Chronic hepatitis B patient portal vein thrombosis prediction model based on virological treatment result and construction method

Abstract

The invention belongs to the technical field of medical data analysis, and relates to a chronic hepatitis B patient portal vein thrombosis prediction model based on virological treatment results and a construction method thereof, and the method comprises the following steps: determining a CHB patient group, and collecting baseline data, laboratory indexes before antiviral treatment, virological indexes and ultrasonic imaging indexes of patients; the patients are grouped according to virology treatment results, and propensity score matching is adopted to control confounding factors; single-variable and multivariable Cox regression analysis is adopted, and independent risk factors of PVT of the CHB patient are screened; and constructing a column graph prediction model based on the independent risk factors, and performing internal verification. According to the method, the high-precision PVT prediction model is constructed by integrating virology dynamic response indexes and clinical parameters, so that the accuracy of PVT risk assessment is remarkably improved, and a reliable tool is provided for individualized risk management of CHB cirrhosis patients. The model shows excellent performance (AUC is greater than 0.95) through internal verification, and the clinical applicability is strong.

Description

Technical Field

[0001] The present invention belongs to the technical field of medical data analysis, and specifically relates to a portal vein thrombosis prediction model and construction method for chronic hepatitis B patients based on virological treatment results, which is suitable for complication risk assessment and clinical management of patients with cirrhosis. Background Art

[0002] Portal vein thrombosis (PVT) is a serious complication that can occur in patients with chronic hepatitis B (CHB), particularly those with cirrhosis. Studies have shown that not only is hepatitis B virus (HBV) infection associated with PVT in CHB patients, but HBsAg and HBeAg positivity are also associated with PVT in CHB patients. Potential mechanisms of PVT in CHB patients include inflammatory responses, endothelial lesions, coagulation and anticoagulation disorders, cardiolipin antibody production, and hepatitis antigen-antibody complex deposition. Existing PVT prediction models are mostly based on traditional clinical indicators (such as liver function and coagulation parameters), but do not fully integrate dynamic virological response indicators and virological response typing after antiviral treatment. Currently, there are no clear predictors of PVT associated with HBV infection, such as latent HBV DNA and HBsAg.

[0003] Based on this, the present invention innovatively integrates virological dynamic data with clinical parameters, analyzes the baseline data, serological indicators, virological indicators and ultrasound data of CHB patients who receive long-term antiviral treatment, evaluates their different PVT outcomes, identifies risk indicators, and establishes and verifies a PVT prediction model. Summary of the Invention

[0004] The present invention aims to address the aforementioned issues in the prior art by proposing a prediction model and method for portal vein thrombosis in patients with chronic hepatitis B based on virological treatment outcomes. This model combines dynamic virological response indicators with clinical parameters to construct a prediction model for portal vein thrombosis in patients with chronic hepatitis B. This model then tracks PVT results and the antiviral virological response of the cohort, identifies independent risk factors for PVT, and establishes a prediction model for PVT in patients with cirrhosis and CHB, providing support for rapid assessment of patients' PVT risk.

[0005] The technical solution of the present invention is:

[0006] A method for constructing a portal vein thrombosis (PVT) prediction model for chronic hepatitis B patients based on virological treatment results comprises the following steps:

[0007] To identify a population of patients with chronic hepatitis B (CHB) and collect their baseline data, laboratory parameters before antiviral treatment, virological parameters, and ultrasound imaging parameters;

[0008] Patients were grouped according to virological treatment outcomes, and propensity score matching was used to control for confounding factors;

[0009] Univariate and multivariate Cox regression analyses were used to screen the independent risk factors for PVT in CHB patients.

[0010] A nomogram prediction model was constructed based on independent risk factors and internal validation was performed.

[0011] Furthermore, the virological indicators include any one or more of the HBV DNA load after antiviral treatment and before virological response, hepatitis B surface antigen (HBsAg), hepatitis B surface antibody (HBsAb), hepatitis B e antigen (HBeAg), hepatitis B e antibody (HBeAb), hepatitis B core antibody (HBcAb), HBV DNA load after virological response, and time to achieve virological response.

