Clinical decision support device, sample analysis system, and liver cancer risk assessment method

Abstract

This application relates to a clinical decision support device, comprising: a parameter acquisition module configured to acquire the measured values ​​of each biomarker in a biomarker combination of a subject, wherein the biomarker combination includes at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and albumin; a risk assessment module configured to input the measured values ​​of each biomarker in the biomarker combination into a computational model, and obtain the output of the computational model as a liver cancer risk prediction result for the subject; and an output module configured to output the liver cancer risk prediction result for the subject. This application also relates to a sample analysis system and a method for assessing the liver cancer risk of a subject. This application enables improved assessment of the liver cancer risk of a subject.

Description

Technical Field

[0001] This application relates to the field of in vitro diagnostics, specifically to clinical decision support devices, sample analysis systems, and methods for assessing the risk of liver cancer in subjects. In particular, this application proposes a novel combination of liver cancer biomarkers and their applications. Background Technology

[0002] Early diagnosis and treatment of high-risk groups for liver cancer can significantly reduce the disease burden, advance the treatment window for liver cancer patients, and thus improve the 5-year survival rate of liver cancer patients.

[0003] Currently, the primary screening and monitoring method for high-risk groups of liver cancer mentioned in guidelines at all levels is ultrasound imaging combined with serum alpha-fetoprotein (AFP) testing. However, in practical application, data shows that this method misses one in three liver cancer patients. Although AFP is a commonly used serological biomarker for liver cancer diagnosis, its diagnostic sensitivity and specificity are not high, especially for early-stage and small liver cancers (lesions ≤3cm), with a detection rate of only 20%. Ultrasound alone has a sensitivity of only 45% for detecting early-stage liver cancer.

[0004] This shows that current screening and monitoring methods for high-risk groups of liver cancer still need to be optimized. Summary of the Invention

[0005] Against this backdrop, the objective of this application is to propose the application of a novel combination of liver cancer biomarkers that can improve the assessment of liver cancer risk in subjects.

[0006] To achieve the above objectives, the first aspect of this application provides a clinical decision support device, comprising:

[0007] The parameter acquisition module is configured to acquire the measured values ​​of each biomarker in a biomarker combination of subjects, which includes at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase and albumin.

[0008] The risk assessment module is configured to input the measurements of each biomarker in the biomarker combination into a computational model, and obtain the output of the computational model as the subject's liver cancer risk prediction result; and

[0009] The output module is configured to output the liver cancer risk prediction results for the subjects.

[0010] To achieve the above objectives, a second aspect of this application provides a sample analysis system, comprising:

[0011] An immunoassay analyzer is configured to test the blood of a subject to obtain the subject's alpha-fetoprotein (AFP) level and abnormal prothrombin level.

[0012] A biochemical analyzer is configured to test the blood of subjects to obtain their gamma-glutamyl transferase and albumin levels; and

[0013] A data processing device, including a non-transitory computer-readable storage medium storing computer-readable instructions and one or more processors, is configured to execute the computer-readable instructions to perform the following steps: acquiring the subject's alpha-fetoprotein (AFP) measurement, abnormal prothrombin measurement, gamma-glutamyl transferase (γ-glutamyltransferase) measurement, and albumin measurement; inputting the AFP measurement, abnormal prothrombin measurement, γ-glutamyl transferase (γ-glutamyltransferase) measurement, and albumin measurement into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result; and outputting the subject's liver cancer risk prediction result.

[0014] To achieve the above objectives, a third aspect of this application provides a method for assessing the risk of liver cancer in a subject, comprising:

[0015] Blood samples from the subjects were tested to obtain the measurements of each biomarker in the subject's biomarker combination, which included at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and albumin.

[0016] The measured values ​​of each biomarker in this biomarker combination are input into a computational model, and the output of the model is used as the prediction result of the subject's liver cancer risk; and

[0017] Output the predicted risk of liver cancer for the subjects.

[0018] To achieve the above objectives, a fourth aspect of this application provides a clinical decision support device, comprising:

[0019] The parameter acquisition module is configured to acquire the measured values ​​of each biomarker in a biomarker combination of subjects, the biomarker combination including at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase and alkaline phosphatase;

[0020] The risk assessment module is configured to input the measurements of each biomarker in the biomarker combination into a computational model, and obtain the output of the computational model as the subject's predicted liver cancer risk; and

[0021] The output module is configured to output the liver cancer risk prediction results for the subjects.

[0022] To achieve the above objectives, a fifth aspect of this application provides a method for assessing the risk of liver cancer in a subject, comprising:

[0023] Blood samples from the subjects were tested to obtain the measurements of each biomarker in the subject's biomarker combination, which included at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and alkaline phosphatase.

[0024] The measured values ​​of each biomarker in the biomarker combination are input into a computational model, and the output of the computational model is used as the subject's liver cancer risk prediction result, especially the liver cancer risk score; and

[0025] Output the predicted risk of liver cancer for the subject.

[0026] To achieve the above objectives, a sixth aspect of this application provides a clinical decision support device, comprising:

[0027] The parameter acquisition module is configured to acquire measurements of each biomarker in a biomarker combination of subjects, the biomarker combination including at least alpha-fetoprotein, abnormal prothrombin and alkaline phosphatase and optionally albumin;

[0028] The risk assessment module is configured to input the measurements of each biomarker in the biomarker combination into a computational model, and obtain the output of the computational model as the subject's predicted liver cancer risk; and

[0029] The output module is configured to output the liver cancer risk prediction results for the subjects.

[0030] To achieve the above objectives, a seventh aspect of this application provides a method for assessing the risk of liver cancer in a subject, comprising:

[0031] Blood samples from the subjects are tested to obtain measurements of each biomarker in the subject’s biomarker combination, which includes at least alpha-fetoprotein, abnormal prothrombin and alkaline phosphatase and optionally albumin.

[0032] The measured values ​​of each biomarker in this biomarker combination are input into a computational model, and the output of the model is used as the prediction result of the subject's liver cancer risk; and

[0033] Output the predicted risk of liver cancer for the subjects.

[0034] To achieve the above objectives, the eighth aspect of this application provides a method for assessing the risk of liver cancer in a subject, comprising:

[0035] Blood samples from the subjects were tested to obtain the measurements of each biomarker in the subject's biomarker combination, which consisted of alpha-fetoprotein, abnormal prothrombin, and albumin.

[0036] The measured values ​​of each biomarker in this biomarker combination are input into a computational model, and the output of the model is used as the prediction result of the subject's liver cancer risk; and

[0037] Output the predicted risk of liver cancer for the subjects.

[0038] To achieve the above objectives, the ninth aspect of this application provides a method for assessing the risk of liver cancer in a subject, comprising:

[0039] Blood samples from the subjects were tested to obtain the measurements of each biomarker in the subject's biomarker combination, which consisted of alpha-fetoprotein, abnormal prothrombin, and gamma-glutamyl transferase.

