MED12 mutation, ratio of collagen fibers) are associated with, for example, the identification information of the patient (e.g., a patient-specific identification ID set for each patient) and the MRI image and are stored in the storage unit 12.

[0062] Next, the output unit 136 outputs the prediction result of the prediction unit 135 to the display unit 15 (S12).

[0063] Next, the display unit 15 displays the prediction result of the prediction unit 135 (S13).

[0064] FIG. 5 is a schematic diagram showing an example of the prediction result displayed on the display unit 15. The figure shows that the subtype of uterine myoma "No.B" of patient "A" is "MED12(+) myoma", the proportion of collagen fibers is "C%", and the proportion of smooth muscle cells is "D%", which is displayed on the display unit 15 as the prediction result.

[0065] Thus, based on the MRI image in which the uterine myoma is photographed, the apparatus 1 can non-invasively predict whether the uterine myoma is a specific subtype (MED12(+) myoma). In addition, based on the MRI image in which the uterine myoma is photographed, the apparatus 1 can non-invasively predict the approximate proportion of collagen fibers (i.e., tissue composition) in the uterine myoma. As a result, the user of the apparatus 1 can determine the treatment policy and drugs to be used for the uterine myoma based on the presence or absence of MED12 mutation.

[0066] Among the therapeutic drugs for uterine myoma, selective progesterone receptor modulators (SPRM) have an antagonistic effect on progesterone receptors, and gonadotropin-releasing hormone analogs (GnRHa) have an effect of suppressing the secretion of estrogen and progesterone. And MED12(+) myomas contain approximately 50% of fibroblasts that proliferate with estrogen only, and most of MED12(-) myomas are composed of smooth muscle cells that proliferate in the presence of both estrogen and progesterone. Therefore, GnRHa is effective against both MED12(+) myomas and MED12(-) myomas, and is particularly suitable for MED12(+) myomas that contain approximately 50% of fibroblasts that proliferate with estrogen only. On the other hand, since SPRM inhibits only progesterone receptors, it is a drug suitable for MED12(-) myomas that are mostly composed of smooth muscle cells.

[0067] Next, as treatment guidelines for uterine fibroids, the perspectives of tumor shrinkage effect and the possibility of long-term administration can be mentioned. The tumor shrinkage effect in drug therapy is higher in MED12(-) fibroids with a low collagen fiber content than in MED12(+) fibroids. In addition, since GnRHa has side effects such as osteoporosis due to decreased estrogen secretion, there is a limitation that it cannot be administered long-term. On the other hand, since SPRM has no effect on estrogen secretion, SPRM can be administered long-term. Thus, the mechanisms of action of GnRHa and SPRM are different, and it is predicted that the treatment effects will differ depending on the presence or absence of MED12 mutations. Therefore, by predicting the presence or absence of MED12 mutations with the present apparatus 1, it becomes possible to predict the treatment effect, and an appropriate treatment guideline and drug to be used for the patient can be selected.

[0068] Note that multiple uterine fibroids can exist in one patient. Therefore, in the present invention, the above-described processing is performed for each uterine fibroid.

[0069] ●Examples● Next, examples of the present invention will be described. In the following examples, 45 specimens of uterine fibroids were used. The base sequences were analyzed for the DNA extracted from each specimen, and the presence or absence of MED12 mutations was determined. As a result, there were 34 specimens of MED12(+) fibroids and 11 specimens of MED12(-) fibroids.

[0070] Also, in the following examples, among the imaging conditions considered to be effective for the quantitative evaluation of the fibrosis (amount of collagen fibers) of uterine fibroids, four conditions (T2WI, ADC, T1map, T2*BOLD) were used. Since the T2 value of T2WI is a relative value among these imaging conditions, the ratio of the T2 value of the uterine fibroid to the value of the skeletal muscle in the same MRI image (T2 value of the uterine fibroid / T2 value of the skeletal muscle) was treated as the quantitative value of T2WI.

