Non-invasive grading method and device for liver inflammation and storage medium

Abstract

The invention discloses a non-invasive grading method and device for liver inflammation and a storage medium, and the method comprises the steps: obtaining at least two evaluation indexes of the liver in a target human body, and the at least two evaluation indexes comprise an apparent dispersion coefficient ADC value and a T2 weighted imaging signal intensity index SII value of the same region in a magnetic resonance imaging MRI image of the liver; calculating a joint value corresponding to the values of all the evaluation indexes, wherein the joint value is positively correlated with the SII value and is negatively correlated with the ADC value; and according to the joint value, determining a liver inflammation grading result of the target human body, so that the accuracy of liver inflammation grading can be effectively improved.

Description

Technical Field

[0001] This article relates to the field of medical technology, and in particular to a non-invasive grading method, device and storage medium for liver inflammation. Background Technology

[0002] The grading of liver inflammation is crucial for disease diagnosis, treatment, and prognosis. Traditionally, liver inflammation grading has relied primarily on liver biopsy, an invasive procedure that can cause complications such as pain and bleeding, and is also costly and time-consuming. With advancements in imaging technology, magnetic resonance imaging (MRI) has emerged as a promising tool for liver inflammation assessment due to its non-invasive nature and high resolution. Therefore, how to identify the non-invasive grading of liver inflammation based on MRI images is a pressing issue that needs to be addressed. Summary of the Invention

[0003] This application provides a non-invasive method, apparatus, and storage medium for grading liver inflammation.

[0004] A non-invasive grading method for liver inflammation includes:

[0005] Obtain at least two evaluation metrics for the liver in the target human body, wherein the at least two evaluation metrics include the apparent diffusion coefficient (ADC) value and the T2-weighted imaging signal intensity index (SII) value of the same region in the magnetic resonance imaging (MRI) image of the liver;

[0006] Calculate the joint value corresponding to the values ​​of all evaluation indicators, wherein the joint value is positively correlated with the SII value and negatively correlated with the ADC value;

[0007] Based on the combined values, the liver inflammation grading result of the target human body is determined.

[0008] A storage medium storing a computer program, wherein the computer program is configured to execute the method described above when run.

[0009] A non-invasive liver inflammation grading device includes a memory and a processor, wherein the memory stores a computer program and the processor is configured to run the computer program to perform the method described above.

[0010] In this embodiment, by combining ADC value and SII value, the pathophysiological changes of liver inflammation can be reflected more comprehensively, thereby improving the accuracy of liver inflammation grading.

[0011] Other features and advantages of this application will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the application. Other advantages of this application can be realized and obtained by means of the solutions described in the description and the accompanying drawings. Attached Figure Description

[0012] The accompanying drawings are used to provide an understanding of the technical solutions of this application and constitute a part of the specification. They are used together with the embodiments of this application to explain the technical solutions of this application and do not constitute a limitation on the technical solutions of this application.

[0013] Figure 1 A flowchart illustrating the non-invasive grading method for liver inflammation provided in this application embodiment;

[0014] Figure 2 A schematic diagram of a calibration curve for predicting grade ≥G2 liver inflammation in CHB patients based on a combined index of ADC and SII values, provided in an embodiment of this application.

[0015] Figure 3 A schematic diagram of a clinical decision curve for predicting ≥G2 grade liver inflammation in CHB patients based on a combined index of ADC and SII values, provided in an embodiment of this application. Detailed Implementation

[0016] This application describes several embodiments, but these descriptions are exemplary and not restrictive, and it will be apparent to those skilled in the art that many more embodiments and implementations are possible within the scope of the embodiments described herein. Although many possible combinations of features are shown in the drawings and discussed in the detailed description, many other combinations of the disclosed features are also possible. Unless specifically limited, any feature or element of any embodiment may be used in combination with, or may replace, any feature or element of any other embodiment.

[0017] This application includes and contemplates combinations of features and elements known to those skilled in the art. The embodiments, features, and elements disclosed in this application can also be combined with any conventional features or elements to form unique inventive solutions. Any feature or element of any embodiment can also be combined with features or elements from other inventive solutions to form another unique inventive solution. Therefore, it should be understood that any feature shown and / or discussed in this application can be implemented individually or in any suitable combination. Therefore, the embodiments are not limited except by the limitations imposed by the appended claims and their equivalents. Furthermore, various modifications and changes can be made within the scope of the appended claims.

[0018] Furthermore, in describing representative embodiments, the specification may have presented methods and / or processes as a specific sequence of steps. However, the method or process should not be limited to the specific order of steps described herein, to the extent that it does not depend on such a specific order. As will be understood by those skilled in the art, other sequences of steps are also possible. Therefore, the specific order of steps set forth in the specification should not be construed as a limitation of the claims. Moreover, the claims concerning the method and / or process should not be limited to the steps performed in the written order, and those skilled in the art will readily understand that these orders can be varied and still remain within the spirit and scope of the embodiments of this application.