[0040] The measured values ​​of each biomarker in this biomarker combination are input into a computational model, and the output of the model is used as the prediction result of the subject's liver cancer risk; and

[0041] Output the predicted risk of liver cancer for the subjects.

[0042] In the technical solutions proposed in this application, a novel combination of liver cancer biomarkers is used to assess the liver cancer risk of subjects. This combination of liver cancer biomarkers includes at least alpha-fetoprotein (AFP), abnormal prothrombin, gamma-glutamyl transferase (GGT), and albumin; or it includes at least AFP, abnormal prothrombin, GGT, and alkaline phosphatase; or it includes at least AFP, abnormal prothrombin, and alkaline phosphatase; or it consists of AFP, abnormal prothrombin, and albumin; or it consists of AFP, abnormal prothrombin, and GGT. This results in superior efficacy in the auxiliary diagnosis of liver cancer, particularly in terms of significantly improved sensitivity and / or specificity. Attached Figure Description

[0043] The present application will now be described more clearly with reference to the embodiments and accompanying drawings. Through the detailed description of the embodiments of the present application, the above-mentioned and other advantages will become clear to those skilled in the art. The accompanying drawings are only for illustrating preferred embodiments and should not be considered as limiting the present application. In the accompanying drawings:

[0044] Figure 1 A schematic block diagram of a clinical decision support device according to some embodiments of this application is shown;

[0045] Figure 2 A schematic block diagram of a sample analysis system according to some embodiments of this application is shown;

[0046] Figure 3A flowchart illustrating a method for assessing the risk of liver cancer in a subject according to some embodiments of this application is shown.

[0047] Figure 4 The specific training process of a computational model according to some embodiments of this application is illustrated;

[0048] Figure 5 The feature importance scores for detecting hepatocellular carcinoma throughout the entire period are shown for each feature parameter;

[0049] Figure 6 The feature importance scores for detecting early-stage liver cancer are shown for each feature parameter;

[0050] Figure 7 The ROC curves of a first evaluation model, a second evaluation model, and a comparative model according to some embodiments of this application are shown for full-stage liver cancer.

[0051] Figure 8 The ROC curves of a third evaluation model according to some embodiments of this application for full-stage hepatocellular carcinoma are shown.

[0052] Figure 9 The ROC curves of the four-three evaluation model according to some embodiments of this application for full-stage hepatocellular carcinoma are shown.

[0053] Figure 10 The ROC curves of the fifth evaluation model according to some embodiments of this application for full-stage hepatocellular carcinoma are shown.

[0054] Figure 11 The ROC curves of the sixth evaluation model according to some embodiments of this application for full-stage hepatocellular carcinoma are shown.

[0055] Figure 12 The ROC curves of the seventh evaluation model according to some embodiments of this application for full-stage hepatocellular carcinoma are shown.

[0056] Figure 13 The ROC curves of a first evaluation model, a second evaluation model, and a comparative model according to some embodiments of this application are shown for early-stage liver cancer.

[0057] Figure 14 The ROC curves of a third evaluation model for early-stage liver cancer according to some embodiments of this application are shown.

[0058] Figure 15 The ROC curves of a 4x3 evaluation model for early-stage liver cancer according to some embodiments of this application are shown.

[0059] Figure 16 The ROC curves of the fifth evaluation model for early-stage liver cancer according to some embodiments of this application are shown.

[0060] Figure 17The ROC curves for a sixth evaluation model for early-stage liver cancer according to some embodiments of this application are shown; and

[0061] Figure 18 The ROC curves for a seventh evaluation model for early-stage liver cancer according to some embodiments of this application are shown. Detailed Implementation

[0062] The embodiments of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0063] It should be noted that the terms "first, second, and third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific order of objects. It can be understood that "first, second, and third" can be interchanged in a specific order or sequence where permitted.

[0064] Those skilled in the art will understand that, unless otherwise defined, all terms used herein (including technical and scientific terms) have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0065] Primary liver cancer mainly includes three different pathological types: hepatocellular carcinoma (HCC), intrahepatic cholangiocarcinoma (ICC), and combined hepatocellular-cholangiocarcinoma (cHCC-CCA). These three types differ significantly in their pathogenesis, biological behavior, pathological histology, treatment methods, and prognosis. HCC accounts for 75%–85%, while ICC accounts for 10%–15%. In this invention, liver cancer refers only to HCC.

[0066] There are two conventional staging schemes for liver cancer: one is the Barcelona Clinic Liver Cancer (BCLC) system, commonly used internationally; the other is the China Liver Cancer Staging (CNLC) system, which is developed based on China's specific national conditions and practical experience, taking into account the patient's physical condition, liver tumor status, and liver function. CNLC includes stages Ia, Ib, IIa, IIb, IIIa, IIIb, and IV. In the embodiments of this application, "early-stage liver cancer" refers to stages Ia and Ib in CNLC, and "full-stage liver cancer" refers to the sum of all CNLC stages.

[0067] In related technologies, abnormal prothrombin (PIVKA-II or DCP) and serum alpha-fetoprotein isoform (AFP-L3) are used as diagnostic markers for liver cancer.

[0068] Furthermore, in related technologies, a biomarker combination (hereinafter also referred to as the ASAP model) for liver cancer assessment has been constructed based on gender, age, AFP, and PIVKA-II. However, this biomarker combination still has the following shortcomings:

[0069] 1) The sensitivity of this biomarker combination for early-stage liver cancer is only 70%–76%;

[0070] 2) The application of this biomarker combination requires obtaining the patient's age and gender. In practical applications, considering medical information security, there may be limitations in obtaining the patient's age and gender.

[0071] 3) This biomarker combination was established for the Barcelona Clinical Stage of Liver Cancer (BCLC). There is currently no model based on the Chinese CNLC stage.

[0072] 4) Some early-stage HCC patients test negative for both AFP and PIVKA-II in serum marker tests, making them undetectable by existing screening methods.

[0073] Besides commonly used serum tumor markers AFP and abnormal prothrombin (PIVKA-II or DCP), liver function parameters in routine biochemical tests can reflect the state of the liver, and monitoring their changes helps in the early detection of potential liver damage and the risk of liver cancer. Coagulation function is an important indicator of liver function. Although coagulation parameters cannot be directly used for the diagnosis of liver cancer, they can reflect the severity and progression of liver disease in patients.

[0074] Based on this, embodiments of this application propose a new combination of liver cancer biomarkers constructed by combining serum biomarkers and biochemical or coagulation parameters to assess the liver cancer risk of subjects.

[0075] The liver cancer biomarker combination proposed in this application is particularly suitable for assessing the liver cancer risk in subjects whose liver ultrasound examination shows no lesions or lesions ≤3cm. Alternatively or additionally, the liver cancer biomarker combination proposed in this application is particularly suitable for assessing the liver cancer risk in subjects with chronic liver diseases, which preferably include chronic hepatitis B, chronic hepatitis C, alcoholic liver disease, and non-alcoholic fatty liver disease.