[0071] First, for each imaging condition, MRI images of 45 uterine fibroids were obtained. Next, the region of interest was extracted from each MRI image, and the average value "Va" of its signal intensity was calculated. Here, the value and unit of the average value "Va" of the signal intensity differ for each imaging condition. Therefore, in the following description, the average value "Va" of the signal intensity in each figure is represented by the name of each imaging condition and its unit.

[0072] Figure 6 is an image diagram showing a comparison of examples of MRI images of MED12(+) fibroids and MED12(-) fibroids for each imaging condition.

[0073] Figure 7 is a graph showing the comparison results of the distributions of the average values "Va" of the signal intensities of MRI images of MED12(+) fibroids and MED12(-) fibroids for each imaging condition. This figure shows box plots of the average values "Va" of the signal intensities of MED12(+) fibroids and MED12(-) fibroids, respectively.

[0074] As shown in Figure 7, significant differences were observed in the values of the average value "Va" of the signal intensity in each of the four imaging conditions (T2WI, ADC, T1map, T2*BOLD). Therefore, these four imaging conditions were selected as specific imaging conditions. Next, for the average value "Va" of the signal intensity of each of the four specific imaging conditions, a "Student and Welch's t-test" was performed between two groups with or without MED12 mutation. As a result, the probability "P" of three specific imaging conditions (T2WI, T1map, T2*BOLD) is "P < 0.01", and the probability "P" of one specific imaging condition (ADC) is "P < 0.05". Therefore, there are significant differences between the two groups in each of the four specific imaging conditions.

[0075] Next, an ROC curve was created for each specific imaging condition, and based on the ROC curve, the cut-off value at which the sensitivity and specificity for the detection of MED12(-) fibroids are the highest was calculated as the threshold "V1".

[0076] Figure 8 is an ROC curve diagram for each specific imaging condition. FIG. 9 shows the threshold value "V1" for each specific imaging condition calculated based on the ROC curve.

[0077] As shown in FIGS. 8 and 9, good sensitivity (0.8 or higher) and specificity (0.7 or higher) were obtained under each specific imaging condition. Thus, the present invention can accurately predict the presence or absence of MED12 mutation based on the MRI images taken under each specific imaging condition.

[0078] Note that the value of each threshold "V1" varies depending on the number of specimens and is not limited to the value shown in FIG. 9.

[0079] Next, αSMA staining was performed on each specimen, and the ratio of smooth muscle cells to the total number of cells was quantitatively calculated. Also, Masson's trichrome staining was performed on each specimen, and the area of collagen fibers with respect to the total area was quantitatively calculated as the ratio of collagen fibers.

[0080] FIG. 10(a) is a graph showing the ratio of smooth muscle cells in the presence or absence of MED12 mutation, and (b) is a graph showing the ratio of collagen fibers in the presence or absence of MED12 mutation. This figure shows box plots of smooth muscle cells and collagen fibers in MED12(+) myomas and MED12(-) myomas, respectively. Also, for comparison, this figure also shows the ratios of smooth muscle cells and collagen fibers in normal myometrium (denoted as MM). As shown in FIG. 10, in MED12(-) myomas, the ratio of smooth muscle cells is high (about 85%, and the remaining 15% is occupied by fibroblasts), and the ratio of collagen fibers is low (about 25%). On the other hand, in MED12(+) myomas, the ratio of smooth muscle cells and fibroblasts is closer (the ratio of smooth muscle cells is about 50 - 60%), and the ratio of collagen fibers is higher than that in MED12(-) myomas (about 50%).

[0081] Next, for each specific imaging condition, the relationship between the average value "Va" of the signal intensity and the ratio of collagen fibers was determined.

[0082] FIG. 11 is a graph showing the relationship (correlation) between the average value “Va” of the signal intensity and the ratio of collagen fibers for each specific imaging condition. The horizontal axis of this figure indicates the ratio of collagen fibers, and the vertical axis indicates the average value “Va” of the signal intensity. As shown in FIG. 11, a correlation was found between the average value “Va” of the signal intensity and the ratio of collagen fibers under three specific imaging conditions (T2WI, T1map, T2*BOLD). The linear function in the previous embodiment is determined based on this correlation, for example. The apparatus 1 can predict the approximate ratio (i.e., tissue composition) of collagen fibers in uterine fibroids based on the MRI images taken under each specific imaging condition.

