Prediction method and device for complete remission of endometrial cancer patient, medium and product

Abstract

The invention discloses a prediction method and device for complete remission of an endometrial cancer patient, a medium and a product, and relates to the technical field of biologication.The method comprises the steps that feature extraction is conducted on an obtained data set, and a plurality of radiomics features are obtained; establishing an MR scoring model according to the plurality of radiomics features; performing prediction effectiveness evaluation on the MR scoring model to obtain an MR score; performing normal distribution test on the plurality of radiomics characteristics to obtain a plurality of radiomics characteristics converted into normal distribution; carrying out variance homogeneity analysis on the plurality of radiomics characteristics converted into normal distribution, and determining the difference of the plurality of radiomics characteristics converted into normal distribution and prediction factors related to complete remission of the patient; and constructing a column graph model according to the MR score and the prediction factor. According to the method, young EC patients suitable for retaining fertility can be accurately identified, the accuracy of a clinical decision-making process is improved, and unnecessary uterine cavity invasive operation is reduced.

Description

Technical Field

[0001] The present invention relates to the field of biotechnology, and in particular to a method, device, medium and product for predicting complete remission of endometrial cancer patients. Background Art

[0002] Endometrial cancer (EC) has become the most common gynecological malignancy, and its mortality rate has increased significantly. According to global cancer statistics, 80,000 new cases of EC were reported in China in 2020. An increasing number of young women are diagnosed with EC, with approximately 25% of cases occurring before menopause, of which 3% to 10% are under 40 years old at diagnosis. Atypical endometrial hyperplasia (AEH) is a non-invasive precancerous lesion of the endometrium. Over time, the risk of AEH developing into EC ranges from 8.2% to 27.5%. With the relaxation of China's fertility policy and the late childbearing caused by socioeconomic factors, up to 70% of young women diagnosed with EC still have fertility needs. Radical surgery (total hysterectomy + bilateral salpingo-oophorectomy) is often faced with the reluctance of young patients; therefore, research and development of fertility-preserving treatments is a top priority. However, clinical practice has so far faced several challenges, such as the inability to determine which treatment regimen induces the best response in patients' endometrial tissue, repeated evaluation of the risk of endometrial damage from hysteroscopic surgery, and the risk of disease progression and recurrence due to delayed radical surgery.

[0003] Complete remission (CR) is defined as pathological examination showing complete disappearance of atypical endometrial hyperplasia or cancerous lesions, replaced by flat or cuboidal glands, stromal decellularization or fibrosis; imaging examinations showed no tumors in the chest, abdomen, or pelvis. Complete remission is crucial to evaluate the safety of fertility-sparing treatment in patients with early EC and AEH. Inspired by these findings, we innovatively designated complete remission as the primary outcome, focusing on patients with AEH and early EC with a strong desire for fertility. We integrated a series of factors, including demographic, clinical, pathological, and imaging characteristics, to construct a prediction model. Our goal is to more accurately identify young patients suitable for fertility-sparing treatment, reduce the possibility of disease progression, minimize unnecessary hysteroscopy, and improve pregnancy outcomes. Summary of the invention

[0004] The purpose of the present invention is to provide a method, device, medium and product for predicting complete remission in patients with endometrial cancer, which can accurately identify young EC patients who are suitable for preserving fertility to improve the accuracy of the clinical decision-making process.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] In a first aspect, the present invention provides a method for predicting complete remission in patients with endometrial cancer, the method comprising:

[0007] A data set is acquired; the data set includes sagittal T2-weighted images of several patients, a region of interest of each sagittal T2-weighted image, and a diffusion-weighted imaging result; the diffusion-weighted imaging result includes: low signal, medium signal, and high signal, corresponding to values ​​of 0, 1, and 2, respectively.

[0008] Feature extraction is performed on the data set to obtain several radiomics features.

[0009] An MR scoring model is established based on several of the imaging genomics features; the MR scoring model is a model established based on the degree of correlation between several of the imaging genomics features and the prognostic outcome.

[0010] The predictive efficacy of the MR scoring model is evaluated to obtain the MR score.