[0076] like Figure 1 As shown, this application embodiment first provides a clinical decision support device 100, which is used to assist clinicians in assessing the liver cancer risk of subjects. The clinical decision support device 100 includes a parameter acquisition module 110, a risk assessment module 120, and an output module 130. The parameter acquisition module 110 is configured to acquire the measured values ​​of each biomarker in the biomarker combination or liver cancer biomarker combination proposed according to this application embodiment of the subject. The risk assessment module 120 is configured to input the measured values ​​of each biomarker in the biomarker combination into a calculation model, and obtain the output of the calculation model as the subject's liver cancer risk prediction result. The output module 130 is configured to output the subject's liver cancer risk prediction result.

[0077] By using the combination of biomarkers or liver cancer biomarkers proposed in the embodiments of this application to assess the liver cancer risk of subjects, the detection rate of early liver cancer can be improved, the detection efficiency of liver cancer throughout the entire process can be enhanced, and doctors can be more effectively assisted in assessing liver cancer risk, enabling more patients to obtain radical treatment opportunities and survival benefits.

[0078] According to the first embodiment of this application, the proposed combination of biomarkers includes at least alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT) and albumin (ALB).

[0079] In a specific example, the proposed biomarker combination consists of AFP, PIVKA-II, γ-GT, and ALB; that is, the proposed biomarker combination includes only these four biomarkers.

[0080] Furthermore, the biomarker combination according to the first embodiment of this application may also include alkaline phosphatase (ALP). For example, the proposed biomarker combination consists of AFP, PIVKA-II, γ-GT, ALB, and ALP, or includes only these five biomarkers.

[0081] It should be noted that the inventors of this application have discovered through research that as long as the biomarker combination according to the first embodiment of this application includes alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT), and albumin (ALB), it is possible to significantly improve the assessment of liver cancer risk. However, this does not mean that the biomarker combination according to the first embodiment of this application only includes the four biomarkers AFP, PIVKA-II, γ-GT, and ALB, but may further include other biomarkers, such as coagulation parameters mentioned below, even if the improvement effect of other biomarkers on liver cancer risk assessment is limited. For example, the biomarker combination according to the first embodiment of this application may include alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT), albumin (ALB), and coagulation parameters, such as prothrombin time (PT) and / or fibrinogen (FIB).

[0082] According to the second embodiment of this application, the proposed biomarker combination includes at least alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT), and alkaline phosphatase (ALP). For example, the proposed biomarker combination consists of AFP, PIVKA-II, γ-GT, and ALP, or includes only these four biomarkers.

[0083] It should also be noted that the inventors of this application have discovered through research that as long as the biomarker combination according to the second embodiment of this application includes alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT), and alkaline phosphatase (ALP), it is possible to significantly improve liver cancer risk assessment. However, this does not mean that the biomarker combination according to the second embodiment of this application only includes the four biomarkers AFP, PIVKA-II, γ-GT, and ALP, but may further include other biomarkers, such as coagulation parameters mentioned below, even if the improvement effect of other biomarkers on liver cancer risk assessment is limited. For example, the biomarker combination according to the second embodiment of this application may include alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), gamma-glutamyl transferase (GGT or γ-GT), alkaline phosphatase (ALP), and coagulation parameters, such as prothrombin time (PT) and / or fibrinogen (FIB).

[0084] According to the third embodiment of this application, the proposed biomarker combination includes at least alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), and alkaline phosphatase (ALP). For example, the proposed biomarker combination consists of AFP, PIVKA-II, and ALP, or includes only these three biomarkers. Optionally, the biomarker combination according to the third embodiment of this application may also include albumin (ALB). For example, the proposed biomarker combination consists of AFP, PIVKA-II, ALP, and ALB, or includes only these four biomarkers.

[0085] It should also be noted that the inventors of this application have discovered through research that as long as the biomarker combination according to the third embodiment of this application includes alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), and alkaline phosphatase (ALP), it is possible to significantly improve liver cancer risk assessment. However, this does not mean that the biomarker combination according to the third embodiment of this application only includes AFP, PIVKA-II, and ALP, but may further include other biomarkers, such as coagulation parameters mentioned below, even if other biomarkers have limited effect on improving liver cancer risk assessment. For example, the biomarker combination according to the third embodiment of this application may include alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP), alkaline phosphatase (ALP), and coagulation parameters, such as prothrombin time (PT) and / or fibrinogen (FIB).

[0086] According to the fourth embodiment of this application, the proposed biomarker combination consists of alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP) and albumin (ALB), that is, the proposed biomarker combination only includes AFP, PIVKA-II and ALB.

[0087] According to the fifth embodiment of this application, the proposed biomarker combination consists of alpha-fetoprotein (AFP), abnormal prothrombin (PIVKA-II or DCP) and gamma-glutamyl transferase (GGT or γ-GT), that is, the proposed biomarker combination only includes AFP, PIVKA-II and γ-GT.

[0088] In some embodiments, the parameter acquisition module 110 can be configured to proactively send the measured values ​​of various biomarkers of the subject to the risk assessment module 120, or can be configured to send the measured values ​​of various biomarkers of the subject to the risk assessment module 120 based on a request from the risk assessment module 120. Similarly, the risk assessment module 120 can be configured to proactively send the subject's liver cancer risk prediction results to the output module 130, or can be configured to send the subject's liver cancer risk prediction results to the output module 130 based on a request from the output module 130.

[0089] In some embodiments, the liver cancer risk prediction results output by the computational model can be quantitative results, such as liver cancer risk score, liver cancer risk probability, or liver cancer risk index.

[0090] In other embodiments, the liver cancer risk prediction results output by the computational model can be qualitative results, such as risk grading (high risk, medium risk, low risk). This risk grading can be obtained, for example, based on the quantitative results described above.

[0091] As one implementation method, the liver cancer risk prediction result can be obtained in the form of a liver cancer risk score. That is, the risk assessment module 120 can be configured to calculate the subject's liver cancer risk score as the liver cancer risk prediction result by using a computational model from the measurements of each biomarker in the proposed biomarker combination. Accordingly, the output module 130 can also be configured to: compare the subject's liver cancer risk score with a risk threshold; and when the subject's liver cancer risk score is greater than a preset risk threshold, output a prompt indicating an increased risk of liver cancer for the subject, and optionally also output suggested clinical treatment measures. This prompt can be, for example, an upward arrow or a text prompt.

[0092] As an alternative implementation, the output module 130 can also be configured to output a liver cancer risk score and its preset risk threshold. This allows physicians to assess a subject's liver cancer risk based on the output liver cancer risk score and the corresponding threshold.