[0083] ● Summary According to the embodiments described above, this program causes the apparatus 1 (control unit 13) to function as an image acquisition unit 131, a region extraction unit 133, a signal intensity acquisition unit 134, and a prediction unit 135. The MRI images are taken under specific imaging conditions that enable evaluation of the degree of fibrosis in human tissues. The region extraction unit 133 extracts, as a region of interest, the region in the MRI image where uterine fibroids are imaged. The signal intensity acquisition unit 134 acquires the signal intensity corresponding to the specific imaging condition based on the pixels constituting the region of interest. The prediction unit 135 predicts the presence or absence of MED12 mutations in the uterine fibroids imaged in the region of interest and the tissue composition of the uterine fibroids imaged in the region of interest based on the signal intensity. According to this configuration, this program (apparatus 1) can predict whether the target uterine fibroids are MED12(+) fibroids based on the MRI images. Also, as described above, the tissue composition of uterine fibroids varies greatly depending on the presence or absence of MED12 mutations. Therefore, if the uterine fibroids are MED12(+) fibroids, this program (apparatus 1) can predict that their tissue composition consists of abundant collagen fibers. On the other hand, if the uterine fibroids are MED12(−) fibroids, this program (apparatus 1) can predict that their tissue composition consists of a small amount of collagen fibers. That is, the apparatus 1 can non-invasively predict the subtype and tissue composition of uterine fibroids. Also, the same effect can be obtained for this method.

[0084] Also, according to the embodiments described above, the specific imaging condition is any one of T2WI, ADC, T1map, and T2*BOLD. According to this configuration, the present apparatus 1 can accurately predict whether the target specimen is a MED12(+) myoma and can also predict its tissue composition (the ratio of collagen fibers).

[0085] Furthermore, according to the embodiments described above, the signal intensity acquisition unit 134 calculates (acquires) the average value "Va" of the signal intensity. The prediction unit 135 predicts whether the uterine myoma is a MED12(+) myoma or a MED12(-) myoma based on the comparison result between the average value "Va" of the signal intensity and the threshold value "V1". According to this configuration, the present apparatus 1 can easily predict the presence or absence of MED12 mutation only by calculating the average value "Va" of the signal intensity in the region of interest.

[0086] Furthermore, according to the embodiments described above, the storage unit 12 stores correspondence information indicating the correspondence between the pixel information of the MRI image and the signal intensity. The signal intensity acquisition unit 134 acquires the signal intensity based on the correspondence information. According to this configuration, the signal intensity acquisition unit 134 can easily acquire the signal intensity based on the pixel information of the MRI image.

[0087] Furthermore, according to the embodiments described above, the storage unit 12 stores the correspondence information for each specific imaging condition. The present apparatus 1 includes a condition identification unit 132 that identifies the specific imaging condition based on the MRI image. According to this configuration, the present apparatus 1 can automatically identify the specific imaging condition based on the MRI image and acquire the signal intensity based on the correspondence information corresponding to the specific imaging condition.

[0088] Furthermore, according to the embodiment described above, the storage unit 12 stores correlation information indicating the correlation between the average value "Va" and the tissue composition (amount of collagen fibers) of uterine fibroids. The prediction unit 135 predicts the tissue composition of uterine fibroids based on the average value "Va" and the correlation information. According to this configuration, the apparatus 1 can easily predict the tissue composition of uterine fibroids based on the pixel information (average value "Va") of the MRI image and the correlation information.

[0089] Furthermore, according to the embodiment described above, the storage unit 12 stores correlation information for each specific imaging condition. The apparatus 1 includes a condition specifying unit 132 that specifies the specific imaging condition based on the MRI image. The apparatus 1 includes a condition specifying unit 132 that specifies the specific imaging condition based on the MRI image. According to this configuration, the apparatus 1 can automatically specify the specific imaging condition based on the MRI image and predict the tissue composition based on the correlation information corresponding to the specific imaging condition.