[0019] The apparent diffusion coefficient (ADC) is a quantitative parameter in diffusion-weighted imaging (DWI) that reflects the diffusion and movement of water molecules within tissues. In the assessment of liver inflammation, inflammation causes changes in intracellular and extracellular water content, thus affecting the ADC value. Generally, the more severe the inflammation and the more pronounced the cellular edema, the lower the ADC value may be.

[0020] The T2-weighted imaging signal intensity index (SII) is calculated by measuring the ratio of the signal intensity of the liver lesion area to that of a standardized reference area (such as the right erector spinae muscle) in T2WI. Changes in the SII value can reflect pathological changes in liver tissue, such as inflammation and edema; the more severe the inflammation, the higher the SII value may be.

[0021] In related technologies, there have been some studies on the application of ADC and SII values ​​in liver inflammation assessment, but a non-invasive grading method that integrates multiple evaluation indicators is lacking. Furthermore, different types of liver inflammation (such as metabolic-associated steatohepatitis and drug-induced hepatitis) have distinct clinical indicators and imaging characteristics, requiring a more universal and accurate assessment scheme. Therefore, this application provides a non-invasive liver inflammation grading method based on a combination of multiple indicators, which has significant clinical value for improving diagnostic accuracy, enabling early intervention, and personalized treatment.

[0022] Figure 1 This is a schematic flowchart illustrating the non-invasive grading method for liver inflammation provided in an embodiment of this application. Figure 1 As shown, the method includes:

[0023] Step 101: Obtain at least two evaluation indicators for the liver in the target human body, wherein the at least two evaluation indicators include the ADC value and SII value of the same region in the MRI image of the liver.

[0024] The same region refers to image data acquired at the same anatomical location in the liver to ensure that the measurements of ADC and SII values ​​are comparable and accurate.

[0025] Specifically, the ADC value of the target human body is obtained based on the DWI sequence, where the b-value (also known as the diffusion sensitivity factor) of the DWI sequence is 800 s / mm. 2 DWI sequence layer number ≥ 40: using a b value of 800 s / mm 2 Using DWI sequences, and requiring a DWI sequence layer count of ≥40, ensures that the acquired ADC values ​​have sufficient sensitivity and accuracy. Higher b-values ​​better reflect the restricted diffusion of water molecules in tissues, while more layers cover a larger liver area, providing more comprehensive information.

[0026] The SII value of the target human body is obtained by fat-suppressed T2-weighted imaging (FS T2WI) sequence. FS T2WI can effectively suppress the interference of fat signals, making the signals of liver lesions clearer and more distinguishable. Therefore, the SII value obtained by FS T2WI sequence can more accurately reflect the inflammation and edema of liver tissue.

[0027] The purpose of this step is to collect key data required for liver inflammation assessment. Obtaining ADC and SII values ​​through MRI scans provides a non-invasive way to reflect the pathological state of the liver, avoiding the invasiveness and associated risks of traditional liver biopsies. This data provides the foundation for subsequent liver inflammation grading.

[0028] Step 102: Calculate the joint value corresponding to the values ​​of all evaluation indicators, wherein the joint value is positively correlated with the SII value and negatively correlated with the ADC value.

[0029] The combined value is a comprehensive indicator calculated by weighting the ADC value, SII value, and other possible evaluation indicators. It is used to comprehensively reflect the degree of liver inflammation and serves as a basis for determining the grade of liver inflammation.

[0030] The combined value is positively correlated with the SII value and negatively correlated with the ADC value. This means that when the SII value increases, the combined value also increases; while when the ADC value increases, the combined value decreases. This relationship reflects the pathophysiological changes in liver inflammation: inflammation-induced cellular edema and tissue destruction increase the signal intensity of T2WI (increased SII value) while restricting the diffusion of water molecules (decreased ADC value). Therefore, this correlation design of the combined value can more accurately reflect the true state of liver inflammation.

[0031] The purpose of this step is to integrate multiple evaluation indicators into a single comprehensive indicator. By combining data from different dimensions through a mathematical model, the assessment results become more comprehensive and reliable. The calculation of the combined value takes into account the relative importance of each indicator in the assessment of liver inflammation, thereby improving the accuracy of the assessment results.

[0032] Step 103: Determine the liver inflammation grading result of the target human body based on the combined value.

[0033] Liver inflammation grading is a classification and assessment of the degree of liver inflammation. It is typically based on pathological characteristics, dividing liver inflammation into different grades, such as G0 (no inflammation), G1 (mild inflammation), G2 (moderate inflammation), and G3 (severe inflammation). This grading helps guide clinical treatment and predict disease progression. It also helps doctors develop more precise treatment plans and assess patient prognosis.

