G-GAAD model for early diagnosis of liver cancer as well as construction method and application of G-GAAD model

Abstract

The invention discloses a G-GAAD model for early diagnosis of liver cancer and a construction method and application thereof, and the construction method comprises the following steps: firstly collecting sample data of a subject, including an oligosaccharide chain detection value (GT), gender, age, a serum alpha fetoprotein detection value (AFP) and a serum abnormal prothrombin detection value (DCP); a data set is randomly divided into a training set and a verification set, GT, gender, age, AFP and DCP are used as independent variables, clinical diagnosis results are used as dependent variables, a G-GAAD model is constructed through a logistic regression method, and the verification set is used for verifying the performance of the model. According to the G-GAAD model disclosed by the invention, by integrating multi-dimensional information, the accuracy of liver cancer diagnosis is remarkably improved, and particularly, the missed diagnosis rate is effectively reduced in patients with negative GT, AFP and DCP. The model provides a new technical means for early diagnosis of liver cancer, and has important clinical application value and social significance.

Description

Technical Field

[0001] The present invention belongs to the technical field of biomedicine, and particularly relates to a G-GAAD model for early diagnosis of liver cancer, a construction method thereof, and an application thereof. Background Art

[0002] Liver cancer, as a highly malignant tumor, is an important health problem globally, with high incidence and mortality rates. Due to the lack of obvious early symptoms of liver cancer, most patients are in the middle and late stages at the time of diagnosis, resulting in poor prognosis. Although certain progress has been made in the diagnosis and treatment of liver cancer in recent years, the low early diagnosis rate and high postoperative recurrence rate are still the main bottlenecks restricting the improvement of the survival rate of liver cancer patients.

[0003] In China, the occurrence of liver cancer usually follows the "hepatitis - liver cirrhosis - liver cancer" trilogy model, and patients with liver cirrhosis are high-risk groups for liver cancer occurrence. Currently, the commonly used methods for liver cancer screening and early diagnosis in clinical practice mainly rely on imaging examinations (such as ultrasound, CT, MRI, etc.) combined with serum marker detection, among which alpha-fetoprotein (AFP) and des - gamma-carboxyprothrombin fragment (DCP) are the most commonly used serum markers. However, these traditional methods have obvious deficiencies in terms of sensitivity and specificity, especially in the diagnosis of early liver cancer, with a relatively high missed diagnosis rate.

[0004] In recent years, oligosaccharide chain marker (GT) as an emerging liver cancer diagnostic marker has shown certain clinical application potential. However, although GT shows relatively high diagnostic value in some liver cancer patients, in actual application, there are still a considerable proportion of GT-negative patients being missed diagnosed. Similarly, AFP and DCP, as traditional liver cancer markers, also have a relatively high false negative rate, especially in early liver cancer or certain special types of liver cancer. Relying solely on these markers for diagnosis often fails to meet the clinical needs. Summary of the Invention

[0005] Aiming at the deficiencies of the prior art, the present invention provides a G-GAAD model for early diagnosis of liver cancer, a construction method thereof, and an application thereof. By integrating multi-dimensional information such as GT, Gender, Age, AFP, and DCP, this model aims to improve the accuracy of liver cancer diagnosis, especially for patients with negative GT, AFP, and DCP, reduce the missed diagnosis rate, and provide a more reliable method for early diagnosis of liver cancer.

[0006] The present invention is achieved through the following technical solutions:

[0007] A construction method of a G-GAAD model for early diagnosis of liver cancer, comprising the following steps:

[0008] Step 1) Collect valid data of the samples of the subjects as a data set. The valid data includes the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value.

[0009] Step 2) Randomly divide the data set into a training set and a validation set. In the data of the training set, use the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value as independent variables, and use the clinical diagnosis result as the dependent variable. Establish a G-GAAD model for early diagnosis of liver cancer through logistic regression method, and then verify the performance of the G-GAAD model through the data of the validation set.