[0090] ● Other Embodiments Note that the apparatus may also serve as an MRI apparatus that captures MRI images. That is, for example, this program may be executed on an MRI apparatus.

[0091] Also, the apparatus may not include a condition specifying unit. In this case, for example, the signal intensity acquisition unit may read out correspondence information corresponding to a specific imaging condition input or selected in advance by the user of the apparatus from the storage unit. Also, for example, the imaging condition information and the correspondence information corresponding to the specific imaging condition may be set in this program in advance.

[0092] Furthermore, in the present invention, the storage unit may not store the correlation information. Also, the prediction unit may not predict the tissue composition based on the correlation information. In this case, the prediction unit may predict the tissue composition based on the prediction result of the presence or absence of the MED12 mutation (that is, if it is a MED12(+) fibroid, the amount of collagen fibers is abundant, and if it is a MED12(-) fibroid, the amount of collagen fibers is small, etc.).

[0093] Furthermore, the output destination of the output unit in the present invention is not limited to the display unit. That is, for example, the output unit in the present invention may output information to a printer. Also, for example, the output unit in the present invention may output information to another external device (such as a portable information processing terminal, etc.) via the communication unit.

[0094] Furthermore, the present device may also have a function of estimating (determining) an appropriate treatment policy and a drug to be used based on the prediction result of the prediction unit.

[0095] Furthermore, in the present invention, the prediction unit only needs to predict the presence or absence of the MED12 mutation, and does not necessarily need to execute the prediction of the tissue composition based on the correlation.

[0096] Furthermore, in the present invention, the specific imaging conditions are not limited to the four types of imaging conditions described in the examples. That is, for example, among the seven types of imaging conditions described in the present embodiment, even if three types of imaging conditions not described in the examples are used, through verification similar to that in the examples, they can become specific imaging conditions. Also, for example, the same applies to imaging conditions not listed in the present embodiment.

[0097] Furthermore, in the embodiment described above, this program (this device 1) predicted the presence or absence of the MED12 mutation based on the signal intensity of the MRI image taken under one type of specific imaging condition. Instead of this, this program (this device) may predict the presence or absence of the MED12 mutation based on the signal intensities of a plurality of MRI images taken under each of a plurality of specific imaging conditions. In this case, for example, the prediction unit may comprehensively determine based on machine learning based on each signal intensity to predict the presence or absence of the MED12 mutation.

[0098] Furthermore, in the embodiments described above, the apparatus 1 was configured by a single computer. Instead, the apparatus may be configured by a plurality of computers. That is, for example, the apparatus may be configured by a plurality of computer groups that function as the apparatus. Specifically, for example, the apparatus (computer group) may be configured by a computer having a storage unit and a computer having a control unit that executes this method. Also, for example, a plurality of computers may be provided with the functions of the image acquisition unit, condition identification unit, region extraction unit, signal intensity acquisition unit, prediction unit, and output unit, respectively, in a distributed manner. In this case, the plurality of computers constituting the computer group may transmit and receive information via a network, or may transfer information using a portable storage medium.

Explanation of Reference Numerals

[0099] 1 Subtype prediction apparatus for uterine fibroids 12 Storage unit 131 Image acquisition unit 132 Condition identification unit 133 Region extraction unit 134 Signal intensity acquisition unit 135 Prediction unit

Claims

1. A uterine myoma subtype prediction program that functions as a uterine myoma subtype prediction device for predicting whether the subtype of the uterine myoma is a specific subtype having a MED12 (Mediator complex subunit 12) mutation and its tissue composition based on an MRI image of the uterine myoma, wherein the MRI image is taken under specific imaging conditions that enable evaluation of the degree of fibrosis in human tissue, and the subtype prediction device includes, at least one processor, and functions at least one of the processors as, an image acquisition unit that acquires the MRI image, a region extraction unit that extracts, as a region of interest, a region in the acquired MRI image where the uterine myoma is imaged, a signal intensity acquisition unit that acquires a signal intensity corresponding to the specific imaging conditions based on the pixels constituting the region of interest in the region of interest, and a prediction unit that predicts the presence or absence of the specific subtype in the uterine myoma imaged in the region of interest and the tissue composition of the uterine myoma imaged in the region of interest based on the signal intensity. A uterine myoma subtype prediction program, characterized by the above.