[0012] Furthermore, the virological treatment results are divided into five categories: virological response (VR), primary nonresponse (PNR), partial virological response (PVR), virological breakthrough (VBT), and sustained off-therapy virological response (SOVT).

[0013] Furthermore, the laboratory indicators before antiviral treatment include any one or more of liver function, alanine aminotransferase, aspartate aminotransferase, glutamyl transpeptidase, blood lipids, triglycerides, high-density lipoprotein, low-density lipoprotein, prothrombin time, activated partial thromboplastin time, international normalized ratio, D-dimer, fibrinogen, platelet count, mean platelet volume, tumor-related indicators, and alkaline phosphatase.

[0014] Furthermore, the ultrasound imaging indicators include any one or more of the spleen size, portal vein main trunk diameter, whether PVT occurs, and whether new hepatocellular carcinoma (HCC) occurs at the end of the patient's follow-up.

[0015] Furthermore, the baseline data include the patient's age, gender, history of hypertension, history of diabetes and degree of liver cirrhosis. The degree of liver cirrhosis includes whether there is varicose vein and bleeding in the portal vein system, whether there is ascites, and whether there is hepatic encephalopathy.

[0016] Furthermore, the independent risk factors include virological breakthrough (VBT), spleen long axis ≥18 cm, HBsAg>250 IU / mL, HBsAb>1000 mIU / mL, and combined primary liver cancer.

[0017] Furthermore, the internal validation was to divide the patient population into a training group and a validation group at a ratio of 7:3, evaluate the model performance, and validate the model using ROC curve, calibration curve and decision curve analysis.

[0018] The present invention also protects a prediction model for portal vein thrombosis in chronic hepatitis B patients based on virological treatment results obtained by any of the construction methods described above.

[0019] Beneficial effects of the present invention:

[0020] (1) The method for constructing a prediction model for portal vein thrombosis in chronic hepatitis B patients based on virological response provided by the present invention integrates virological dynamic response indicators for the first time, that is, the virological treatment results after antiviral treatment (such as VBT) are used as core predictive factors, breaking through the limitation of traditional models that only rely on static indicators. At the same time, through PSM analysis, cohort matching is optimized, the interference of confounding factors such as degree of cirrhosis and age on the results is reduced, the confounding factors are accurately controlled, and a multidimensional visualization model that can be interactively predicted is constructed.

[0021] (2) This study constructed a highly accurate PVT prediction model by integrating dynamic virological treatment outcome indicators with clinical parameters, significantly improving the accuracy of PVT risk assessment and providing a reliable tool for individualized risk management in patients with chronic hepatitis B (CHB) cirrhosis. Internal validation demonstrated excellent performance (AUC > 0.95) and strong clinical applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Figure 3 shows the KM curves for the effect of different antiviral treatment outcomes on PVT. Virological response (VR), primary nonresponse (PNR), partial virological response (PVR), and virological breakthrough (VBT) were included. When P < 0.05, the risk of PVT was considered different between groups. All PVTs occurred after patients achieved a virological response.

[0023] Figure 2 KM curves of different spleen sizes in response to PVT.

[0024] Spleen size = 0: indicates spleen thickness <4 cm; Spleen size = 1: indicates spleen thickness between 4 and 5 cm; Spleen size = 2: indicates spleen thickness >5 cm. When P < 0.05, the risk of PVT was considered different between the groups.

[0025] Figure 3 Figure 3 is the KM curve for the response of patients with or without HCC to PVT. When P < 0.05, the risk of PVT was considered different between the groups.

[0026] Figure 4 Figure 3 shows the KM curve for the relationship between HBV DNA level after antiviral treatment and PVT. HBV DNA level = Low represents a low HBV DNA level after treatment; HBV DNA level = Moderate represents a moderate level; and HBV DNA level = High represents a high level. When P < 0.05, the risk of PVT was considered different between the groups.

[0027] Figure 5 Visualization of the SMD before and after PSM: Linear distribution and probability density distribution. Treatment = 0: represents the unadjusted distribution before PSM, and Treatment = 1: represents the adjusted distribution after PSM. The matching ratio is 1:3, and the caliper value is 0.05.