[0034] The method provided in this application provides a non-invasive method for grading liver inflammation, which achieves quantitative assessment of liver inflammation by combining indicators such as ADC value and SII value. The following is a detailed description:

[0035] Offering a non-invasive assessment method: Traditional methods for grading liver inflammation typically rely on liver biopsy, an invasive procedure that can cause complications such as pain and bleeding, and is also costly and time-consuming. This method, however, calculates ADC and SII values ​​by acquiring MRI images of the liver, eliminating the need for a liver biopsy and avoiding the risks and discomfort associated with invasive procedures, thus achieving a non-invasive assessment of liver inflammation.

[0036] Improving assessment accuracy: ADC values ​​reflect the diffusion of water molecules in tissues, while SII values ​​are related to signal intensity in the liver. During liver inflammation, hepatocytes undergo pathological changes such as edema and necrosis, which affect water molecule diffusion and signal intensity. Combining ADC and SII values ​​provides a more comprehensive reflection of the pathophysiological changes in liver inflammation, thereby improving the accuracy of liver inflammation grading.

[0037] Early diagnosis and monitoring: Liver inflammation may not present with obvious symptoms in its early stages, but subtle changes in liver tissue can be sensitively detected by obtaining ADC and SII values ​​from MRI images, enabling early diagnosis of liver inflammation. During treatment, these indicators can be repeatedly acquired to monitor changes in liver inflammation, allowing for timely adjustments to the treatment plan and improving treatment outcomes.

[0038] This protocol can be applied to the liver inflammation grading of chronic hepatitis B (CHB), fatty liver (including metabolic-associated steatohepatitis), and drug-induced hepatitis. The following are specific instructions:

[0039] The reason why CHB hepatitis grading is applicable is that the pathological basis of CHB hepatitis is hepatocellular damage and inflammatory response caused by hepatitis B virus (HBV) infection. ADC and SII values ​​can reflect pathological changes such as hepatocellular edema, necrosis, and fibrosis, which are key indicators for CHB inflammation grading.

[0040] The reason this applies to fatty liver (metabolic steatohepatitis) is that the pathological basis of fatty liver is hepatocyte steatosis, inflammation, and fibrosis. ADC and SII values ​​can also reflect these pathological changes. For example, inflammatory cellular edema and tissue destruction increase T2WI signal intensity (increased SII value) while restricting water molecule diffusion (decreased ADC value).

[0041] The reason for applying these values ​​to drug-induced hepatitis is that drug-induced hepatitis is liver damage caused by drugs or their metabolites, and its pathological features include hepatocellular necrosis, inflammation, and cholestasis. ADC and SII values ​​can reflect these pathological changes and therefore can be used for inflammation grading in drug-induced hepatitis.

[0042] This method achieves non-invasive grading of liver inflammation by acquiring the patient's ADC and SII values, calculating a combined value, and comparing it with a preset cutoff value. Since ADC and SII values ​​can reflect various liver pathological changes (such as cellular edema, necrosis, and fibrosis), this method is applicable to the assessment of various types of liver inflammation, including CHB, fatty liver, and drug-induced hepatitis. Its non-invasiveness and accuracy make it a valuable clinical tool, aiding in early diagnosis, treatment monitoring, and prognostic assessment.

[0043] By classifying hepatitis caused by different etiologies into inflammation levels, this approach can help doctors better understand disease progression, develop personalized treatment plans, and improve patients' treatment outcomes and quality of life.

[0044] In clinical applications, the diagnostic efficacy and value range of DWI, SII, and the combination of DWI and SII in predicting grade G2 liver inflammation in CHB patients were compared. See Table 1 for details:

[0045] Table 1

[0046]

[0047]

[0048] As shown in Table 1, receiver operating characteristic (ROC) curve analysis reveals that the area under the curve (AUC) of the combined method is 0.904, which is higher than that of DWI alone (AUC = 0.860) and SII alone (AUC = 0.812), indicating that this approach has higher accuracy in predicting liver inflammation grades ≥G2. Furthermore, the sensitivity (87.8%) and specificity (84.0%) of the combined method are also higher than those of DWI or SII alone.

[0049] The method provided in the embodiments of this application is described below:

[0050] In one exemplary embodiment, the evaluation metrics further include clinical metrics of the target human body.

[0051] Specifically, clinical indicators include:

[0052] Hepatitis B virus (HBV) DNA content: reflects the replication level of hepatitis B virus in the blood;

[0053] The amount of Hepatitis B surface antigen (HBsAg): a marker of Hepatitis B virus infection;

[0054] Alanine transaminase (ALT) level: reflects the degree of hepatocellular damage;

[0055] Aspartate transaminase (AST): reflects hepatocellular damage;

[0056] Alkaline phosphatase (ALP): is involved in the liver's bile excretion function;

[0057] Gamma-glutamyl transferase (GGT): reflects liver or biliary tract disease;

[0058] Platelet count (PLT): associated with liver disease;

[0059] International Normalized Ratio (INR): Reflects the liver's coagulation function;

[0060] Albumin (ALB): reflects liver's synthetic function;

[0061] Total bilirubin (TB): reflects liver metabolism and excretion function.