[0010] Preferably, the sample in Step 1) is the blood, serum, and plasma of the venous blood or peripheral blood of the subject.

[0011] Preferably, the oligosaccharide chain detection value is obtained by the following method:

[0012] First, add 2 μL of the subject's sample to a test tube, then add 5 μL of 5% SDS, mix well, and perform high-temperature denaturation treatment by heating at 95 °C for 5 min, and cool to 4 °C to obtain a sample; add 3 μL of 2.2 units / μL glycamidase to the sample, mix well and centrifuge, keep it at 37 °C in an incubator for 3 h, then cool to 4 °C, add 50 μL of deionized water to the cooled sample to terminate the reaction for 1 min, and store it at low temperature -20 °C; take 10 μL of the sample stored at low temperature and dry it in a metal bath for 90 min, then cool to 4 °C, add 2 μL of a mixed solution of 100 mM trisulfonic acid trisodium salt fluorescent label and 1 M organic reducing agent prepared according to a volume ratio of 1:1 to the dried sample, centrifuge, and keep it at 37 °C in an incubator for 16 h for fluorescent labeling, then cool to 4 °C, and then add 100 μL of deionized water to terminate the reaction for 1 min, mix well and centrifuge, and store it at low temperature -20 °C; take 2 μL of the sample after terminating the reaction and add 2 μL of sialidase, mix well and centrifuge, keep it at 37 °C in an incubator for 16 h, then add 40 μL of deionized water to terminate the reaction for 1 min, mix well and centrifuge, and take 10 μL of the sample for oligosaccharide chain fragment separation detection by a gene sequencer to obtain the oligosaccharide chain detection value.

[0013] For the G-GAAD model constructed by the above construction method, the G-GAAD model uses the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value of the subject as input variables, and the G-GAAD model calculates the score using the following formula:

[0014] G-GAAD score = exp(0.053 × age - 0.748 × gender + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335) / [1 + exp(0.053 × age - 0.748 × gender + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335)];

[0015] Wherein: male and female genders are represented by 0 and 1 respectively; GT is the oligosaccharide chain detection value; AFP is the serum alpha-fetoprotein detection value; DCP is the serum abnormal prothrombin detection value.

[0016] The application of the above G-GAAD model in the preparation of products for early diagnosis of liver cancer is to judge whether the subject has liver cancer through the optimal threshold. When the G-GAAD score ≥ the optimal threshold, it indicates that the subject has liver cancer, and the optimal threshold is 0.236.

[0017] Preferably, the liver cancer includes HBV-related liver cancer, HCV-related liver cancer, and non-viral liver cancer.

[0018] Preferably, the product is a kit or a chip.

[0019] A computer device includes a memory and a processor. The memory stores a computer program, and the processor executes the following steps:

[0020] Step 1) Obtain the valid data of the subject's sample. The valid data includes the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value;

[0021] Step 2) Input the valid data into the above G-GAAD model to calculate the G-GAAD score of the subject;

[0022] Step 3) Diagnose whether the subject has liver cancer according to the G-GAAD score.

[0023] A computer-readable storage medium stores a computer program, and the computer program is executed by the processor to perform the following steps:

[0024] Step 1) Obtain the valid data of the subject's sample. The valid data includes the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value;

[0025] Step 2) Input the valid data into the above G-GAAD model to calculate the G-GAAD score of the subject;

[0026] Step 3) Diagnose whether the subject has liver cancer according to the G-GAAD score.

[0027] A computer program product includes a computer program, and the computer program is executed by a processor to perform the following steps:

[0028] Step 1) Obtain valid data of a sample of a subject, where the valid data includes oligosaccharide chain detection values, gender, age, serum alpha-fetoprotein detection values, and serum abnormal prothrombin detection values;

[0029] Step 2) Input the valid data into the above G-GAAD model to calculate the G-GAAD score of the subject;

[0030] Step 3) Diagnose whether the subject has liver cancer according to the G-GAAD score.