[0011] A normal distribution test is performed on several of the imaging omics features to obtain several imaging omics features converted into normal distribution.

[0012] A variance homogeneity analysis is performed on several of the imaging omics features after transformation into a normal distribution to determine the differences among the several of the imaging omics features after transformation into a normal distribution.

[0013] According to the differences of several radiomic features after transformation into normal distribution, predictive factors associated with complete remission of patients are determined; the predictive factors include: clinical characteristics, pathological characteristics, metabolic risk score and menstrual regularity.

[0014] A nomogram model was constructed according to the MR score and the predictive factors.

[0015] Based on the nomogram model, a nomogram is determined; the nomogram is used to evaluate the possibility of complete remission in women with early endometrial cancer and atypical hyperplasia.

[0016] In a second aspect, the present invention provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described methods for predicting complete remission in patients with endometrial cancer.

[0017] In a third aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for predicting complete remission in patients with endometrial cancer.

[0018] In a fourth aspect, the present invention provides a computer program product, comprising a computer program, which, when executed by a processor, implements any of the above-mentioned methods for predicting complete remission in patients with endometrial cancer.

[0019] According to the specific embodiments provided by the present invention, the present invention discloses the following technical effects:

[0020] The present invention provides a prediction method, device, medium and product for complete remission of patients with endometrial cancer. Several imaging genomics features are obtained by extracting features from an acquired data set. The data set includes sagittal T2-weighted images of several patients, a region of interest of each sagittal T2-weighted image and diffusion-weighted imaging results. The diffusion-weighted imaging results include: low signal, medium signal and high signal, corresponding to values ​​of 0, 1 and 2, respectively. An MR scoring model is established based on several of the imaging genomics features, and the predictive efficacy of the MR scoring model is evaluated to obtain an MR score. The MR score can promote the precise implementation of conservation therapy. A normal distribution test is performed on several of the imaging omics features to obtain several imaging omics features converted to normal distribution, and variance homogeneity analysis is performed on several of the imaging omics features converted to normal distribution to determine the differences of several of the imaging omics features converted to normal distribution, and the predictive factors related to complete remission of patients are determined based on the differences of several of the imaging omics features converted to normal distribution; a nomogram model is constructed based on the MR score and the predictive factors; finally, a nomogram is determined based on the nomogram model; the nomogram is used to evaluate the possibility of complete remission in women with early endometrial cancer and atypical hyperplasia. The present invention can accurately identify young EC patients who are suitable for preserving fertility to improve the accuracy of the clinical decision-making process. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0022] Figure 1 A schematic flow chart of a method for predicting complete remission in patients with endometrial cancer provided by one embodiment of the present invention.

[0023] Figure 2 A schematic diagram of a project flow overview provided for an embodiment of the present invention; wherein: Figure 2 (a) is a schematic diagram of inclusion and exclusion criteria; Figure 2 (b) Schematic diagram of extracting imaging genomics features.

[0024] Figure 3 A schematic diagram of a calculation formula for constructing an MR score provided in one embodiment of the present invention; wherein, Figure 3 (a) Schematic diagram of the algorithm for screening imaging omics features; Figure 3 (b) Schematic diagram for selecting the optimal λ value; Figure 3 (c) is a schematic diagram of obtaining each characteristic coefficient.

[0025] Figure 4 A schematic diagram of establishing and evaluating an MR scoring model provided by an embodiment of the present invention; wherein: Figure 4 (a) Schematic diagram of the distribution differences of baseline information under six imaging omics features; Figure 4 (b) Schematic diagram of the relationship between the six imaging omics features; Figure 4 (c) Schematic diagram of the correlation coefficients among the six imaging omics features; Figure 4 (d) Schematic diagram showing that the patients in the training group were evenly divided into the low MR score group and the high MR score group; Figure 4 (e) Schematic diagram showing that the patients in the validation group were evenly divided into low MR score group and high MR score group; Figure 4 (f) is a schematic diagram of the CR possibility of patients in the training group; Figure 4 (g) is a schematic diagram of the possibility of CR in patients in the validation group; Figure 4 (h) is a schematic diagram of the prediction effect of the training group; Figure 4 (i) Schematic diagram of the prediction effect of the validation group.