[0028] Figure 6 Figure 3 shows the KM curves for the effect of different antiviral treatment outcomes after PSM on PVT. Virological response (VR), primary nonresponse (PNR), partial virological response (PVR), and virological breakthrough (VBT) were included. When P < 0.05, the risk of PVT was considered different between groups. All PVTs occurred after patients achieved a virological response.

[0029] Figure 7 Nomogram for median survival prediction based on risk indicators.

[0030] Figure 8 Receiver operating characteristic curves for median survival of the training and validation groups.

[0031] Figure 9 Nomogram calibration curve.

[0032] Figure 10 This is a nomogram decision curve analysis based on clinical features. The blue line represents the nomogram; the green line represents all negative samples; and the red line represents all positive samples. When the nomogram line lies above the other two lines, it is considered to have a good predictive effect. DETAILED DESCRIPTION

[0033] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0034] In order to further understand the present invention, the present invention will be further described with reference to the accompanying drawings and embodiments.

[0035] The present invention provides a method for constructing a portal vein thrombosis prediction model for chronic hepatitis B patients based on virological treatment results, comprising the following steps:

[0036] Step 1: Research subjects and inclusion criteria

[0037] A longitudinal retrospective cohort study of adults with new-onset chronic hepatitis B was conducted, with data collected from patients who were first diagnosed with chronic hepatitis B between January 1, 2012, and December 31, 2023.

[0038] 1. Selection criteria

[0039] All patients who meet all of the following criteria will be included in the cohort:

[0040] a. Presence of hepatitis B surface antigen (HBsAg) in serum for more than 6 months, accompanied by at least one documented detectable serum HBV DNA test (≥100 IU / mL).

[0041] b. After the HBV DNA level stabilizes, routine liver color Doppler ultrasound examination should be performed at least three times a year before the end of follow-up.

[0042] 2. Exclusion criteria

[0043] Patients who meet any of the following criteria will be excluded from the cohort:

[0044] a. received antiviral treatment before enrollment;

[0045] b. Under 18 years of age;

[0046] c. Anticoagulant drugs, including heparin, low molecular weight heparin, and warfarin, were administered during follow-up;

[0047] d. History of HCV, HDV, and HIV infection, or positive infection markers;

[0048] e. Hepatocellular carcinoma (HCC) diagnosis before enrollment;

[0049] f. A clear history of liver surgery or liver interventional surgery during the follow-up period;

[0050] g. A clear history of splenectomy before or during follow-up.

[0051] h. PVT occurs before virological response.

[0052] Step 2: Data Collection

[0053] 1. Baseline data:

[0054] The patient's age, gender, history of hypertension, history of diabetes, and degree of cirrhosis (whether there is varicose vein and bleeding in the portal vein, ascites, or hepatic encephalopathy).

[0055] 2. Laboratory indicators before antiviral treatment:

[0056] Liver function (total bilirubin T-Bil), alanine aminotransferase (ALT), aspartate aminotransferase (AST), glutamyl transpeptidase (GGT), blood lipids (total cholesterol (TC)), triglycerides (TG), high-density lipoprotein (HDL), low-density lipoprotein (LDL), blood coagulation (prothrombin time (PT)), activated partial thromboplastin time (APTT), international normalized ratio (INR), D-dimer (D-dimer), fibrinogen (Fib), platelet count (PLT), mean platelet volume (MPV), tumor-related indicators (alpha-fetoprotein (AFP), alkaline phosphatase (ALP)).

[0057] 3. Virological indicators:

[0058] HBV DNA load after antiviral treatment and before virological response, five hepatitis B test items (hepatitis B surface antigen (HBsAg), hepatitis B surface antibody (HBsAb), hepatitis B e antigen (HBeAg), hepatitis B e antibody (HBeAb), hepatitis B core antibody (HBcAb)), HBV DNA load after virological response, and time to achieve virological response.

[0059] 4. Imaging indicators:

[0060] The spleen size (maximum long axis diameter, maximum thickness diameter), portal vein trunk diameter (diameter of the portal vein trunk measured during digestive system ultrasound examination), occurrence of PVT, and development of new HCC were evaluated at the end of follow-up.