[0093] In some embodiments, the clinical decision support device 100 further includes a threshold setting module configured to set a preset risk threshold for determining whether a subject is likely to have liver cancer. For example, the manufacturer or physician can manually adjust the preset risk threshold through the threshold setting module so that the threshold can be matched to a specific use case.

[0094] In some embodiments, the aforementioned biomarker combination can also be combined with other information of the subject, such as age and gender, to assess the subject's liver cancer risk. That is, the parameter acquisition module 110 can also be configured to acquire at least one of the subject's age and gender. Accordingly, the risk assessment module 120 can also be configured to input the subject's age and / or gender, along with the measured values ​​of each biomarker in the biomarker combination according to any one of the first to fifth embodiments of this application, into the calculation model to obtain the output of the calculation model as the subject's liver cancer risk prediction result.

[0095] For example, the risk assessment module 120 can also be configured to input the subject's age and / or sex along with alpha-fetoprotein (AFP) measurements, abnormal prothrombin measurements, gamma-glutamyl transferase (γ-glutamyltransferase) measurements, and albumin measurements into the calculation model to obtain the output of the calculation model as the subject's liver cancer risk prediction result. Alternatively, the risk assessment module 120 can also be configured to input the subject's age and / or sex along with AFP measurements, abnormal prothrombin measurements, gamma-glutamyl transferase (γ-glutamyltransferase) measurements, and alkaline phosphatase (ALP) measurements into the calculation model to obtain the output of the calculation model as the subject's liver cancer risk prediction result. Alternatively, the risk assessment module 120 can also be configured to input the subject's age and / or sex along with AFP measurements, abnormal prothrombin measurements, and ALP measurements into the calculation model to obtain the output of the calculation model as the subject's liver cancer risk prediction result. Alternatively, the risk assessment module 120 can be configured to input the subject's age and / or sex along with alpha-fetoprotein (AFP) measurements, abnormal prothrombin (AP) measurements, and albumin measurements into the computational model to obtain the model's output as the subject's liver cancer risk prediction result.

[0096] In some alternative embodiments, the clinical decision support device 100 proposed in this application does not use information such as the subject's age and / or gender. That is, the clinical decision support device 100 can be configured to not use the subject's age and / or gender as input to the computational model when obtaining the subject's liver cancer risk prediction results using the computational model.

[0097] In other embodiments, the clinical decision support device 100 may be configured to not use any of the subject's routine blood parameters as input to the computational model when obtaining the subject's liver cancer risk prediction results with the aid of a computational model.

[0098] In particular, the clinical decision support device 100 proposed in this application uses only the proposed combination of biomarkers to obtain the liver cancer risk prediction results for the subjects.

[0099] As some implementation methods, the clinical decision support device 100 proposed in the embodiments of this application uses only the combination of biomarkers proposed according to the first embodiment of this application to obtain the liver cancer risk prediction results of the subject, for example, only the biomarkers composed of AFP, PIVKA-II, γ-GT and ALB are used to obtain the liver cancer risk prediction results of the subject.

[0100] In some other implementations, the clinical decision support device 100 proposed in the embodiments of this application uses only the combination of biomarkers proposed according to the second embodiment of this application to obtain the liver cancer risk prediction results of the subject, for example, only the biomarkers composed of AFP, PIVKA-II, γ-GT and ALP are used to obtain the liver cancer risk prediction results of the subject.

[0101] As another implementation, the clinical decision support device 100 proposed in the embodiments of this application uses only the combination of biomarkers proposed according to the third embodiment of this application to obtain the liver cancer risk prediction results of the subject, for example, only the biomarkers composed of AFP, PIVKA-II and ALP are used to obtain the liver cancer risk prediction results of the subject.

[0102] As another implementation, the clinical decision support device 100 proposed in this application uses only the combination of biomarkers consisting of AFP, PIVKA-II and ALB as proposed in the fourth embodiment of this application to obtain the liver cancer risk prediction results of the subject.

[0103] As another implementation, the clinical decision support device 100 proposed in this application uses only the combination of biomarkers consisting of AFP, PIVKA-II and γ-GT proposed according to the fifth embodiment of this application to obtain the liver cancer risk prediction results of the subject.

[0104] In some embodiments, the computational model can be a pre-trained machine learning model, particularly a neural network model. It is understood that the machine learning model is trained using measurements of individual biomarkers from a combination of biomarkers from multiple known samples. Preferably, the multiple known samples are respectively derived from patients newly diagnosed with liver cancer and benign liver disease at their first visit.

[0105] As one implementation method, machine learning models can be built based on the random forest algorithm.

[0106] As another implementation, machine learning models can be built based on the support vector machine classification algorithm (SVM algorithm).

[0107] As another implementation method, machine learning models can be built based on logistic regression algorithms.

[0108] In the embodiments of this application, the various modules 110, 120, and 130 of the clinical decision support device 100 can be implemented in software and / or hardware. In particular, the parameters or code of the computational model, especially the machine learning model, can be deployed in a cloud environment or on a local analyzer.

[0109] In some embodiments, the clinical decision support device 100 may be implemented as middleware.

[0110] In some embodiments, the clinical decision support device 100 may be integrated into a laboratory information system (LIS) or a hospital information system (HIS) in the form of software.

[0111] In other embodiments, the clinical decision support device 100 may be integrated into the LIS or HIS system in the form of hardware, such as a processor or computer-readable storage medium, with its respective modules 110, 120 and 130 implemented in the hardware in the form of code or software.

[0112] In some other embodiments, the clinical decision support device 100 may also be a stand-alone hardware device.

[0113] This application also provides a sample analysis system 200, such as... Figure 2 As shown, the sample analysis system 200 includes an immunoassay analyzer 210, a biochemical analyzer 220, and a data processing device 230.

[0114] The immunoassay analyzer 210 is configured to test the blood of a subject to obtain the subject's alpha-fetoprotein (AFP) level and abnormal prothrombin level. The immunoassay analyzer 210 may be, for example, a chemiluminescent immunoassay analyzer.

[0115] Here, commercially available alpha-fetoprotein (AFP) test kits and abnormal prothrombin test kits can be used in an immunoassay analyzer 210, such as a chemiluminescence immunoassay analyzer, to test the subject's blood sample.

[0116] The biochemical analyzer 220 is configured to test the blood of a subject to obtain at least one of the subject's gamma-glutamyl transferase, albumin, and alkaline phosphatase values.

[0117] Here, commercially available γ-glutamyltransferase, albumin, and alkaline phosphatase assay kits can be used to test the subject's blood samples in the biochemical analyzer 220.

[0118] The data processing device 230 includes a non-transitory computer-readable storage medium 231 storing computer-readable instructions and one or more processors 232. Here, the processors 232 are configured to execute the computer-readable instructions to perform the following steps: acquiring measurements of each biomarker in a biomarker combination according to one of the first to fifth embodiments of this application for the subject; inputting the measurements of each biomarker into a computational model to obtain the output of the computational model as a liver cancer risk prediction result for the subject, particularly a liver cancer risk score; and outputting the liver cancer risk prediction result for the subject.