[0026] Figure 5 A schematic diagram of establishing and evaluating a nomogram model after incorporating MR scoring provided in one embodiment of the present invention; wherein: Figure 5 (a) is a schematic diagram of the nomogram model after incorporating MR score; Figure 5 (b) is a schematic diagram of the prediction accuracy of the nomogram model; Figure 5 (c) Schematic diagram of model 1 in which patients are divided equally into three groups to evaluate the possibility of CR; Figure 5 (d) Schematic diagram of model 1 in which patients are divided equally into three groups to evaluate the possibility of CR; Figure 5 (e) Schematic diagram of the ROC curve of the nomogram excluding the MR score in model 1; Figure 5 (f) is a schematic diagram of the ROC curve of the nomogram of model 2 excluding the MR score; Figure 5 (g) is a schematic diagram of the standardized net benefit of Model 1; Figure 5 (h) Schematic diagram of the standardized net benefit of Model 2.

[0027] Figure 6 A schematic diagram of the structure of a computer device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0030] Magnetic resonance imaging (MRI) plays a vital role in diagnosis and staging because of its significant advantages in assessing the spread of intrauterine tumors, assessing invasion of the myometrium and cervix, and detecting pelvic tumor deposits. Radiomics is an advanced technology that extracts numerous features from medical images, including analyzing size, contour, and texture details, providing important information about pixel or voxel arrangements, and then integrating these extracted features into statistical models to help develop personalized diagnostic and treatment strategies. Radiomics effectively reveals a wide range of hidden objective features in medical images that are usually not available to radiologists through film reading.

[0031] The study of this invention aims to integrate a series of factors, including demographic, clinicopathological and molecular lesion characteristics, with complete remission after endometrial lesions as the primary outcome. The purpose of this invention is to quantify the impact of each factor on fertility in order to accurately identify young EC patients who are suitable for preserving fertility and improve the accuracy of the clinical decision-making process.

[0032] In an exemplary embodiment, Figure 1 As shown, a method for predicting complete remission in patients with endometrial cancer is provided, comprising the following steps S1 to S8. Among them:

[0033] S1: Acquire a data set; the data set includes sagittal T2-weighted images of several patients, a region of interest of each sagittal T2-weighted image, and diffusion-weighted imaging results; the diffusion-weighted imaging results include: low signal, medium signal, and high signal, corresponding to values ​​of 0, 1, and 2, respectively.

[0034] S2: extracting features from the data set to obtain a number of radiomics features.

[0035] S3: Establishing an MR scoring model based on the plurality of radiomics features; the MR scoring model is a model established based on the degree of correlation between the plurality of radiomics features and the prognostic outcome.

[0036] S4: Evaluate the predictive efficacy of the MR scoring model to obtain a MR score.

[0037] S5: performing a normal distribution test on the plurality of imaging omics features to obtain a plurality of imaging omics features converted into a normal distribution.

[0038] S6: performing variance homogeneity analysis on the plurality of imaging omics features transformed into normal distribution to determine the differences among the plurality of imaging omics features transformed into normal distribution.

[0039] S7: Based on the differences in several radiomic features after transformation into normal distribution, predictive factors associated with complete remission of patients are determined; the predictive factors include: clinical characteristics, pathological characteristics, metabolic risk score and menstrual regularity.

[0040] S8: constructing a nomogram model according to the MR score and the predictive factors.

[0041] S9: Based on the nomogram model, determine a nomogram; the nomogram is used to evaluate the possibility of complete remission in women with early endometrial cancer and atypical hyperplasia.

[0042] By implementing the above steps S1 to S8, a prediction model for evaluating the reliability of CR in EC patients was constructed, an “MR score” was developed through radiomics features, and this score was incorporated into the nomogram model to facilitate the precise implementation of conservation therapy.