[0061] Step 3: Specific steps and results of virological testing

[0062] 1. Diagnosis of CHB patients:

[0063] Chronic hepatitis B requires HBsAg positivity for at least 6 months

[0064] A two-step immunoassay using Chemiflex technology, combining chemiluminescent microparticle immunoassay (CMIA) and flexible detection formats, was used to quantitatively measure HBsAg and HBsAb, and qualitatively measure HBeAg, HBeAb, and HBcAb.

[0065] According to the test criteria, when the HBsAg concentration is <0.05 IU / mL, the sample is considered unreactive and marked as negative; when the HBsAg concentration is >250.00 IU / mL, the sample is marked as reaching the upper threshold and marked as “>250.00 IU / mL”. Similarly, when the HBsAb concentration is >1000 mIU / mL, the sample is marked as reaching the upper threshold.

[0066] For HBeAg, HBeAb, and HBcAb, the system calculates and stores the cutoff relative optical unit (RLU) value based on the average PLU values obtained from the three test calibrators. The sample S / CO is then calculated using the formula "S / CO = sample - RLU / Cutoff - RLU." When the HBeAb test value is <1.00 S / CO, the sample is considered reactive and marked as positive. When the HBeAg and HBcAb test values are >1.00 S / CO, the sample is considered reactive and marked as positive.

[0067] 2. Evaluation of virological treatment outcomes and subsequent treatment plans

[0068] The virological treatment outcomes of the cohort were divided into five categories according to the EASL 2017 clinical practice guidelines.

[0069] (1) Virological response (VR) was defined as the undetectable serum HBV DNA by sensitive polymerase chain reaction (PCR) assay with a detection limit of 10 IU / ml.

[0070] (2) Primary nonresponse (PNR) is defined as a decrease in serum HBV DNA of less than 1 log10 after 3 months of treatment.

[0071] (3) Partial virological response (PVR) was defined as a decrease in HBV DNA of more than 1 log10 IU / ml, but detectable HBV DNA, in compliant patients after at least 12 months of treatment.

[0072] (4) Virological breakthrough (VBT) is defined as an increase of HBV DNA level by more than 1 log10 IU / ml compared with the lowest value of HBV DNA level.

[0073] (5) Sustained off-treatment virological response (SOVT) is defined as serum HBV DNA level <2000 IU / ml for at least 12 months after the end of treatment.

[0074] Because all patients in the cohort continued to take antiviral drugs until the end of follow-up, there was no response outcome of SOVT.

[0075] For patients with hepatitis B cirrhosis receiving nucleos(t)ide analogues (NAs) therapy, if HBV DNA is still detectable after 24 weeks (HBV DNA > 100 IU / mL), NA therapy should be adjusted (for patients with entecavir (ETV) and lamivudine (LAM) resistance, switch to tenofovir disoproxil fumarate (TDF). For patients with adefovir (ADV) resistance, if LAM is sensitive, switch to LAM; if LAM is resistant, use TDF; and if HBV DNA quantitative determination reaches a stable plateau, switch to ETV or add ETV. For TDF resistance, if LAM is sensitive, switch to ETV; if LAM is resistant, add ETV. For multidrug resistance, switch to ETV combined with TDF or tenofovir-alafenamide (TAF) combination therapy. It can also be used in combination with alfa interferon (IFN-α) or pegylated interferon (Peg-IFN).

[0076] Step 4. Specific steps and conclusions of the analysis method

[0077] 1. Baseline data characteristics and differences among different patients

[0078] A total of 252 patients were enrolled in the study and completed follow-up. Of the enrolled patients, 32 chose not to receive antiviral therapy; all others received conventional and long-term antiviral therapy. Initially, 183 patients received ETV alone; 17 received TDF alone; 10 received ADF alone; 1 received LVD alone; 2 received adefovir combined with lamivudine; and 7 received Peg-IFN-α. The median duration of PVT in the entire cohort was 2571 days.

[0079] The average age of the subjects was 53.7 years old, and 67.59% of them were male.

[0080] The mean baseline serum HBV DNA load was 132,500 IU / ml (Q1 = 5,587.50, Q3 = 3,737,500 IU / ml). During follow-up, 28 patients developed PVT (11.11%). Of these, 109 (43.25%) were HBeAg-positive and 125 (49.60%) were HBeAb-positive. All patients were HBcAb-positive, and HBsAb-positive after treatment.