[0119] For example, processor 232 is configured to execute computer-readable instructions to perform the following steps: acquiring the subject's alpha-fetoprotein (AFP) value, abnormal prothrombin value, gamma-glutamyl transferase (γ-glutamyl transferase) value, and albumin value; inputting the AFP value, abnormal prothrombin value, γ-glutamyl transferase (γ-glutamyl transferase) value, and albumin value into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result, especially the liver cancer risk score; and outputting the subject's liver cancer risk prediction result.

[0120] For example, processor 232 is configured to execute computer-readable instructions to perform the following steps: acquiring the subject's alpha-fetoprotein (AFP) value, abnormal prothrombin value, gamma-glutamyl transferase (γ-glutamyltransferase) value, and alkaline phosphatase (ALP) value and inputting them into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result, especially the liver cancer risk score; and outputting the subject's liver cancer risk prediction result.

[0121] For example, processor 232 is configured to execute computer-readable instructions to perform the following steps: acquiring only the subject's alpha-fetoprotein (AFP) measurement, abnormal prothrombin measurement, and albumin measurement and inputting them into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result, especially the liver cancer risk score; and outputting the subject's liver cancer risk prediction result.

[0122] For example, processor 232 is configured to execute computer-readable instructions to perform the following steps: acquiring only the subject's alpha-fetoprotein (AFP) measurement, abnormal prothrombin measurement, and gamma-glutamyl transferase (γ-glutamyltransferase) measurement and inputting them into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result, especially the liver cancer risk score; and outputting the subject's liver cancer risk prediction result.

[0123] In some embodiments, the processor includes, but is not limited to, devices for interpreting computer instructions and processing data in computer software, such as a central processing unit (CPU), a micro controller unit (MCU), a field-programmable gate array (FPGA), and a digital signal processor (DSP).

[0124] In some embodiments, the data processing device 230 may be independent of the LIS system, HIS system, immunoassay analyzer, and biochemical analyzer, or it may be integrated into one of the LIS system, HIS system, immunoassay analyzer, and biochemical analyzer.

[0125] In some embodiments, the data processing device 230 may include a clinical decision support device 100 according to any of the above embodiments.

[0126] In some embodiments, such as Figure 2 As shown, the immunoassay analyzer 210 and the biochemical analyzer 220 are connected to the LIS in communication, and the LIS system is in turn connected to the HIS in communication. At this point, a specific testing process of the sample analysis system 200 according to the embodiments of this application is as follows: The clinician prescribes a test order for the patient, specifying the combination of biomarkers to be tested. The HIS system transmits the testing requirements for the biomarker combination to the LIS system. The LIS system issues test instructions, which are received by the immunoassay analyzer 210 (e.g., a chemiluminescence analyzer) and the biochemical analyzer 220 to complete the testing of the corresponding biomarkers and transmit the measured values ​​of the corresponding biomarkers to the data processing device 230. The data processing device 230 may be integrated into the laboratory middleware or laboratory management system or be a standalone device. Then, the data processing device 230 completes the liver cancer risk assessment for the patient to obtain a liver cancer risk score and determines whether the risk is high or low based on the corresponding liver cancer risk threshold, while providing suggested clinical treatment measures. Subsequently, the data processing device 230 transmits the liver cancer risk score, liver cancer risk threshold, and clinical treatment measures to the LIS system and the HIS system, which are also reflected on the patient's test report. By deploying the data processing device 230 in middleware or laboratory intelligent management systems, reports can be sent directly to clinicians or patients in the test reports.

[0127] This application also provides a method 300 for assessing the risk of liver cancer in a subject, such as... Figure 3 As shown, method 300 includes steps S310, S320 and S330.

[0128] In step S310, a blood sample from the subject is tested to obtain the measured values ​​of each biomarker in the subject's biomarker combination.

[0129] Here, the biomarker combination may be a combination of biomarkers according to one of the first to fifth embodiments of this application.

[0130] Specifically, the proposed biomarker combination may include at least AFP, PIVKA-II, γ-GT, and ALB, and optionally ALP. Alternatively, the proposed biomarker combination may include at least AFP, PIVKA-II, γ-GT, and ALP. Alternatively, the proposed biomarker combination may include at least AFP, PIVKA-II, and ALP, and optionally ALB. Alternatively, the proposed biomarker combination may include only AFP, PIVKA-II, and ALB. Alternatively, the proposed biomarker combination may include only AFP, PIVKA-II, and γ-GT.

[0131] In some embodiments, in step S310, the subject's blood sample is tested on an immunoassay analyzer and a chemical analyzer using a corresponding test kit to obtain the test results for the corresponding biomarkers.

[0132] In step S320, the measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the subject's liver cancer risk prediction result, especially the liver cancer risk score.

[0133] The computational model can be, for example, a machine learning model, such as one built on a random forest algorithm, a support vector machine classification algorithm (SVM algorithm), or a logistic regression algorithm.

[0134] In step S330, the liver cancer risk prediction results of the subject are output, and optionally, clinical treatment measures are output based on the liver cancer risk prediction results.

[0135] In some embodiments, in step S320, when obtaining the subject's liver cancer risk prediction results using the computational model, no blood routine parameters of the subject are used as input to the computational model.

[0136] In particular, in step S320, only the proposed combination of biomarkers is used to obtain the subject's liver cancer risk prediction results.

[0137] As some implementation methods, in step S320, only the combination of biomarkers proposed according to the first embodiment of this application is used to obtain the subject's liver cancer risk prediction results, for example, only the biomarkers composed of AFP, PIVKA-II, γ-GT and ALB are used to obtain the subject's liver cancer risk prediction results.

[0138] In some other implementations, in step S320, only the combination of biomarkers proposed according to the second embodiment of this application is used to obtain the subject's liver cancer risk prediction results, for example, only the biomarkers composed of AFP, PIVKA-II, γ-GT and ALP are used to obtain the subject's liver cancer risk prediction results.

[0139] As another implementation, in step S320, only the combination of biomarkers proposed according to the third embodiment of this application is used to obtain the subject's liver cancer risk prediction results, for example, only the biomarkers composed of AFP, PIVKA-II and ALP are used to obtain the subject's liver cancer risk prediction results.

[0140] In some other implementations, in step S320, only the combination of biomarkers consisting of AFP, PIVKA-II and ALB proposed according to the fourth embodiment of this application is used to obtain the liver cancer risk prediction results of the subject.

[0141] As another implementation, in step S320, only the combination of biomarkers consisting of AFP, PIVKA-II and γ-GT proposed according to the fifth embodiment of this application is used to obtain the liver cancer risk prediction results of the subject.

[0142] This application also provides a method for constructing a machine learning model for assessing the risk of liver cancer, including a data collection step, a feature variable screening step, and an algorithm training and evaluation step.