[0043] like Figure 2 As shown, in step S1, it specifically includes:

[0044] S101: Based on the inclusion and exclusion criteria, patients were screened to obtain the screened patients; the inclusion criteria included: pathological diagnosis of atypical endometrial hyperplasia or stage IA grade 1 endometrial adenocarcinoma, strong willingness to preserve the child, signed informed consent and acceptable follow-up; the exclusion criteria included: history of serious medical diseases, advanced cancer stage or invasive characteristics, drug allergy or contraindications, other malignancies or recent conservative treatment, previous surgery or alternative therapy and history of major cardiovascular events.

[0045] S102: Obtain a sagittal T2-weighted image of the screened patient.

[0046] S103: Based on the sagittal T2 weighted image, determine a region of interest corresponding to the sagittal T2 weighted image.

[0047] In this example, we collected data from 187 EC patients seeking fertility preservation who were treated at Peking University People's Hospital from January 1, 2013 to December 31, 2022. The inclusion and exclusion criteria of the present invention are as follows: Figure 2 As shown in (a), the inclusion criteria were: pathological diagnosis of atypical endometrial hyperplasia or stage IA grade 1 endometrial adenocarcinoma, strong willingness to preserve, signed informed consent, and acceptable follow-up; exclusion criteria were: history of severe medical illness, advanced cancer stage or invasive characteristics, drug allergy or contraindications, other malignancies or recent conservative treatment, previous surgery or alternative therapy, and history of major cardiovascular events. These patients were then randomly divided into a training group and a validation group in a 2:1 ratio. According to the above inclusion and exclusion criteria, a total of 155 patients were included in our study, including 97 in the training group and 58 in the validation group. The demographic and clinical characteristics of each group were analyzed separately, and the baseline characteristics of patients with incomplete response (IR) and patients with CR results were compared (Tables 1 and 2). Overall, the characteristics of the patients were relatively consistent. In the training group, differences were found in body mass index (BMI) (p=0.003), insulin levels (p=0.001), hypertension (HP) status (p=0.031), menstrual regularity (p=0.013), and pathology (p=0.019), while in the validation group, differences were found only in pathology (p=0.004).

[0048] Table 1 Basic information and clinical data of the training cohort

[0049]

[0050] Table 2 Basic information and clinical data of the validation cohort

[0051]

[0052]

[0053] It should be noted that the data in Tables 1 and 2 are expressed as mean ± standard deviation or number (percent %).

[0054] Patients were asked to fast for at least 4 hours before receiving MRI. All preoperative MRIs were obtained on a Disco very7503.0T (GE Healthcare, Waukesha, WI, USA) MRI image scanner. In order to construct the imaging omics model in the present invention, the patient's sagittal T2-weighted imaging (T2WI) was used and imported into the open source software ITK-SNAP (http: / / www.itksnap.org / , version 3.8.0). Two senior doctors were responsible for manually delineating the region of interest (ROI) layer by layer along the edge of the tumor, such as Figure 2 (b) and imported them into 3D slicer (https: / / www.slicer.org / , version 4.13.0), and reconstructed them into three-dimensional structures through an automated process.

[0055] In an exemplary embodiment, in step S2, it specifically includes:

[0056] S201: Standardizing the data set using z-score to obtain a standardized data set.

[0057] S202: Using the PyRadiomics algorithm to extract features from the standardized data set to obtain a number of radiomics features.

[0058] In this embodiment, before extracting imaging features, the signal intensity is normalized by z-score conversion to address the inconsistency of magnetic resonance scan intensity. 1510 imaging features of patient MRI were extracted by PyRadiomics algorithm (https: / / pyradiomics.readthedocs.io / en / latest / features.html).

[0059] In another exemplary embodiment of the present invention, three algorithms, namely machine learning, Cox regression and random forest, were used to identify and filter the radiological features that were most closely associated with the patient's CR outcome. Subsequently, we selected 82 features that met the criteria of all three algorithms for inclusion in Lasso regression. Lasso regression is an extension of generalized linear regression, which reduces the variance and prediction error of the regression coefficients by adding a penalty term for model complexity to the log-likelihood function, effectively solving the problem of multicollinearity between parameters. Lasso feature selection reduces the model coefficients by adjusting the penalty coefficient λ, thereby obtaining an MR scoring formula containing independent variables that are strongly correlated with the dependent variable and its corresponding regression coefficient.