[0081] There were significant differences in the following indicators between patients with PVT and those without PVT: age (P<0.001), portal vein diameter (P<0.001), spleen long axis (P<0.001), virological response time (P<0.001), PT (P=0.004), PLT (P<0.001), AST (P=0.049), GGT (P<0.025), ALP (P<0.001), AFP (P<0.001), D-Dmier (P<0.001), INR (P<0.001), LDL (P=0.010), HDL (P<0.015), TC (P=0.014), HBV DNA load and HBV after antiviral treatment. DNA level (P < 0.001), combined HCC (P < 0.001), cirrhosis level (P < 0.001 (using Fisher's exact test)), and antiviral response (P < 0.001) (Table 1)

[0082] Table 1 Baseline data characteristics and differences between PVT and non-PVT patients

[0083]

[0084]

[0085]

[0086]

[0087] Note: High (H), medium (M), and low (L) viral loads are defined as baseline serum HBV DNA levels ≥8.00, 5.00-7.99, and 2.00-4.99 log10 IU / ml, respectively. Clinical cirrhosis levels were categorized according to the following criteria: echogenic fibrosis on ultrasound; 1. No varices or other complications; 2. Varices without bleeding or ascites; 3. Gastroesophageal varices without ascites or hepatic encephalopathy; 4. Complications other than gastroesophageal varices; 5. More than two complications; 6. Recurrent infections, extrahepatic organ dysfunction, chronic or acute liver failure (ACLF), refractory ascites, persistent hepatic encephalopathy, and intractable jaundice. Grades 0-3 are considered compensated cirrhosis. Grades 4-6 are considered decompensated. Spleen size was categorized by spleen thickness: normal: spleen thickness <4 cm; slightly larger: 4-5 cm; and large: ≥5 cm. Differences between groups were considered statistically significant when P < 0.05. “*” indicates that the analysis was performed using Fisher’s exact probability method.

[0088] Among all patients receiving antiviral therapy, 188 (74.60%) achieved VR, 30 (11.90%) achieved PVR, 5 (1.98%) achieved PNR, and 29 (11.51%) achieved VBT. Analysis of differences in patients with different virological response groups showed significant differences in the degree of liver cirrhosis, age, and HCC. (Table 2)

[0089] Table 2 Baseline data characteristics and differences of patients with different virological responses

[0090]

[0091]

[0092]

[0093] Note: When P < 0.05, the difference between the groups was considered statistically significant. "*" indicates that the analysis was performed using Fisher's exact probability method.

[0094] 2. Risk factors for PVT in CHB patients

[0095] Univariate Cox regression analysis showed that the following factors were considered to be risk factors for PVT in CHB patients:

[0096] The patients were significantly associated with moderate HBV DNA level after antiviral treatment (P < 0.001, hazard ratio (HR) = 5.21; 95% confidence interval (95% CI) = 2.18-12.45); PVR virological response (P < 0.001, HR = 22.52; 95% CI = 4.54-111.66), PNR virological response (P = 0.006, HR = 30.09; 95% CI = 2.70-335.92) and VBT virological response (P < 0.001, HR = 65.60; 95% CI = 14.79-290.95); and newly diagnosed HCC (P < 0.001, HR = 14.29; 95% CI = 6.03-33.89); combined with decompensated cirrhosis (P < 0.001, HR = 2.37; 95% CI = 1.72-3.27); older age (P = 0.002, HR = 1.06; 95% CI = 1.02-1.10); larger portal vein diameter (P < 0.001, HR = 13.20; 95% CI = 5.32-32.75); larger spleen long axis (P < 0.001, HR = 1.37; 95% CI = 1.26-1.49); longer virological response time (P < 0.001, H R = 1.01; 95% CI = 1.01–1.01); higher levels of D-Dmier (P < 0.001, HR = 1.37; 95% CI = 1.26–1.49), INR (P = 0.003, HR = 5.24; 95% CI = 1.75–15.7), PT (P = 0.006, HR = 1.15; 95% CI = 1.04–1.27), MPV (P = 0.023, HR = 1.49; 95% CI = 1.06–2.11), and ALP (P < 0.001, HR = 1.01; 95% CI = 1.0 1-1.02), Tbil (P=0.013, HR=1.01; 95%CI=1.01-1.01), HDL (P=0.03, HR=1.13; 95%CI=1.01-1.27), HBsAg (P=0.002, HR=1.01; 95%CI=1.01-1.02) and HBsAb after antiviral treatment (P<0.001, HR=1.01; 95%CI=1.01-1.02); and lower PLT (P<0.001, HR=0.98; 95%CI=0.97-0.99).