[0062] The clinical indicators corresponding to different types of liver inflammation are explained below:

[0063] Clinical indicators for chronic hepatitis B (CHB) include hepatitis B virus (HBV) DNA content, hepatitis B virus surface antigen (HBsAg) quantity, alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), platelet count (PLT), international normalized ratio (INR), albumin (ALB), and total bilirubin (TB).

[0064] Clinical indicators of metabolic-related steatohepatitis include alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), platelet count (PLT), international normalized ratio (INR), albumin (ALB), total bilirubin (TB), and body mass index (BMI).

[0065] Clinical indicators of drug-induced hepatitis include alanine aminotransferase (ALT), aspartate aminotransferase (AST), alkaline phosphatase (ALP), gamma-glutamyl transferase (GGT), platelet count (PLT), international normalized ratio (INR), albumin (ALB), and total bilirubin (TB).

[0066] The advantages of adding clinical indicators in the above exemplary embodiments include:

[0067] Improving the comprehensiveness of assessment: Different clinical indicators can reflect various aspects of the liver, such as hepatocellular damage, bile excretion function, and bile synthesis function. Increasing the number of clinical indicators can make the assessment more comprehensive and help to more accurately reflect the pathological state of the liver.

[0068] Enhanced assessment accuracy: By integrating multiple indicators, false positive or false negative results that may occur with a single indicator can be reduced, thereby improving the accuracy of liver inflammation grading.

[0069] Facilitating Personalized Treatment: The abnormal clinical indicators may vary depending on the etiology of liver inflammation. Including additional clinical indicators helps identify different types of liver inflammation, providing a more accurate basis for developing personalized treatment plans.

[0070] Facilitates monitoring of treatment effectiveness: By monitoring changes in multiple clinical indicators during treatment, the treatment effect can be more comprehensively evaluated, and the treatment plan can be adjusted in a timely manner.

[0071] In an exemplary embodiment, the process of obtaining the calculation expression for the joint value is as follows:

[0072] A computational model was constructed by using the liver inflammation grading results as the dependent variable and all evaluation indicators as independent variables.

[0073] The computational model is determined using a sample dataset to obtain the computational expression for the joint value. The sample dataset includes the values ​​of evaluation indicators and the corresponding liver inflammation grading results of the individuals in the case group and the control group.

[0074] In this model, the dependent variable is the outcome that needs to be predicted or explained; in this case, it is the grade of liver inflammation. For example, the inflammation grade could be G0 (no inflammation), G1 (mild inflammation), G2 (moderate inflammation), G3 (severe inflammation), etc. Using the inflammation grade as the dependent variable means that the goal of the model is to predict or explain the severity of liver inflammation.

[0075] In this context, independent variables are factors or indicators used to predict or explain the dependent variable. The evaluation indicators mentioned above include ADC and SII values; additionally, clinical indicators such as HBV DNA levels, HBsAg levels, and ALT levels can be referenced. Using these evaluation indicators as independent variables means that the model will utilize information from these indicators to predict the grade of liver inflammation.

[0076] The sample dataset is a set of data containing known evaluation index values ​​and corresponding liver inflammation grading results. By using statistical analysis or machine learning methods, such as logistic regression, decision trees, and random forests, the sample data is trained to enable the model to learn the relationship between the evaluation index and the liver inflammation grading, thereby determining the specific form and parameters of the computational model.

[0077] Through the process of constructing and determining the computational model described above, the calculation expression for the joint value can be obtained. This expression combines the various evaluation indicators through mathematical operations (such as weighted summation, product, etc.) to reflect their comprehensive impact on the grading of liver inflammation.

[0078] The advantages of using the above method to determine the calculation expression for the joint value include:

[0079] Data-driven personalized assessment: A computational model is built based on a sample dataset, which can automatically determine the calculation expression for joint values ​​according to actual data. This data-driven approach makes the assessment more personalized, better adaptable to the differences among different populations and individuals, and improves the accuracy of liver inflammation grading.

[0080] Improving model reliability and repeatability: Training and validating models using sample datasets ensures their reliability and repeatability. Compared to traditional empirical formulas or expert judgment, data-driven models more objectively reflect reality, reduce the influence of subjective factors, and improve the consistency of evaluation results.

[0081] Flexibility to adapt to different clinical scenarios: Different computational models can be built based on different sample datasets, making them suitable for various clinical scenarios. For example, models can be built separately for different types of liver inflammation (such as chronic hepatitis B, metabolic-associated fatty liver disease, drug-induced hepatitis, etc.), improving the specificity and effectiveness of the models.