[0031] The G-GAAD model provided by the present invention constructs a new liver cancer diagnosis model by integrating multi-dimensional information such as oligosaccharide chain marker (GT), gender, age, alpha-fetoprotein detection value (AFP), and abnormal prothrombin detection value (DCP), and has the following beneficial effects:

[0032] (1) The G-GAAD model of the present invention effectively solves the problem of missed diagnosis existing when GT, AFP, and DCP are used alone or in combination in the prior art. Especially in patients with negative GT, AFP, and DCP, the missed diagnosis rate is significantly reduced. By introducing clinical features such as gender and age, the G-GAAD model can more comprehensively reflect the individual differences of patients, thereby improving the accuracy and reliability of diagnosis.

[0033] (2) The G-GAAD model of the present invention has important value in improving the early diagnosis rate of liver cancer. Since the early symptoms of liver cancer are hidden and the sensitivity and specificity of traditional diagnosis methods are insufficient in the early stage, many patients miss the best treatment opportunity when they are diagnosed. The G-GAAD model can identify potential liver cancer patients in high-risk populations earlier through multi-index combined analysis, providing a valuable time window for early intervention and treatment, thereby improving the prognosis of patients.

[0034] (3) The G-GAAD model of the present invention has high clinical application feasibility. The model is based on existing clinical detection indicators (AFP, DCP) and easily accessible clinical information (gender, age), supplemented by the GT indicator, without introducing complex detection technologies or high costs, and is convenient for popularization and use in medical institutions at all levels. This not only helps to improve the popularity of liver cancer diagnosis, but also provides an efficient and economical diagnostic tool for areas with relatively scarce medical resources.

[0035] In summary, the G-GAAD model of the present invention significantly improves the accuracy of liver cancer diagnosis and reduces the missed diagnosis rate through multi-dimensional information integration. Its application value is particularly prominent, especially in patients negative for GT, AFP, and DCP. This model provides a new technical means for the early diagnosis of liver cancer and has important clinical significance and social value. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 ROC curve diagrams for the G-GAAD model, GT, AFP, and DCP in differentiating liver cancer patients from non-liver cancer patients in the training set;

[0037] Figure 2 ROC curve diagrams for the G-GAAD model, GT, AFP, and DCP in differentiating liver cancer patients from non-liver cancer patients in the validation set. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0038] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0039] Unless otherwise specified, the technical means used in the following embodiments are all conventional means well-known to those skilled in the art. The experimental methods without specific conditions mentioned are all conventional methods in the art.

[0040] The materials, reagents, etc. used in the following embodiments can be obtained from commercial channels unless otherwise specified.

[0041] Example 1

[0042] A method for constructing a G-GAAD model for the early diagnosis of liver cancer, the specific steps are as follows:

[0043] 1. Sample collection

[0044] In this example, relevant samples were collected from the Second People's Hospital of Tianjin, including blood samples of 150 patients with liver cirrhosis and 98 patients with liver cancer. The clinical diagnosis results of the above samples of patients have been determined by the methods recommended in the clinical practice guidelines. The following experiments have been reported to the ethics committee for filing and approval.

[0045] 2. Data detection

[0046] The effective data collected from the above samples was used as a data set. The effective data includes oligosaccharide chain detection value (GT), gender (Gender), age (Age), serum alpha-fetoprotein detection value (AFP), and serum abnormal prothrombin detection value (DCP).