[0060] Univariate Cox regression analysis was performed to identify prognostic factors in the patient cohort, including the MR score derived from the above model. The nomogram was constructed using independent risk factors, and its performance was evaluated by calculating the discrimination criterion and correction criterion. The nomogram discrimination performance was evaluated by receiver operating characteristic (ROC) curve analysis, providing indicators such as the area under the ROC curve (AUC) and 95% confidence interval. We used calibration curves to validate the prediction model and evaluate the selected factor set. A calibration curve that closely followed the ideal line was considered to represent the prediction accuracy of the nomogram. In addition, we performed decision curve analysis (DCA) to visualize the net benefit of clinical decision making.

[0061] The calculation formula of the MR scoring model is:

[0062] MR score=-0.15235×lbp3Dk_ngtdm_Busyness-0.04730×square_glrlm_LongRu nLowGrayLevelEmphasis-0.00789×squareroot_glrlm_RunLengthNonUniformity-0.02973×waveletHHL_ngtdm_Bus yness-0.02907*waveletHLL_ngtdm_Busyness-0.04663×waveletLHL_glrlm_RunLengthNonUniformity-1.581*DWI;

[0063] Among them, MR score is MR score, lbp3Dk_ngtdm_Busyness is the first busyness, squ are_glrlm_LongRunLowGrayLevelEmphasis is the long run low gray level emphasis, squareroot_glrl m_RunLengthNonUniformity is the first run non-uniformity, waveletHHL_ngtdm_Busynes s is the second busyness, waveletHLL_ngtdm_Busyness is the third busyness, waveletLHL_glrl m_RunLengthNonUniformity is the second run non-uniformity, and DWI is diffusion weighted imaging. The busyness means the degree of change from one pixel to its adjacent area; the long run low gray level emphasis means the difference distribution between different gray values; the run non-uniformity means the similarity of the run length of the entire image. The lower the value, the stronger the homogeneity of the entire image.

[0064] In another exemplary embodiment of the present invention, in step S4, specifically including:

[0065] S401: Based on a plurality of the radiomics features, screen out radiomics features having a correlation greater than a preset threshold.

[0066] S402: Inputting the radiomics features greater than a preset correlation threshold into an MR scoring model to obtain an MR score.

[0067] In another exemplary embodiment of the present invention, in step S5, the following steps are specifically included:

[0068] S501: When the plurality of radiomics features are continuous variables, a t-test or a rank sum test is used to obtain the plurality of radiomics features after conversion to a normal distribution.

[0069] S502: When several of the radiomics features are categorical variables, a chi-square test is used to obtain several radiomics features after conversion into a normal distribution.

[0070] In another exemplary embodiment of the present invention, in step S6, specifically including:

[0071] S601: When the variances are equal, a t-test is used to determine the differences of several radiomics features after transformation into a normal distribution.

[0072] S602: When the variances are not uniform, the Mann-Whitney u test is used to determine the differences of several radiomics features after transformation into a normal distribution.

[0073] In the clinical data, the numerical data were first tested for normal distribution and variance homogeneity, with t-test or rank-sum test for continuous variables and chi-square test for categorical variables. After the imaging features were transformed into a normal distribution, the t-test (for homogeneous variance) or Mann-Whitney u-test (for unequal variance) was used to determine their differences. Subsequently, univariate Cox regression analysis was performed to determine the predictive factors associated with CR. All the above tests considered the differences to be significant when P < 0.05. Statistical analysis and data visualization were performed using the R4.3.3 software package (https: / / www.r-project.org / , the R Foundation) and Empower Statistical Software (http: / / www.empowerstats.com, X&Y Solutions, Inc., Boston, MA).