[0143] In the data collection step, sample data from multiple known samples are collected. These known samples are respectively from newly diagnosed HCC patients and patients with benign liver disease during their first visit. The sample data includes the patient's diagnostic results and the measured values ​​of each biomarker in the biomarker combination. The patient's diagnostic results include whether the patient has been diagnosed with liver cancer or benign liver disease. The biomarker combination may include, for example, AFP, PIVKA-II, γ-GT, ALP, ALB, CA125, CA199, CA50, HE4, CYFRA21-1, etc.

[0144] In the feature variable screening step: using the sample data of the multiple known samples, a combination of biomarkers according to one of the first to fifth embodiments of this application is selected as the feature variable for constructing the machine learning model.

[0145] In the algorithm training and evaluation steps, the measured values ​​of each biomarker in the selected biomarker combination of the multiple known samples and the patient's diagnostic results are used to train and validate the machine learning model in order to construct a machine learning model for assessing the risk of liver cancer.

[0146] In some embodiments, during the algorithm training and evaluation steps, a machine learning model can be built based on a random forest algorithm, a support vector machine classification algorithm, or a logistic regression algorithm.

[0147] This application also provides an application of a clinical decision support device 100 in the manufacture of a sample analysis system for assessing the risk of liver cancer in a specific subject. The sample analysis system includes an immunoassay analyzer and a biochemical analyzer. The immunoassay analyzer is configured to test the subject's blood to obtain the subject's alpha-fetoprotein (AFP) and abnormal prothrombin (APP) values. The biochemical analyzer is configured to test the subject's blood to obtain at least one of the subject's gamma-glutamyl transferase (GGT), albumin, and alkaline phosphatase (ALP) values.

[0148] This application also provides the application of an alpha-fetoprotein (AFP) detection kit, an abnormal prothrombin detection kit, a gamma-glutamyl transferase (γ-glutamyl transferase) detection kit, and an albumin detection kit in the preparation of analytical reagents for assessing the liver cancer risk of a subject. The AFP, abnormal prothrombin, γ-glutamyl transferase, and albumin detection kits are used to test the subject's blood sample to obtain AFP, abnormal prothrombin, γ-glutamyl transferase, and albumin values. These values ​​are then input into a computational model, particularly a pre-trained machine learning model, to obtain the output of the computational model as the subject's liver cancer risk score. An increase in the subject's liver cancer risk score relative to a preset risk threshold is associated with an increased risk of liver cancer in the subject.

[0149] The following describes specific examples of the construction and verification of the computational model in the embodiments of this application.

[0150] Example 1: Screening Biomarkers

[0151] 1. Inclusion criteria for research subjects and establishment of the database

[0152] HCC group: newly diagnosed HCC patients aged ≥18 years, including those with various causes such as HBV / HCV infection, alcoholic liver disease (ALD), non-alcoholic fatty liver disease (NAFLD), etc.

[0153] Benign liver disease group: Patients with chronic liver diseases, including various types of liver diseases such as chronic hepatitis B (CHB), chronic hepatitis C (CHC), alcoholic liver disease (ALD), and non-alcoholic fatty liver disease (NAFLD).

[0154] Patient samples and staging diagnosis conclusions were collected, and immune, biochemical, and coagulation indicators were tested on the patient samples. The final patient and staging information is shown in Table 1.

[0155] Table 1 Patient and Stage Information

[0156] *Two cases were excluded during data analysis due to missing AFP test results.

[0157] A total of 34 candidate characteristic variables related to the incidence and progression of liver cancer were considered, and their data distribution is shown in Table 2. Among them, WBC represents white blood cell count, RBC represents red blood cell count, PLT represents platelet count, HB represents hemoglobin concentration, APTT represents activated partial thromboplastin time, FIB represents fibrinogen, INR represents international normalized ratio, PT represents prothrombin time, TT represents thrombin time, AFP represents alpha-fetoprotein, CⅣ represents type IV collagen, CA125 represents tumor-associated antigen CA125, CA19-9 represents carbohydrate antigen CA19-9, CA50 represents carbohydrate antigen 50, and CYFRA represents... 21-1 represents cytokeratin 19 fragment, FERR represents ferritin, HA represents hyaluronic acid, HE4 represents human epididymal protein 4, LN represents laminin, PⅢNP represents type III procollagen amino-terminal peptide, PIVKA-II represents abnormal prothrombin, ALB represents albumin, ALT represents alanine aminotransferase, AST represents aspartate aminotransferase, CR represents creatinine, DBIL represents direct bilirubin, γ-GT represents γ-glutamyl transferase, LDH represents lactate dehydrogenase, TBIL represents total bilirubin, TP represents total protein, ALP represents alkaline phosphatase, and A / G represents albumin / globulin ratio.

[0158] Table 2 Candidate feature variables and their related data

[0159] 2. Correlation analysis

[0160] Spearman correlation analysis was used to screen out co-correlated variables, where a correlation coefficient > 0.7 was considered a strong correlation. This resulted in the selection of strongly cross-correlated variables, as shown in Table 3. Strongly cross-correlated variables do not coexist in the model; that is, they are not simultaneously included as input variables in the computational model.

[0161] Table 3 Characteristic variables of strong cross-correlation

[0162] 3. Training and Establishment of the Computational Model

[0163] Patients with liver cancer were labeled as 1, and patients with benign liver disease were labeled as 0. To distinguish between patients with liver cancer and patients with benign liver disease, a binary classification task model was constructed. The random forest algorithm in machine learning was used for parameter selection and model training. Finally, based on feature importance scores and clinical relevance, 4 to 6 feature variables were selected as variables for calculating the model.

[0164] The dataset, composed of patient classification (i.e., liver cancer patients are labeled 1, and benign liver disease patients are labeled 0) and feature variable results, is divided into training and testing sets in a 7:3 ratio. The training set is used for training with a five-fold crossover method, while the testing set is only used to evaluate the model's generalization performance. The specific training process is as follows: Figure 4 As shown.

[0165] The random forest algorithm is used for model building, and the training process can be carried out in a conventional manner.

[0166] 4. Variable selection and model performance analysis

[0167] Here, model variables are selected based on two dimensions: the feature importance score for detecting full-stage hepatocellular carcinoma and the feature importance score for detecting early-stage hepatocellular carcinoma. The feature importance score for detecting full-stage hepatocellular carcinoma is as follows: Figure 5 As shown, the characteristic importance score for detecting early-stage liver cancer is as follows: Figure 6 As shown.