[0047] The detection method of the oligosaccharide chain detection value in the blood sample can refer to Patent CN104807998B, specifically as follows:

[0048] ① Protein denaturation: First, add 2 μL of the subject's sample to a test tube, then add 5 μL of 5% SDS. After thorough mixing, perform high-temperature denaturation treatment by heating at 95 °C in a PCR for 5 min, and then cool to 4 °C to obtain a sample;

[0049] ② Glycosidase treatment: Add 3 μL of 2.2 units / μL glycamidase to the sample, mix well and centrifuge. Keep it at 37 °C in an incubator for 3 h, then cool to 4 °C. Add 50 μL of deionized water to the cooled sample to terminate the reaction for 1 min, and store it at -20 °C;

[0050] ③ Fluorescent labeling: Take 10 μL of the sample stored at low temperature and dry it in a metal bath for 90 min, then cool to 4 °C. Add 2 μL of fluorescent labeling mixture (prepared by mixing 100 mM trisulfonate sodium salt fluorescent label and 1 M organic reducing agent sodium borohydride in a volume ratio of 1:1) to the dried sample, centrifuge, and keep it at 37 °C in an incubator for 16 h for fluorescent labeling. Then cool to 4 °C, and add 100 μL of deionized water to terminate the reaction for 1 min. After mixing and centrifuging, store it at -20 °C;

[0051] ④ Oligosaccharide chain map detection: Take 2 μL of the sample after terminating the reaction and add 2 μL of sialidase, mix well and centrifuge. Keep it at 37 °C in an incubator for 16 h, then add 40 μL of deionized water to terminate the reaction for 1 min. After mixing and centrifuging, take 10 μL of the sample and perform oligosaccharide chain fragment separation detection using a gene sequencer.

[0052] ⑤ Data collection

[0053] After the blood sample undergoes the above protein denaturation, glycosidase treatment, fluorescent labeling, and oligosaccharide chain map detection, the oligosaccharide chain detection value (GT) is obtained.

[0054] Meanwhile, collect the patient's clinical information, including gender, age, and the detection results of AFP and DCP.

[0055] 3. Construction of the G-GAAD model

[0056] Randomly divide the dataset into a training set and a validation set. In the data of the training set, use GT, gender, age, AFP, and DCP as independent variables and the clinical diagnosis result as the dependent variable. Establish a G-GAAD model for early diagnosis of liver cancer through logistic regression, and then verify the performance of the G-GAAD model using the data of the validation set.

[0057] (1) Conduct an inter-group comparative analysis of the clinical indicators of patients with liver cirrhosis and liver cancer

[0058] The clinical information indexes of the above 150 patients with liver cirrhosis and 98 patients with liver cancer were compared and analyzed, and the indexes with a statistical result of P < 0.1 were selected as the factors included in the model, as shown in Table 1 below.

[0059] Table 1 Comparison results of each index data between patients with liver cirrhosis and liver cancer

[0060]

[0061] (2) Dataset division

[0062] The patients were randomly divided into a training set and a validation set in a ratio of 6:4. There were 90 patients with liver cirrhosis and 59 patients with liver cancer in the training set. There were 60 patients with liver cirrhosis and 39 patients with liver cancer in the validation set.

[0063] The selected indexes included gender, age, GT, AFP, and DCP, which were used as independent variables; the clinical diagnosis result was used as the dependent variable. In the training set, the logistic regression equation calculation method was used to construct the G-GAAD model. The calculation formula of the G-GAAD model is as follows:

[0064] G-GAAD score = exp(0.053 × age - 0.748 × gender + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335) / [1 + exp(0.053 × age - 0.748 × gender + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335)];

[0065] In the formula: gender male and female are represented by 0 and 1 respectively.

[0066] At the same time in the training set, the G-GAAD model, GT, AFP, and DCP were separately used as markers to distinguish patients with liver cancer and liver cirrhosis, the ROC curve was drawn, and the data of the validation set was used to verify the performance of the G-GAAD model. The results are as Figure 1 、 2 shown.