[0074] In another exemplary embodiment of the present invention, a patient MR score calculation formula is constructed as follows: Figure 3 In the study of this embodiment,

[0075] Three different methods were used to extract 1468 out of 1510 potential radiomic features: machine learning extracted 767, Cox regression extracted 444, and random forest extracted 374. After these crossovers, we identified and selected 82 features that were highly associated with EC patients, such as Figure 3 (a) is shown. In the Lasso regression method, the optimal λ value (λ = 6) is determined, and the focus is on minimizing the partial likelihood deviation ( Figure 3 b). Then, a risk assessment formula was developed incorporating six core radiomic features, a DWI classification variable and their respective coefficients, allowing calculation of a personalized MR score for each patient ( Figure 3 c).

[0076] The calculation formula of the MR scoring model is:

[0077] MR score=-0.15235×lbp3Dk_ngtdm_Busyness-0.04730×square_glrlm_LongRu nLowGrayLevelEmphasis-0.00789×squareroot_glrlm_RunLengthNonUniformity-0.02973×waveletHHL_ngtdm_Bus yness-0.02907*waveletHLL_ngtdm_Busyness-0.04663×waveletLHL_glrlm_RunLengthNonUniformity-1.581*DWI;

[0078] Among them, MR score is MR score, lbp3Dk_ngtdm_Busyness is the first busyness, squ are_glrlm_LongRunLowGrayLevelEmphasis is the long run low gray level emphasis, squareroot_glrl m_RunLengthNonUniformity is the first run non-uniformity, waveletHHL_ngtdm_Busynes s is the second busyness, waveletHLL_ngtdm_Busyness is the third busyness, waveletLHL_glrl m_RunLengthNonUniformity is the second run non-uniformity, and DWI is diffusion weighted imaging. The busyness means the degree of change from one pixel to its adjacent area; the long run low gray level emphasis means the difference distribution between different gray values; the run non-uniformity means the similarity of the run length of the entire image. The lower the value, the stronger the homogeneity of the entire image.

[0079] In another exemplary embodiment of the present invention, the establishment and evaluation of the MR scoring model is as follows: Figure 4 As shown in the heat map, the six radioactive features showed different enrichments. In addition, the baseline characteristics of these groups were different, suggesting that combining various indicators may enhance the predictive power of fertility-sparing treatment in EC patients, such as Figure 4 (a). In addition, these six radiomic features showed high positive correlation, such as Figure 4 (b) The correlation between these features is within an acceptable range, except that there are high correlation coefficients between some grayscale features, such as Figure 4 (c) According to the mean score of LASSO regression, the two groups of patients were divided into high MR score and low MR score subgroups. The MR score and CR status of each individual were sorted and displayed on a dot plot, showing that there were significant differences in CR between the groups, as shown in Figure 4 (d) and Figure 4 (e) The MR score effectively predicted the probability of CR in the training group (p = 7.745e-04) and the validation group (p = 9.414e-06), where the higher the MR score, the greater the possibility of achieving CR, as shown in Figure 4 (f) and Figure 4 In addition, the model achieved impressive area under the curve (AUC) values ​​of 0.871 and 0.833, respectively, which strongly suggests that the six identified MR features and DWI manifestations are key indicators of potential treatment efficacy, such as Figure 4 (h) and Figure 4 (i) as shown.