[0082] Facilitates model updates and optimization: As more sample data accumulates, the computational model can be continuously updated and optimized to better reflect the latest clinical experience and research findings. This ability to continuously improve helps maintain the model's sophistication and effectiveness, adapting to evolving clinical needs.

[0083] It provides interpretable assessment evidence: by constructing a computational model, the calculation expression for the joint value is obtained, where the weight and role of each evaluation indicator can be clearly explained. This provides clinicians with a deeper understanding and interpretation of the assessment results, enhancing the transparency and credibility of the assessment process.

[0084] The following example, using a logistic regression model and ADC and SII values ​​as all evaluation indicators, illustrates how to obtain the expression for the joint value:

[0085] Logistic regression is a statistical model used for binary classification problems. It can predict the probability that a dependent variable (liver inflammation grade) belongs to a certain category based on given independent variables (evaluation indicators). In this approach, it constructs a predictive model by analyzing the relationship between ADC and SII values ​​and liver inflammation grade, which is used to calculate the joint value, thereby achieving a non-invasive assessment of liver inflammation grade.

[0086] Logistic regression models are suitable for handling binary classification problems, where the dependent variable is binary. In this approach, liver inflammation grading can be considered a binary classification problem. For example, inflammation grades ≥ G2 can be classified into one category, and inflammation grades < G2 into another. Therefore, logistic regression models are suitable for the prediction needs of this approach.

[0087] The methods for determining the joint value calculation expression include:

[0088] Based on logistic regression analysis: Logistic regression analysis is a statistical method used to analyze the relationship between independent variables and a binary dependent variable. In this scheme, logistic regression analysis is performed using ADC and SII values ​​as independent variables and liver inflammation grade (≥G2 or <G2) as the dependent variable to obtain the regression coefficients of the model. These regression coefficients reflect the degree of influence of each independent variable on the liver inflammation grade.

[0089] Constructing a predictive model: Based on the results of the logistic regression analysis, a predictive model was constructed. The model takes the form: Joint value = 11.732 + 3.860 × SII value - 13.834 × ADC value. In this model, 11.732 is the intercept term, and 3.860 and -13.834 are the regression coefficients of the SII and ADC values, respectively, representing the contribution of each independent variable to the joint value.

[0090] Model validation and optimization: The model's performance, such as accuracy, sensitivity, and specificity, is evaluated by validating it on both training and validation sets. Based on the validation results, the model is optimized and tuned to improve its predictive and generalization abilities. For example, ROC curve analysis can be used to evaluate the model's classification performance and determine the optimal cutoff value for achieving the best classification results.

[0091] Typically, liver inflammation assessment models are built primarily based on adult data, which may not accurately reflect the characteristics and patterns of liver inflammation in children. This can lead to biases in assessing children, affecting the accuracy of diagnosis and treatment. Therefore, this application proposes the following solution:

[0092] The sample dataset also includes sample data from children;

[0093] Specifically, when the target human body is a child, the joint value is determined by a joint calculation expression based on a sample dataset that includes sample data of children.

[0094] The above exemplary embodiments further define the composition and application of the sample dataset, as follows:

[0095] Expanding the sample dataset means including sample data from children. Compared to a dataset containing only adult samples, the expanded sample dataset covers a wider age range and includes more children. This allows the model to learn the characteristics and patterns of liver inflammation in children during training, thereby improving the accuracy of the model in grading liver inflammation in children.

[0096] By incorporating data from children's samples, the resulting joint calculation expression better reflects the characteristics of liver inflammation in children. When grading liver inflammation in children, this expression allows for a more accurate assessment of their liver inflammation status, enabling personalized liver inflammation grading.

[0097] The technical advantages achieved by the above exemplary embodiments include:

[0098] Improving the accuracy of liver inflammation grading in children: Since children's livers differ from adults in terms of anatomical structure, physiological function, and response to inflammation, joint computational expressions specifically determined for sample datasets containing children's samples can more accurately capture the characteristics of liver inflammation in children and reduce the risk of misjudgment caused by using adult models.

[0099] Improving the overall reliability and applicability of the model: By adding children's data to the sample dataset, the model training is more comprehensive, which can improve the reliability and applicability of the entire model when facing different age groups, making it more valuable in clinical applications.

[0100] Facilitating Early Diagnosis and Treatment of Liver Diseases in Children: Accurate liver inflammation grading is crucial for the early diagnosis and treatment of liver diseases in children. By using joint computational expressions determined specifically for sample datasets containing pediatric samples, doctors can develop appropriate treatment plans for pediatric patients more promptly and accurately, improving treatment outcomes.

[0101] Optimizing management strategies for childhood liver diseases: In the management of childhood liver diseases, accurate inflammation grading helps doctors better understand disease progression, rationally arrange follow-up and monitoring plans, optimize management strategies, and improve the quality of life of children.

[0102] In practical applications, there are differences in the image generation process between different MRI equipment manufacturers and MRI equipment with different field strengths. These differences may affect the measurement results of ADC value and SII value, and thus affect the accuracy of liver inflammation grading.