[0067] (3) Experimental results

[0068] As Figure 1 shown, the AUC values of the areas under the ROC curves of the G-GAAD model, GT, AFP, and DCP for distinguishing patients with liver cancer and liver cirrhosis in the training set were 0.930, 0.914, 0.694, and 0.823 respectively. As Figure 2As shown, the AUC values under the ROC curves of the G-GAAD model, GT, AFP, and DCP for differentiating liver cancer patients from liver cirrhosis patients in the validation set were 0.950, 0.922, 0.706, and 0.713, respectively. The above experimental results indicate that the diagnostic performance of the G-GAAD model of the present invention is superior to that of GT, AFP, and DCP.

[0069] Through ROC curve analysis, the Youden index determined the optimal threshold of the G-GAAD model to be 0.236. When the G-GAAD score ≥ 0.236, it indicates liver cancer, and when the G-GAAD score < 0.236, it indicates no liver cancer.

[0070] The comparisons of the sensitivity, specificity, coincidence rate, and AUC values of the G-GAAD model, GT, AFP, and DCP in the training set and the validation set are shown in Table 2 below.

[0071] Table 2 Sensitivity, specificity, coincidence rate, and AUC values of the G-GAAD model, GT, AFP, and DCP in the training set and the validation set

[0072]

[0073] As can be seen from Table 2, the coincidence rate of the G-GAAD model with clinical diagnosis in the training set reached 84.56%, which is superior to GT (81.21%), AFP (67.11%), and DCP (77.18%). The coincidence rate of the G-GAAD model with clinical diagnosis in the validation set reached 89.90%, which is superior to GT (85.86%), AFP (67.68%), and DCP (64.65%).

[0074] The sensitivity of the G-GAAD model in liver cancer patients negative for GT, AFP, and DCP is shown in Table 3 below.

[0075] Table 3 Sensitivity of the G-GAAD model in liver cancer patients negative for GT, AFP, and DCP

[0076] Clinical diagnosis results of liver cancer Sample size n Sensitivity of the G-GAAD model GT-negative liver cancer 14 64.29％(9 / 14) AFP-negative liver cancer 57 91.23％(52 / 57) DCP-negative liver cancer 39 89.74％(35 / 39) Liver cancer negative for both AFP and DCP 30 86.67％(26 / 30) Liver cancer negative for GT, AFP, and DCP 6 83.33％(5 / 6)

[0077] As can be seen from Table 3, in liver cancer patients negative for GT, AFP, and DCP respectively, the sensitivities of the G-GAAD model were 64.29%, 91.23%, and 89.74%, respectively. In addition, in patients negative for both AFP and DCP, the sensitivity of the G-GAAD model reached 86.67%; and in patients negative for GT, AFP, and DCP, the sensitivity of the G-GAAD model was still as high as 83.33%. The G-GAAD model of the present invention significantly reduces the missed diagnosis rate of liver cancer patients negative for GT, AFP, and DCP through multi-dimensional information integration, providing an efficient and reliable detection method for the early diagnosis of this part of patients, and having important clinical application value.

[0078] The embodiments described above are only a part of the embodiments of the present invention, rather than all embodiments. The detailed description of the embodiments of the present invention is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. The scope of protection of the present invention shall be subject to the scope claimed in the claims. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

Claims

1. A method for constructing a G-GAAD model for early diagnosis of liver cancer, characterized in that: The following steps are involved: Step 1) collecting valid data of the subject's sample as a data set, wherein the valid data includes oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value; Step 2) The data set is randomly divided into a training set and a validation set, wherein the oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value are used as independent variables in the training set data, and the clinical diagnosis result is used as the dependent variable. A G-GAAD model for early diagnosis of liver cancer is established by a logistic regression method, and the performance of the G-GAAD model is verified by the data of the validation set.

2. The method for constructing a G-GAAD model for early diagnosis of liver cancer according to claim 1, characterized in that: Step 1) The sample is venous blood or peripheral blood, serum and plasma of the subject.