2. The specific imaging conditions are any one of T2WI, ADC, T1map, and T2*BOLD. The uterine myoma subtype prediction program according to Claim 1.

3. The signal intensity acquisition unit calculates an average value of the signal intensity, and the prediction unit predicts whether the uterine myoma is the specific subtype based on a comparison result between the average value and a predetermined threshold value. The uterine myoma subtype prediction program according to Claim 1 or 2.

4. The uterine myoma subtype prediction device includes a storage unit that stores correspondence information indicating a correspondence between pixel information of the MRI image and the signal intensity. The signal intensity acquisition unit acquires the signal intensity based on the correspondence information. The uterine myoma subtype prediction program according to Claim 3.

5. The storage unit stores the correspondence information for each specific imaging condition, and at least one of the processors functions as a condition identification unit that identifies the specific imaging conditions based on the MRI image. The signal intensity acquisition unit acquires the signal intensity based on the correspondence information corresponding to the specific imaging conditions identified by the condition identification unit. ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ ​ The uterine fibroids subtype prediction program according to claim 4.

6. The uterine fibroids subtype prediction device, a storage unit that stores correlation information indicating the correlation between the average value and the tissue composition of the uterine fibroids; comprises the prediction unit predicts the tissue composition of the uterine fibroids based on the average value and the correlation information; The uterine fibroids subtype prediction program according to claim 3.

7. the storage unit stores the correlation information for each of the specific imaging conditions, at least one of the processors a condition identification unit that identifies the specific imaging conditions based on the MRI image; functions as the prediction unit predicts the tissue composition of the uterine fibroids based on the correlation information corresponding to the specific imaging conditions identified by the condition identification unit; The uterine fibroids subtype prediction program according to claim 6.

8. A method for predicting the subtype of uterine fibroids, which is executed by a device for predicting the subtype of uterine fibroids that predicts whether the uterine fibroids have a specific subtype having a MED12 (Mediator complex subunit 12) mutation based on an MRI image of the uterine fibroids, the MRI image is taken under specific imaging conditions that enable evaluation of the degree of fibrosis in human tissue, the uterine fibroids subtype prediction device an image acquisition step of acquiring the MRI image; a region extraction step of extracting, as a region of interest, the region in the acquired MRI image where the uterine fibroids are imaged; a signal intensity acquisition step of acquiring, in the region of interest, the signal intensity corresponding to the specific imaging conditions based on the pixels constituting the region of interest; a prediction step of predicting the presence or absence of the specific subtype in the uterine fibroids imaged in the region of interest and the tissue composition of the uterine fibroids imaged in the region of interest based on the signal intensity; including A method for predicting the subtype of uterine fibroids, characterized in that.

9. A device for predicting the subtype of uterine fibroids that predicts whether the subtype of the uterine fibroids has a specific subtype having a MED12 (Mediator complex subunit 12) mutation based on an MRI image of the uterine fibroids, the MRI image is taken under specific imaging conditions that enable evaluation of the degree of fibrosis in human tissue, an image acquisition unit that acquires the MRI image; In the obtained MRI image, an area extraction unit that extracts, as a region of interest, the region in which the uterine myoma is photographed; In the region of interest, a signal intensity acquisition unit that acquires a signal intensity corresponding to the specific imaging condition based on the pixels constituting the region of interest; A prediction unit that predicts the presence or absence of the specific subtype in the uterine myoma photographed in the region of interest and the tissue composition of the uterine myoma photographed in the region of interest based on the signal intensity; comprising: A uterine myoma subtype prediction device characterized by the above.

Images