[0103] Specifically, MRI devices with different field strengths (such as 1.5T and 3.0T) differ in signal intensity and image resolution, which may lead to variations in the measured ADC and SII values. Furthermore, MRI devices from different manufacturers may differ in hardware design, imaging techniques, and post-processing algorithms, which can also affect the measurement results.

[0104] To better correct these differences and ensure the reliability of the evaluation results, embodiments of this application propose the following solutions, including:

[0105] Acquire MRI configuration data, wherein the MRI configuration data includes at least one of the field strength used to generate the MRI image and the manufacturer information of the image generation device used to generate the MRI image;

[0106] Determine the cutoff value corresponding to the MRI configuration data;

[0107] When the combined value is greater than or equal to the cutoff value, the liver inflammation grade is determined to be ≥G2; otherwise, the liver inflammation grade is determined to be <G2.

[0108] In the exemplary embodiments described above, by establishing a database or mapping table containing different MRI configuration data and corresponding cutoff values, the cutoff values ​​can be determined quickly and accurately based on the MRI configuration data. This database can be built by collecting a large amount of clinical and experimental data, and continuously updated and optimized as data accumulates.

[0109] The technical advantages achieved by the above exemplary embodiments include:

[0110] Improving the accuracy of grading: This scheme can effectively correct the influence of different MRI equipment and field strength on the measurement results, thereby improving the accuracy of liver inflammation grading.

[0111] Ensuring comparability of results: This approach ensures data consistency and comparability when comparing data in multi-center studies or across different devices.

[0112] Enhancing the reliability of clinical applications: By using cutoff values ​​that match MRI configuration data, misjudgments caused by equipment differences can be reduced, thereby enhancing the reliability and credibility of clinical applications.

[0113] In practical applications, doctors need a method to quickly and accurately assess treatment effectiveness during patient care. Traditional assessment methods may require lengthy observations and multiple tests, while obtaining changes in combined values ​​before and after treatment allows for a preliminary assessment of treatment efficacy in a shorter time.

[0114] Specifically, the combined values ​​of the target human body are acquired at time points before and after treatment. The efficacy of antiviral treatment is evaluated by comparing changes in these combined values ​​before and after treatment. The specific implementation method is as follows:

[0115] Before treatment, patients underwent MRI scans to obtain ADC and SII values, and the combined value was calculated as the baseline value before treatment. At T1 and T2 time points after treatment, MRI scans were performed again to obtain new ADC and SII values, and the new combined value was calculated.

[0116] The combined values ​​after treatment were compared with those before treatment. If the combined values ​​after treatment were lower than those before treatment, it indicated that the liver inflammation had been relieved and the antiviral treatment was effective for the patient; conversely, if the combined values ​​did not decrease significantly or increased, it may indicate that the treatment effect was not good and the treatment plan needed to be adjusted.

[0117] The technical advantages achieved by the above exemplary embodiments include:

[0118] Early efficacy assessment: This method can provide efficacy assessment in a short period of time after treatment, enabling doctors to understand the treatment effect in a timely manner and provide a basis for adjusting the subsequent treatment plan.

[0119] Non-invasive: Compared with traditional methods, this method is non-invasive and safe, and does not require multiple liver biopsies, reducing patient pain and risks.

[0120] Objective quantification: By comparing the combined values, objective efficacy assessment indicators are provided, avoiding errors from subjective judgment.

[0121] Personalized treatment: This helps to achieve personalized treatment, adjusting the treatment plan in a timely manner based on the patient's treatment response, thereby improving the success rate of treatment and the patient's survival rate.

[0122] The predictive model is used to synthesize information from all evaluation indicators to determine the probability that the target human liver inflammation grade is ≥G2. It integrates these indicators into a unified scoring system by assigning weights and scoring scales to each indicator.

[0123] In practical applications, based on a large amount of clinical data, statistical analysis or machine learning methods (such as logistic regression, decision trees, etc.) can be used to determine the weight and scoring scale of each evaluation indicator. The weight represents the relative importance of each evaluation indicator in predicting the grade of liver inflammation. For example, the SII value may be assigned a larger weight because it is highly sensitive to inflammation. The scoring scale is used to convert the actual measured value of each evaluation indicator into a score. For example, the SII value scoring scale might be designed to range from 0 to 100 based on its correlation with liver inflammation.

[0124] The total score is obtained by summing the scores of each evaluation indicator. This total score integrates information from all evaluation indicators and reflects the overall status of liver inflammation. Based on the total score, a pre-determined mapping relationship (such as a mapping table determined through ROC curve analysis or a calculated expression) is used to convert the total score into the probability of liver inflammation grade ≥G2.