3. The method for constructing a G-GAAD model for early diagnosis of liver cancer according to claim 2, characterized in that: The oligosaccharide chain detection value is obtained in the following manner: First, add 2 μL of the subject's sample to a test tube, then add 5 μL of 5% SDS, mix thoroughly, heat at 95°C for 5 minutes for high-temperature denaturation treatment, and cool to 4°C to obtain a sample; add 3 μL of 2.2 units / μL of glycosidase to the sample, mix and centrifuge, keep it in an incubator at 37°C for 3 hours, and then cool to 4°C. Add 50 μL of deionized water to the cooled sample to terminate the reaction for 1 minute, and store it at -20°C; take 10 μL of the low-temperature stored sample and dry it in a metal bath for 90 minutes, then cool it to 4°C, and add 2 μL of 100 mM trisulfonic acid trisulphonate prepared at a volume ratio of 1:1 to the dried sample. The mixture of sodium salt fluorescent marker and 1M organic reducing agent is centrifuged and kept in an incubator at 37°C for 16 hours for fluorescent labeling, then cooled to 4°C, and 100 μL of deionized water is added to terminate the reaction for 1 minute. After mixing and centrifugation, it is stored at -20°C. Take 2 μL of the sample after the termination of the reaction, add 2 μL of sialidase, mix and centrifuge, keep in an incubator at 37°C for 16 hours, and then add 40 μL of deionized water to terminate the reaction for 1 minute. After mixing and centrifugation, 10 μL of the sample is sampled for oligosaccharide chain fragment separation and detection through a gene sequencer.

4. The G-GAAD model constructed by the construction method according to any one of claims 1 to 3, characterized in that: The G-GAAD model uses the test value of oligosaccharide chains, gender, age, serum alpha-fetoprotein test value, and serum abnormal prothrombin test value of the subject as input variables, and the G-GAAD model uses the following formula to calculate the score: G-GAAD score = exp(0.053 × age - 0.748 × sex + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335) / [1 + exp(0.053 × age - 0.748 × sex + 0.965 × GT + 0.002 × AFP + 0.0002 × DCP - 8.335)]; Wherein: gender is represented by 0 for male and 1 for female respectively; GT is the oligosaccharide chain detection value; AFP is the serum alpha-fetoprotein detection value; DCP is the serum abnormal prothrombin detection value.

5. Use of the G-GAAD model according to claim 4 in preparing a product for early diagnosis of liver cancer, characterized in that: The optimal threshold is used to determine whether the subject has liver cancer. When the G-GAAD score ≥ the optimal threshold, it indicates that the subject has liver cancer. The optimal threshold is 0.

236.

6. The use according to claim 5, characterized in that: The liver cancer includes HBV-related liver cancer, HCV-related liver cancer and non-viral liver cancer.

7. The use according to claim 5, characterized in that: The product is a kit or a chip.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: The processor performs the following steps: Step 1) obtaining valid data of the subject's sample, wherein the valid data includes oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value; Step 2) inputting the valid data into the G-GAAD model as described in claim 4 to calculate the G-GAAD score of the subject; Step 3) diagnosing whether the subject has liver cancer based on the G-GAAD score.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: The computer program is executed by the processor to perform the following steps: Step 1) obtaining valid data of the subject's sample, wherein the valid data includes oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value; Step 2) inputting the valid data into the G-GAAD model as described in claim 4 to calculate the G-GAAD score of the subject; Step 3) diagnosing whether the subject has liver cancer based on the G-GAAD score.

10. A computer program product, comprising a computer program, characterized in that The computer program is executed by the processor to perform the following steps: Step 1) obtaining valid data of the subject's sample, wherein the valid data includes oligosaccharide chain detection value, gender, age, serum alpha-fetoprotein detection value, and serum abnormal prothrombin detection value; Step 2) inputting the valid data into the G-GAAD model as described in claim 4 to calculate the G-GAAD score of the subject; Step 3) diagnosing whether the subject has liver cancer based on the G-GAAD score.

Images