[0097] Multivariate Cox regression analysis showed that the following factors were considered to be independent risk factors for PVT in CHB patients:

[0098] VBT antiviral response (P < 0.001, HR = 42.48; 95% CI = 5.02-359.89), concurrent new HCC (P < 0.001, HR = 30.69; 95% CI = 5.50-171.15), larger spleen long axis (P < 0.001, HR = 1.49; 95% CI = 1.21-1.84), higher HBsAg level (P = 0.007, HR = 1.03; 95% CI = 1.01-1.05), and higher HBsAb level after antiviral treatment (P = 0.004, HR = 1.02; 95% CI = 1.01-1.03) (Table 3).

[0099] Table 3 Single and multivariate Cox regression analysis of PVT in CHB patients; and multicollinearity analysis of related factors

[0100]

[0101]

[0102]

[0103]

[0104] Note: HR: hazard ratio, CI: confidence interval; when P < 0.05, the variable was considered to be associated with the risk of PVT in CHB patients. When the variance inflation factor (VIF) was > 3.00, the variable was considered to have significant collinearity.

[0105] To ensure the independence of each variable, a multicollinearity test was performed. After removing the variables with strong collinearity from Table 3, a multivariate Cox regression analysis was performed, and the results were similar (Table 4).

[0106] Table 4 Single and multivariate Cox regression analysis of PVT in CHB patients after removing variables with obvious collinearity

[0107]

[0108]

[0109] Note: HR: hazard ratio, CI: confidence interval; when P < 0.05, the variable was considered to be associated with the risk of PVT in CHB patients.

[0110] 3. The more severe the cirrhosis, the worse the effect of antiviral treatment

[0111] Univariate and multivariate ordered binary logistic regression revealed an association between the degree of cirrhosis and virological response. Decompensated cirrhosis (P < 0.001, HR = 4.60; 95% CI = 2.17–9.75) and advanced age (P = 0.044, HR = 1.03; 95% CI = 1.01-1.06) were independent risk factors for adverse virological responses (PVR, PNR, and VBT) in patients with CHB (Table 5).

[0112] Table 5 Univariate and multivariate logistic regression analysis of different virological response outcomes in CHB patients

[0113]

[0114]

[0115] Note: When P < 0.05, the variable is considered to be associated with the risk of adverse virological response (PVR, PNR, and VBT) in CHB patients.

[0116] 4. The risk of PVT varies among different CHB patients

[0117] The Kaplan-Meier method was used to draw survival curves and compare the cumulative risk of PVT with different virological treatment outcomes (VR, PVR, PNR, VBT). The Kaplan-Meier curve showed that patients with adverse virological responses to antiviral treatment ( Figure 1 ), splenomegaly ( Figure 2 ), combined with HCC ( Figure 3 ) and moderate HBV DNA levels after antiviral treatment ( Figure 4 ) have an increased risk of PVT.

[0118] Patients were then divided into groups based on virological response and their baseline data were analyzed for differences. Variables with significant differences in the baseline data were identified as possible confounders and included in the PSM analysis (Table 6).

[0119] Table 6 Differences and SMD of variables before and after PSM analysis

[0120]

[0121]

[0122]

[0123]

[0124] Note: VR indicates that the patient achieved virological response, and no VR indicates that the patient achieved major non-response, partial response, or virological breakthrough. When P < 0.05 and SMD < 0.2, confounding factors were considered to be controlled.