[0125] The technical effects achieved by the above exemplary embodiments include:

[0126] Provides quantitative assessment: By converting evaluation indicators into scores and summing them to obtain a total score, the final score is mapped to a probability, thereby achieving a quantitative assessment of liver inflammation grading and improving the accuracy and objectivity of the assessment.

[0127] Supporting clinical decision-making: It provides clinicians with more accurate information, which helps in developing personalized treatment plans and prognostic assessments.

[0128] Applicable to different clinical scenarios: Different evaluation indicators can be selected according to specific needs, making it suitable for the assessment of various types of liver inflammation.

[0129] In practical applications, calibration curves and clinical decision curves (DCA) can be used to validate the prediction model.

[0130] The calibration curve, derived from the relationship between predicted probabilities (e.g., nomograms) and actual observed event rates, is primarily used to verify the accuracy of the prediction model and ensure consistency between the model's predicted probabilities and actual observed event rates. By comparing the model's predicted recurrence probabilities with actual recurrence events, the calibration curve visually demonstrates the model's predictive performance. For example, in hepatitis inflammation grading, if the prediction model estimates an 80% probability that a patient's inflammation grade is ≥G2, the calibration curve will show what percentage of patients with a predicted probability of 80% actually achieved an inflammation grade ≥G2. This helps assess the model's reliability; if the curve is close to a 45-degree diagonal (ideally), it indicates accurate model prediction.

[0131] The clinical decision curve (DCA) is calculated based on the net benefit of the predictive model at different decision thresholds. It helps clinicians evaluate the value of the predictive model in actual clinical decision-making, particularly whether using the model at different thresholds brings net benefit to patients. By analyzing the net benefit of the predictive model at different decision thresholds, doctors can determine under what circumstances using the predictive model will bring the greatest clinical benefit to patients. For example, in hepatitis inflammation grading, DCA can show the net benefit of using the predictive model to guide the decision on whether to administer antiviral therapy at different probability thresholds (e.g., 50%, 60%). If DCA shows that the model's net benefit is higher at a certain threshold than at other thresholds, doctors can use this information to select the most appropriate decision threshold to maximize patient treatment outcomes and reduce unnecessary treatments.

[0132] Figure 2 This is a schematic diagram of a calibration curve for predicting ≥G2 grade liver inflammation in CHB patients based on a combined index of ADC and SII values, provided in an embodiment of this application. Figure 2As shown, the horizontal axis represents the predicted probability, indicating the probability that the prediction model predicts the occurrence of the target event (e.g., liver inflammation grade ≥ G2). The vertical axis represents the actual proportion, indicating the proportion of the target event that actually occurs.

[0133] Figure 2 In the results shown, B = 500 repetitions, boot: indicates that 500 bootstrap samplings were performed to evaluate the stability and accuracy of the model; Mean absolute error = 0.028, n = 100: indicates that the mean absolute error is 0.028 and the sample size is 100. The mean absolute error measures the average difference between the predicted and actual values. The smaller the value, the more accurate the model prediction.

[0134] exist Figure 2 In the diagram, the black dashed line represents the model prediction curve under ideal conditions, which is the reference line where the predicted probability is exactly the same as the actual occurrence rate; the red dashed line represents the uncorrected model prediction curve, reflecting the model's performance on the original data; and the blue solid line represents the model prediction curve after bias correction, which corrects the optimistic bias of the model through methods such as guided methods, making the prediction results closer to the real situation.

[0135] from Figure 2 As can be seen, the calibration curve shows that the predicted probability of the prediction model is in good agreement with the actual event occurrence rate.

[0136] Figure 3 This is a schematic diagram illustrating a clinical decision curve for predicting ≥G2 grade liver inflammation in CHB patients based on a combined index of ADC and SII values, as provided in an embodiment of this application. Figure 3 As shown, the horizontal axis represents the threshold probability, which is the probability threshold that doctors or patients use to make treatment decisions. For example, if the threshold probability is 0.2, it means that when the model predicts a liver inflammation grade ≥G2 with a probability greater than 20%, corresponding treatment measures will be taken. The vertical axis represents the net benefit: representing the overall clinical benefit of the model at a specific threshold after considering true positive, false positive, and false negative results.

[0137] The red curve represents the net benefit when using the predictive model; the cyan curve represents the net benefit when all patients are considered positive (i.e., all patients are considered to have ≥G2 grade liver inflammation); and the blue-green curve represents the net benefit when no patients are identified as positive (i.e., no patients are considered to have ≥G2 grade liver inflammation).

[0138] At a threshold probability of 6%, the net benefit of using the predictive model (red curve) begins to exceed that of the strategy assuming all patients are positive (cyan curve). This indicates that using the model for prediction can yield a greater net benefit when the threshold probability is higher than 6%.

[0139] At a threshold probability of 89%, the net benefit of using the predictive model intersects with the strategy assuming no patients are positive (blue-green curve). Beyond the threshold probability of 89%, the net benefit of using the model is no longer significant.