[0168] Based on the feature importance scores for detecting full-stage and early-stage hepatocellular carcinoma (HCC), AFP, PIVKA-II, ALB, ALP, and γ-GT were identified as effective feature variables for detecting both. The correlation between these feature variables and HCC development and progression, as well as clinicians' prescribing habits, was evaluated sequentially. Ultimately, seven feature variables were selected as the feature variables for constructing the HCC risk calculation model. These seven feature variables include: the first group of feature variables (AFP, PIVKA-II, ALB, and γ-GT), and the calculation model built based on this group is called the first assessment model; the second group of feature variables (AFP, PIVKA-II, ALB, γ-GT, age, and sex), and the calculation model built based on this group is called the second assessment model; the third group of feature variables (AFP, PIVKA-II, ALB, and γ-GT), and the calculation model built based on this group is called the third assessment model; the fourth group of feature variables (AFP, PIVKA-II, and ALP), and the calculation model built based on this group is called the third assessment model; and the fourth group of feature variables (AFP, PIVKA-II, and ALP), and the calculation model built based on this group is called the fifth assessment model. The fourth evaluation model; the fifth set of characteristic variables, AFP, PIVKA-II, and ALB, and the computational model built based on the fifth set of characteristic variables is called the fifth evaluation model; the sixth set of characteristic variables, AFP, PIVKA-II, and γ-GT, and the computational model built based on the sixth set of characteristic variables is called the sixth evaluation model; the seventh set of characteristic variables, AFP, PIVKA-II, ALB, and ALP, and the computational model built based on the seventh set of characteristic variables is called the seventh evaluation model; and the eighth set of characteristic variables, AFP, PIVKA-II, ALB, γ-GT, and ALP, and the computational model built based on the eighth set of characteristic variables is called the eighth evaluation model.

[0169] Example 2: Validation of the diagnostic and therapeutic efficacy of the liver cancer risk assessment model

[0170] The diagnostic and therapeutic efficacy of the eight assessment models were validated using the test set described above (which included 257 cases).

[0171] Here, ROC curves and the area under the curve (AUC) are used to evaluate the effectiveness of each evaluation model. The term "ROC curve (receiver operator characteristic curve)" used in this application refers to a receiver operating characteristic curve, which is a curve plotted with the true positive rate on the ordinate and the false positive rate on the abscissa based on a series of different binary classification methods (cutoff thresholds). ROC_AUC (area under the curve) represents the area enclosed by the ROC curve and the horizontal coordinate axis.

[0172] The principle behind creating an ROC curve is to set multiple different critical values ​​for a continuous variable, calculate the corresponding sensitivity and specificity at each critical value, and then plot a curve with sensitivity as the vertical axis and 1-specificity as the horizontal axis.

[0173] Since the ROC curve is composed of multiple critical values ​​representing their respective sensitivity and specificity, it can be used to select the optimal diagnostic threshold for a particular diagnostic method. The closer the ROC curve is to the upper left corner, the higher the sensitivity and the lower the false positive rate, indicating better performance of the diagnostic method. It is known that the point on the ROC curve closest to the upper left corner has the highest sum of sensitivity and specificity; this point, or its neighboring points, is often used as a diagnostic reference value (also called a diagnostic threshold, judgment threshold, preset condition, or preset range).

[0174] In addition, the Delong test is used to determine whether there are significant differences between different assessment models. The smaller the p-value, the more significant the difference. Generally, a p-value < 0.05 is considered to indicate a difference between the two assessment models.

[0175] Here, the liver cancer risk assessment thresholds for the first to eighth assessment models and the ASAP model (constructed based on gender, age, AFP, and PIVKA-II) used as a comparative model are determined based on the maximum Youden index. The liver cancer risk assessment threshold for AFP (single parameter) used as a comparative model is 20 ng / mL, and the liver cancer risk assessment threshold for PIVKA-II (single parameter) used as a comparative model is 40 mAU / mL.

[0176] Table 4 shows the diagnostic efficacy of the above models for all stages of liver cancer. Tables 5 and 6 show the significant differences among the above models for all stages of liver cancer. Figures 7 to 12 The corresponding ROC curves and area under the curve (AUC) for each of the above models for full-stage liver cancer are shown.

[0177] Table 4. Diagnostic efficacy of various models for full-stage liver cancer

[0178] Table 5. Significant differences among the first assessment model, the second assessment model, and the comparative model for whole-stage liver cancer.

[0179] Table 6. Significant differences between the third to seventh assessment models and the ASAP model for whole-stage liver cancer.

[0180] From Tables 4 to 6 and Figures 7 to 12 As can be seen, for all stages of liver cancer, the various assessment models proposed in this application (i.e., the first to the eighth assessment models) are superior to existing assessment methods, and this advantage is statistically significant. Compared with traditional serum tumor markers, they significantly improve sensitivity while maintaining similar specificity, and their efficacy is superior to the ASAP model.

[0181] Table 7 shows the diagnostic efficacy of the above models for early-stage liver cancer. Tables 8 and 9 show the significant differences among the models for early-stage liver cancer. Figures 13 to 18 The ROC curves and area under the curve (AUC) for each of the above models for early-stage liver cancer are shown accordingly.

[0182] Table 7. Diagnostic efficacy of various models for early-stage liver cancer

[0183] Table 8. Significant differences in early-stage liver cancer among the first assessment model, the second assessment model, and the comparative model.

[0184] Table 9. Significant differences between the third to seventh assessment models and the ASAP model for early-stage liver cancer.

[0185] From Tables 7 to 9 and Figures 13 to 18 It is evident that for the highly occult early-stage liver cancer, the various assessment models proposed in this application (i.e., the first to the eighth assessment models) are superior to existing assessment methods, and this advantage is statistically significant. Compared to traditional serum tumor markers, they significantly improve sensitivity while maintaining similar specificity, and their efficacy is superior to the ASAP model. In particular, compared to the ASAP model, they can significantly improve specificity, even by more than 10%, without sacrificing sensitivity. To maintain the same 90% specificity as serum markers, the risk threshold can be further lowered, enabling the various assessment models proposed in this application to achieve even higher sensitivity in practical use, ensuring that more high-risk patients are detected.

[0186] In summary, the various assessment models proposed in this application innovatively combine traditional tumor markers and biochemical liver function test parameters, enabling more accurate and efficient detection of high-risk individuals for liver cancer, with performance superior to existing models. Furthermore, these tumor markers and biochemical liver function test parameters are readily available diagnostic parameters, and patients are typically prescribed these tests during their medical visits; therefore, this application does not incur additional medical costs.

[0187] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, embodiments of the present invention can take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware aspects. Furthermore, embodiments of the present invention can take the form of computer program products implemented on one or more computer-usable storage media (including disk storage and optical storage, etc.) containing computer-usable program code.

[0188] Embodiments of the present invention are described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program operations. These computer programs can be provided to operate on a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that operations performed by the processor of the computer or other programmable data processing device produce implementations in the flowchart. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0189] These computer program operations may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the operations stored in the computer-readable storage medium produce an article of manufacture including an operating device, the operating device being implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0190] These computer program operations can also be loaded onto a computer or other programmable data processing equipment, causing a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing the operations performed on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0191] All features or combinations of features mentioned above in the specification, drawings and claims may be used in any combination or individually, as long as they are meaningful within the scope of this invention and do not contradict each other.