[0125] The purpose of the analysis in Table 6 was to balance the differences in these variables between different virological response groups as much as possible to reduce the interference of these confounding factors on the results. The PSM analysis basically controlled the SMD and the probability density of confounding factors affecting virological response, including the degree of cirrhosis, age, HCC, PT, PLT, etc. (Table 7, Figure 5 ) The KM curves after PSM of different antiviral treatment virological responses showed that Figure 1 Similar results. ( Figure 6 )

[0126] Step 5: Construct a risk prediction model for PVT in CHB patients

[0127] Based on the independent risk factors identified above, the following Figure 7 The nomogram prediction model is shown.

[0128] The constructed nomogram prediction model was internally validated: the cohort was divided into a training group (176 patients) and a validation group (76 patients) in a 7:3 ratio to evaluate model performance. Receiver operating characteristic (AUC) curves, calibration curves, and decision curve analysis (DCA) were used to verify the model's predictive efficacy and clinical applicability.

[0129] The difference analysis showed that there was no significant difference in any variable between the two groups (Table 7). The nomogram AUC of the training group and the validation group were 0.98 (95% CI = 0.96-1.00) and 0.97 (95% CI = 0.93-1.01), respectively. Figure 8 As shown, the calibration curves for median survival probabilities showed moderate agreement between the predicted and observed results for both the training and validation cohorts. Figure 9 and Figure 10 The DCA curve showed that if the patient's threshold probability was between 0.044 and 0.71, it had a better predictive effect on the occurrence of PVT in CHB patients.

[0130] Table 7 Analysis of the differences in variables between the validation group and the training group

[0131]

[0132]

[0133] Note: When P < 0.05, it is considered that there is a significant difference in the variable between the two groups.

[0134] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art may still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein with equivalents. Any modifications, equivalent replacements, and modifications made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A method for constructing a portal vein thrombosis prediction model for chronic hepatitis B patients based on virological treatment results, characterized in that: The following steps are involved: To identify a group of patients with chronic hepatitis B and collect their baseline data, laboratory parameters before antiviral treatment, virological parameters, and ultrasound imaging parameters; Patients were grouped according to virological treatment outcomes, and propensity score matching was used to control for confounding factors; Univariate and multivariate Cox regression analyses were used to identify independent risk factors for portal vein thrombosis in patients with chronic hepatitis B. A nomogram prediction model was constructed based on independent risk factors and internal validation was performed.

2. The construction method according to claim 1, characterized in that The virological indicators include any one or more of HBV DNA load, HBsAg, HBsAb, HBeAg, HBeAb, HBcAb after antiviral treatment and before virological response, HBV DNA load after virological response, and time to achieve virological response.

3. The construction method according to claim 1, wherein The virological treatment results are divided into five categories: virological response, primary non-response, partial virological response, virological breakthrough, and sustained non-treatment virological response.

4. The construction method according to claim 1, characterized in that The laboratory indicators before antiviral treatment include any one or more of liver function, alanine aminotransferase, aspartate aminotransferase, glutamyl transpeptidase, blood lipids, triglycerides, high-density lipoprotein, low-density lipoprotein, prothrombin time, activated partial thromboplastin time, international normalized ratio, D-dimer, fibrinogen, platelet count, mean platelet volume, tumor-related indicators, and alkaline phosphatase.

5. The construction method according to claim 1, characterized in that The ultrasound imaging indicators include any one or more of the spleen size, portal vein main trunk diameter, whether portal vein thrombosis occurs, and whether new hepatocellular carcinoma occurs at the end of the patient's follow-up.

6. The construction method according to claim 1, characterized in that The baseline data include the patient's age, gender, history of hypertension, history of diabetes and degree of liver cirrhosis. The degree of liver cirrhosis includes whether there is varicose vein and bleeding in the portal vein system, whether there is ascites, and whether there is hepatic encephalopathy.

7. The construction method according to claim 1, characterized in that The independent risk factors include virological breakthrough, spleen long axis ≥18cm, HBsAg>250IU / mL, HBsAb>1000mIU / mL, and concurrent primary liver cancer.

8. The construction method according to claim 1, wherein: The internal validation was to divide the patient population into a training group and a validation group at a ratio of 7:3, evaluate the model performance, and validate the model using ROC curve, calibration curve, and decision curve analysis.

9. A prediction model for portal vein thrombosis in chronic hepatitis B patients based on virological treatment results obtained according to the construction method according to any one of claims 1 to 8.

Images