[0140] exist Figure 3 In the middle, the relevant region of the model is Figure 3 The red rectangle represents the area between 6% and 89% on the horizontal axis. In this area, the net benefit of using the predictive model is higher than the other two hypothetical strategies, meaning that within this threshold range, using the model can help make better clinical decisions. The blue rectangle represents the area where the model is irrelevant, meaning that within the horizontal axis ranges of 0% to 6% and 89% to 100%, the net benefit of using the predictive model is lower than the other hypothetical strategies. This indicates that within these threshold ranges, the model may not be the optimal decision-making tool.

[0141] from Figure 3 It is evident that using this predictive model yields significant net benefits within a threshold probability range of 6% to 89%. This implies that for predicting liver inflammation grades ≥G2, when the decision threshold falls within this range, using this model can help physicians make decisions that are more beneficial to patients, improving treatment outcomes and resource utilization efficiency.

[0142] This application embodiment also provides a storage medium storing a computer program, wherein the computer program is configured to execute the method described above when running.

[0143] This application also provides a non-invasive grading device for liver inflammation, including a memory and a processor. The memory stores a computer program, and the processor is configured to run the computer program to perform the method described above.

[0144] It will be understood by those skilled in the art that all or some of the steps, systems, or apparatuses disclosed above, and their functional modules / units, can be implemented as software, firmware, hardware, or suitable combinations thereof. In hardware implementations, the division between functional modules / units mentioned above does not necessarily correspond to the division of physical components; for example, a physical component may have multiple functions, or a function or step may be performed collaboratively by several physical components. Some or all components may be implemented as software executed by a processor, such as a digital signal processor or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit (ASIC). Such software may be distributed on a computer-readable medium, which may include computer storage media (or non-transitory media) and communication media (or transient media). As is known to those skilled in the art, the term "computer storage medium" includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information (such as computer-readable instructions, data structures, program modules, or other data). Computer storage media include, but are not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc (DVD) or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and can be accessed by a computer. Furthermore, it is well known to those skilled in the art that communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

Claims

1. A non-invasive grading method for liver inflammation, comprising: Obtain at least two evaluation metrics for the liver in the target human body, wherein the at least two evaluation metrics include the apparent diffusion coefficient (ADC) value and the T2-weighted imaging signal intensity index (SII) value of the same region in the magnetic resonance imaging (MRI) image of the liver; Calculate the joint value corresponding to the values ​​of all evaluation indicators, wherein the joint value is positively correlated with the SII value and negatively correlated with the ADC value; Based on the combined values, the liver inflammation grading result of the target human body is determined.

2. The method according to claim 1, characterized in that, The evaluation indicators also include the clinical indicators of the target human body.

3. The method according to claim 1 or 2, characterized in that, The process of obtaining the calculation expression for the joint value is as follows: A computational model was constructed by using the liver inflammation grading results as the dependent variable and all evaluation indicators as independent variables. The computational model is determined using a sample dataset to obtain the computational expression for the joint value. The sample dataset includes the values ​​of evaluation indicators and the corresponding liver inflammation grading results of the individuals in the case group and the control group.

4. The method according to claim 3, characterized in that, When the calculation model is a logistic regression model, if all evaluation indicators are ADC value and SII value, the calculation expression of the joint value is joint value = 11.732 + 3.860 × SII value - 13.834 × ADC value.

5. The method according to claim 3, characterized in that: The sample dataset also includes sample data from children; Specifically, when the target human body is a child, the joint value is determined by a joint calculation expression based on a sample dataset that includes sample data of children.

6. The method according to claim 1, characterized in that, The step of determining the liver inflammation grade of the target human body based on the combined value includes: Acquire MRI configuration data, wherein the MRI configuration data includes at least one of the field strength used to generate the MRI image and the manufacturer information of the image generation device used to generate the MRI image; Determine the cutoff value corresponding to the MRI configuration data; When the combined value is greater than or equal to the cutoff value, the liver inflammation grade is determined to be ≥G2; otherwise, the liver inflammation grade is determined to be <G2.

7. The method according to claim 1, characterized in that, The method further includes: The probability of the target human body having a liver inflammation grade ≥ G2 is determined using a predictive model; The prediction model is used to obtain a total score by summing the scores of all evaluation indicators of the target human body, and to map the probability of liver inflammation grade ≥ G2 based on the total score. Each evaluation indicator has its own weight and scoring scale.

8. The method according to claim 7, characterized in that, The method further includes: The prediction model was validated using calibration curves and clinical decision curves.

9. A storage medium, characterized in that, The storage medium stores a computer program, wherein the computer program is configured to execute the method described in any one of claims 1 to 8 when it is run.

10. A non-invasive grading device for liver inflammation, comprising a memory and a processor, characterized in that, The memory stores a computer program, and the processor is configured to run the computer program to perform the method as described in any one of claims 1 to 8.

Images