[0192] The above description is merely a preferred embodiment of the present invention and does not limit the patent scope of the present invention. All equivalent transformations made based on the inventive concept of the present invention and the contents of the specification and drawings of the present invention, or direct / indirect applications in other related technical fields, are included within the patent protection scope of the present invention.

Claims

1. A clinical decision support device, comprising: The parameter acquisition module is configured to acquire the measured values ​​of each biomarker in the subject's biomarker combination, which includes at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and albumin. The risk assessment module is configured to input the measured values ​​of each biomarker in the biomarker combination into the calculation model, and obtain the output of the calculation model as the liver cancer risk prediction result of the subject. as well as The output module is configured to output the liver cancer risk prediction results of the subject.

2. The device according to claim 1, wherein, The risk assessment module is configured to use the computational model to calculate the subject's liver cancer risk score from the measurements of each biomarker in the biomarker combination as the liver cancer risk prediction result; The output module is also configured to: The liver cancer risk scores of the subjects were compared with risk thresholds; as well as When the subject's liver cancer risk score is greater than a preset risk threshold, an alert indicating an increased risk of liver cancer is output, and optionally, recommended clinical treatment measures are output. Optionally, the output module is also configured to output the preset risk threshold.

3. The device according to claim 1, wherein, The biomarker combination also includes alkaline phosphatase.

4. The device according to claim 1 or 2, wherein, The parameter acquisition module is also configured to acquire the subject's age and / or gender; and The risk assessment module is also configured to input the subject's age and / or gender, along with the alpha-fetoprotein (AFP) measurement, abnormal prothrombin (APP) measurement, gamma-glutamyl transferase (γ-glutamyltransferase) measurement, and albumin measurement, into the calculation model to obtain the output of the calculation model as the subject's liver cancer risk prediction result.

5. The device according to claim 1 or 2, wherein, The device is configured to not use the subject's age and / or gender as input to the computational model when obtaining the subject's liver cancer risk prediction results using the computational model.

6. The device according to any one of claims 1 to 5, wherein, The device is configured to not use any of the subject's routine blood parameters as input to the computational model when obtaining the subject's liver cancer risk prediction results using the computational model.

7. The device according to any one of claims 1 to 5, wherein, The subject's liver ultrasound examination showed no lesions or the lesions were ≤3cm; and / or the subject suffered from a chronic liver disease, preferably including chronic hepatitis B, chronic hepatitis C, alcoholic liver disease, or non-alcoholic fatty liver disease.

8. The device according to any one of claims 1 to 7, wherein, The computational model is a pre-trained machine learning model.

9. A sample analysis system, comprising: An immunoassay analyzer is configured to test the blood of a subject to obtain the subject's alpha-fetoprotein (AFP) level and abnormal prothrombin level. A biochemical analyzer is configured to test the blood of the subject to obtain the subject's gamma-glutamyl transferase and albumin levels; and A data processing device, including a non-transitory computer-readable storage medium storing computer-readable instructions and one or more processors, is configured to execute the computer-readable instructions to perform the following steps: acquiring the subject's alpha-fetoprotein (AFP) value, abnormal prothrombin value, gamma-glutamyl transferase (γ-glutamyltransferase) value, and albumin value; inputting the AFP value, abnormal prothrombin value, γ-glutamyl transferase (γ-glutamyltransferase) value, and albumin value into a computational model to obtain the output of the computational model as the subject's liver cancer risk prediction result, particularly a liver cancer risk score; And output the liver cancer risk prediction results of the subjects.

10. A method for assessing the risk of liver cancer in a subject, comprising: Blood samples from the subjects were tested to obtain the measured values ​​of each marker in the subject's biomarker combination, which included at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and albumin. The measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as Output the predicted risk of liver cancer for the subject.

11. The method according to claim 10, wherein, The biomarker combination also includes alkaline phosphatase.

12. The method according to claim 10 or 11, wherein, When using the computational model to obtain the predicted liver cancer risk of the subject, no routine blood parameters of the subject are used as input to the computational model.

13. The method according to any one of claims 10 to 12, wherein, The subject's liver ultrasound examination showed no lesions or the lesions were ≤3cm; and / or the subject suffered from a chronic liver disease, preferably including chronic hepatitis B, chronic hepatitis C, alcoholic liver disease, or non-alcoholic fatty liver disease.

14. A clinical decision support device, comprising: The parameter acquisition module is configured to acquire the measured values ​​of each biomarker in the subject's biomarker combination, which includes at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and alkaline phosphatase. The risk assessment module is configured to input the measured values ​​of each biomarker in the biomarker combination into the calculation model, and obtain the output of the calculation model as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as The output module is configured to output the liver cancer risk prediction results of the subject.

15. A method for assessing the risk of liver cancer in a subject, comprising: Blood samples from the subjects were tested to obtain the measured values ​​of each marker in the subject's biomarker combination, which included at least alpha-fetoprotein, abnormal prothrombin, gamma-glutamyl transferase, and alkaline phosphatase. The measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as Output the predicted risk of liver cancer for the subject.

16. A clinical decision support device, comprising: The parameter acquisition module is configured to acquire the measured values ​​of each biomarker in a biomarker combination of the subject, the biomarker combination including at least alpha-fetoprotein, abnormal prothrombin and alkaline phosphatase and optionally albumin; The risk assessment module is configured to input the measured values ​​of each biomarker in the biomarker combination into the calculation model, and obtain the output of the calculation model as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as The output module is configured to output the liver cancer risk prediction results of the subject.

17. A method for assessing the risk of liver cancer in a subject, comprising: Blood samples from the subject are tested to obtain the measured values ​​of each marker in the subject's biomarker combination, which includes at least alpha-fetoprotein, abnormal prothrombin, and alkaline phosphatase, and optionally also albumin. The measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as Output the predicted risk of liver cancer for the subject.

18. A method for assessing the risk of liver cancer in a subject, comprising: Blood samples from the subjects were tested to obtain the measured values ​​of each marker in the subject's biomarker combination, which consisted of alpha-fetoprotein, abnormal prothrombin, and albumin. The measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the prediction result of the subject's liver cancer risk, especially the liver cancer risk. as well as Output the predicted risk of liver cancer for the subject.

19. A method for assessing the risk of liver cancer in a subject, comprising: Blood samples from the subjects were tested to obtain the measured values ​​of each marker in the subject's biomarker combination, which consisted of alpha-fetoprotein, abnormal prothrombin, and gamma-glutamyl transferase. The measured values ​​of each biomarker in the biomarker combination are input into the calculation model, and the output of the calculation model is used as the liver cancer risk prediction result of the subject, especially the liver cancer risk score; as well as Output the predicted risk of liver cancer for the subject.

Images