[0080] In another exemplary embodiment of the present invention, a nomogram model for accurately evaluating the possibility of CR is constructed as follows: Figure 5 Initially, univariate analysis identified significant differences in clinical characteristics between patients with CR and IR in the training group (Table 3). This analysis led to the selection of four variables: metabolic risk score (MRS) (p = 0.0077), menstrual regularity (p = 0.0199), pathology (p = 0.0007), and MR score (p = 0.0056) for inclusion in the nomogram, as shown in Figure 3. Figure 5 (a) shows. A nomogram was established to predict the possibility of achieving CR and the time required to achieve CR to assess the feasibility of fertility-sparing treatment for patients. Our model showed commendable prediction accuracy at various time periods to achieve CR, as shown in Figure 5 (b) As shown. Subsequently, we combined the training group and the validation group to jointly evaluate the predictive performance of the modal graph. Even when the population was divided into three levels, the difference in CR was still significant, with a p-value of 0.0151 for the training cohort and a p-value of less than 0.001 for the validation cohort, as shown in Figure 5 (c) and Figure 5 (d) For clarity, the nomogram without MR score is referred to as model 1, while the original nomogram with MR score is referred to as model 2. Most importantly, the inclusion of radiomic feature scores significantly improved the predictive ability of the model, with the AUC of the training group increasing from 0.827 to 0.935 and the AUC of the validation group increasing from 0.804 to 0.916, as shown in Figure 2. Figure 5 (e) and Figure 5 As shown in (f), it can be seen from the decision curve analysis (DCA) diagram that compared with model 1, regardless of the proportion of CR patients in the population, model 2 always shows a higher standardized net benefit, as shown in Figure 5 (g) and Figure 5 (h) shown.

[0081] Table 3 Univariate COX analysis of clinical characteristics of the training set

[0082]

[0083]

[0084] The present invention constructs a predictive model for evaluating the reliability of CR in EC patients, develops an "MR score" through radiomics features, and incorporates this score into the nomogram model to promote the precise implementation of conservation therapy.

[0085] The present invention also provides an application scenario, which applies the above-mentioned method for predicting complete remission of endometrial cancer patients. Specifically: The prediction method for complete remission of endometrial cancer patients provided in this embodiment can be applied in the prediction performance scenario of endometrial cancer conservation patients, and the prediction performance scenario of endometrial cancer conservation patients includes: data acquisition link, data processing link and nomogram model construction link; feature extraction of the data set to obtain a number of imaging genomics features; then entering the data processing link, establishing an MR scoring model based on the several imaging genomics features; predictive efficacy evaluation of the MR scoring model to obtain an MR score; normal distribution test of the several imaging genomics features to obtain a number of imaging genomics features converted to a normal distribution; variance homogeneity analysis of the several imaging genomics features converted to a normal distribution to determine the differences of the several imaging genomics features converted to a normal distribution; based on the differences of the several imaging genomics features converted to a normal distribution, determine the predictive factors related to the complete remission of the patient; finally, according to the MR score and the predictive factors, a nomogram model can be constructed; the nomogram model is used to evaluate the possibility of complete remission of the patient.

[0086] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 6 As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. Among them, the processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store a data set, and the data set includes sagittal T2 weighted images of several patients and the region of interest of each sagittal T2 weighted image. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for predicting complete remission of endometrial cancer patients is implemented.

[0087] Those skilled in the art will understand that Figure 6 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0088] In an exemplary embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the above-mentioned method embodiments are implemented when the processor executes the computer program.

[0089] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program, and when the computer program is executed by a processor, the above-mentioned method embodiments are implemented.

[0090] In an exemplary embodiment, a computer program product is provided, including a computer program, and when the computer program is executed by a processor, the above-mentioned method embodiments are implemented.

[0091] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0092] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided by the present invention can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0093] The database involved in each embodiment provided by the present invention may include at least one of a relational database and a non-relational database. The non-relational database may include a distributed database based on blockchain, etc., but is not limited thereto. The processor involved in each embodiment provided by the present invention may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, etc., but is not limited thereto.

[0094] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0095] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A method for predicting complete remission in patients with endometrial cancer, characterized in that: The method for predicting complete remission in patients with endometrial cancer includes: Acquire a data set; the data set includes sagittal T2-weighted images of several patients, a region of interest of each sagittal T2-weighted image, and a diffusion-weighted imaging result; the diffusion-weighted imaging result includes: low signal, medium signal, and high signal, corresponding to values ​​of 0, 1, and 2, respectively; Performing feature extraction on the data set to obtain a number of radiomics features; According to the plurality of radiomics features, an MR scoring model is established; the MR scoring model is a model established according to the degree of correlation between the plurality of radiomics features and the prognosis outcome; Evaluating the predictive effectiveness of the MR scoring model to obtain a MR score; Performing a normal distribution test on a number of the imaging omics features to obtain a number of imaging omics features converted into a normal distribution; Performing variance homogeneity analysis on the several radiomics features transformed into normal distribution to determine the differences among the several radiomics features transformed into normal distribution; Determining predictive factors associated with complete remission of patients based on the differences of several radiomic features transformed into normal distribution; the predictive factors include: clinical characteristics, pathological characteristics, metabolic risk score and menstrual regularity; constructing a nomogram model according to the MR score and the predictive factors; Based on the nomogram model, a nomogram is determined; the nomogram is used to evaluate the possibility of complete remission in women with early endometrial cancer and atypical hyperplasia.

2. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: Get the dataset, including: Based on the inclusion and exclusion criteria, the patients were screened to obtain the screened patients; the inclusion criteria included: pathological diagnosis of atypical endometrial hyperplasia or stage IA grade 1 endometrial adenocarcinoma, strong willingness to preserve, signed informed consent, and acceptable follow-up; the exclusion criteria included: history of severe medical diseases, advanced cancer stage or invasive characteristics, drug allergy or contraindications, other malignancies or recent conservative treatments, previous surgical or alternative therapies, and history of major cardiovascular events; Obtain sagittal T2-weighted images and MRI report results of screened patients; Based on the sagittal T2 weighted image, determining a region of interest corresponding to the sagittal T2 weighted image; Based on the nuclear magnetic resonance report results, the diffusion weighted imaging results are determined.

3. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: Feature extraction is performed on the data set to obtain several radiomics features, including: The data set is standardized by using a z-score to obtain a standardized data set; The PyRadiomics algorithm was used to extract features from the standardized data set to obtain several radiomics features.

4. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: A normal distribution test is performed on several of the imaging omics features to obtain several imaging omics features converted into normal distribution, specifically including: When several of the radiomics features are continuous variables, a t-test or a rank sum test is used to obtain several radiomics features after conversion to a normal distribution; When several of the radiomics features are categorical variables, a chi-square test is used to obtain several radiomics features after conversion to a normal distribution.

5. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: The variance homogeneity analysis is performed on the several radiomics features after being transformed into a normal distribution to determine the differences of the several radiomics features after being transformed into a normal distribution, specifically including: When the variances were equal, the t-test was used to determine the differences of several radiomic features after transformation into normal distribution; When variances were unequal, the Mann-Whitney u test was used to determine the differences in several radiomics features after transformation to a normal distribution.

6. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: The predictive effectiveness of the MR scoring model is evaluated to obtain the MR score, specifically including: Based on the plurality of radiomics features, screening out radiomics features having a correlation greater than a preset threshold; The radiomics features greater than a preset correlation threshold are input into the MR scoring model to obtain the MR score.

7. The method for predicting complete remission in patients with endometrial cancer according to claim 1, characterized in that: The calculation formula of the MR scoring model is: MR score=-0.15235×lbp3Dk_ngtdm_Busyness-0.04730×square_glrlm_LongRunLowGrayLevelEmphasis-0.00789×squareroot_glrlm_RunLengthNonUniform ity-0.02973×waveletHHL_ngtdm_Busyness-0.02907*waveletHLL_ngtdm_Busyness-0.04663×waveletLHL_glrlm_RunLengthNonUniformity-1.581*DWI; Among them, MR score is the MR score, lbp3Dk_ngtdm_Busyness is the first busyness, squ are_glrlm_LongRunLowGrayLevelEmphasis is the long run low gray level emphasis, squareroot_glrl m_RunLengthNonUniformity is the first run unevenness, waveletHHL_ngtdm_Busynes s is the second busyness, waveletHLL_ngtdm_Busyness is the third busyness, waveletLHL_glrl m_RunLengthNonUniformity is the second run unevenness, and DWI is diffusion weighted imaging.

8. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the method for predicting complete remission of endometrial cancer patients according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for predicting complete remission in endometrial cancer patients according to any one of claims 1 to 7 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for predicting complete remission in endometrial cancer patients according to any one of claims 1 to 7 is implemented.

Images