System for predicting prognosis of liver cancer patient based on HED

Abstract

The application discloses a system for predicting the prognosis of a liver cancer patient based on HED. The system comprises a sample data acquisition module, an HED calculation module, a tumor-related antigen peptide prediction module and a prediction result output module. The system for predicting the prognosis of a liver cancer patient based on HED can more accurately predict the prognosis of liver cancer. It is found that high HED is related to good prognosis of liver cancer, corresponding to more TAA peptide segments capable of binding to HLA-I and stronger anti-tumor T cell response, and the above principle can be applied to predict the recurrence risk of early, medium and late liver cancer patients after treatment.

Description

Technical Field

[0001] This invention belongs to the field of liver cancer patient prognosis technology, specifically relating to a system for predicting the prognosis of liver cancer patients based on HED. Background Technology

[0002] Predicting the prognosis of liver cancer patients has always been a core issue in clinical diagnosis and research. Accurate assessment is crucial for developing individualized treatment plans, evaluating efficacy, and improving survival. The development of prognostic prediction models has evolved from single clinical indicators to multi-dimensional integrated models. Early models primarily relied on traditional clinicopathological features, such as tumor size, number, vascular invasion, Barcelona stage (BCLC), and TNM stage. While these indicators are intuitive and easily obtained, they often fail to fully reflect the high heterogeneity of tumors and the complex systemic condition of patients. With a deeper understanding of the biological behavior of liver cancer, tumor markers, such as alpha-fetoprotein (AFP), have been incorporated into the assessment system, but their sensitivity and specificity remain limited.

[0003] In recent years, research focus has shifted to comprehensive predictive systems integrating multi-omics data and cutting-edge technologies. Liquid biopsy technology, especially the detection of circulating tumor DNA (ctDNA), can reflect the gene mutation profile and molecular residual lesions of tumors in real time, showing great potential in predicting recurrence, metastasis, and treatment efficacy. Furthermore, intelligent analysis based on radiomics and deep learning can extract tumor texture and heterogeneity features that are invisible to the human eye from conventional images such as CT and MRI in high throughput. The models constructed have made significant progress in the preoperative non-invasive prediction of microvascular invasion, pathological grading, and postoperative recurrence risk. In the era of immunotherapy, tumor microenvironment-related biomarkers, such as PD-L1 expression, tumor-infiltrating lymphocyte characteristics, and systemic inflammatory markers (such as the neutrophil-lymphocyte ratio, NLR), have also become emerging biomarkers for predicting prognosis and immunotherapy efficacy. Future prognostic prediction models will develop towards greater dynamics and personalization. By integrating patients' clinicopathological baselines, multi-omics molecular characteristics, real-time dynamic monitoring data, and the continuous learning capabilities of artificial intelligence, a "digital twin"-style panoramic prediction system can be constructed. This system holds promise for earlier and more precise risk stratification and intervention, thereby truly improving the long-term survival outcomes of liver cancer patients. However, achieving standardized integration of multi-source data, completing large-scale prospective clinical validation, and ensuring the model's universality across different populations remain the main challenges.

[0004] Evolutionary divergence (HED) is a concept gaining increasing attention in evolutionary biology and ecology, particularly in phylogenetic ecology. It aims to measure the average evolutionary history or phylogenetic relationship among all species in an ecological community. HLA evolutionary divergence (HED) measures the degree of evolutionary difference between two HLA alleles (for a specific locus, such as HLA-A, HLA-B, DRB1, etc.) carried by an individual. The core of HED lies in the heterozygous advantage hypothesis, which states that individuals possessing two significantly different versions of key immune genes can recognize a wider range of pathogen peptides, thus gaining a survival advantage. Currently, there are no reports on predicting the prognosis of liver cancer patients using HED. Summary of the Invention

[0005] The purpose of this invention is to provide a system for predicting the prognosis of liver cancer patients based on HED.

[0006] A system for predicting the prognosis of liver cancer patients based on HED (Hepatocellular Carcinoma Evidence), the system comprising a sample data acquisition module, an HED calculation module, a tumor-associated antigen peptide prediction module, and a prediction result output module.

[0007] The sample data included early-stage liver cancer patients who received TACE combined with ablation therapy and intermediate-to-advanced liver cancer patients who received TKI combined with ICI therapy.

[0008] The HED calculation module includes the calculation of HLA evolutionary divergence, based on the Gratham distance of the amino acid difference between two alleles within the peptide-binding region. The specific calculation method is: Gratham distance = d i,j =ρ((α(c i −c j ) 2 +β(p i −p j ) 2 + γ(v i −v j ) 2 ) 1 / 2 , where d is the Grantham distance after pairwise sequence alignment, i and j are amino acids at homologous positions; c, p and v represent the biochemical composition, polarity and volume of the amino acid, respectively, and α, β and γ are constants.

[0009] The tumor-associated antigen peptide prediction module predicts the 9-mer peptides that bind to AFP, MAGE-A1, MAGE-A3, NY-ESO-1, SALL4, and SSX-2 and the individual patient's HLA-I allele; the HLA-I allele includes HLA-A, HLA-B, and HLA-C alleles.

[0010] The prediction result output module outputs the HED value and indicates whether the prognosis risk is high or low; among them, the HED-C HED cutoff value is 5.88, the HED-A HED cutoff value is 1.16, and the Mean HED cutoff value is 5.35. Values ​​higher than these cutoff values ​​indicate a low prognosis risk.

[0011] High HED is associated with a good prognosis for liver cancer.

[0012] High HED corresponds to more TAA peptides that can bind to HLA-I.

[0013] High HED supports a stronger anti-tumor T cell response.

[0014] The beneficial effects of this invention: This invention is based on a system for predicting the prognosis of liver cancer patients using high hepatocellular carcinoma (HEC) levels, which can predict the prognosis of liver cancer relatively accurately. This invention finds that high HEC is associated with a good prognosis in liver cancer, corresponding to more TAA peptides that can bind to HLA-I and support a stronger anti-tumor T cell response. Applying these principles, the risk of recurrence after treatment in early, intermediate, and advanced liver cancer patients can be predicted. Attached Figure Description

[0015] Figure 1 This is a HED feature map of HLA-A, -B, and -C sites in early and late hepatocellular carcinoma.

[0016] Figure 2 Survival analysis of early-stage hepatocellular carcinoma patients based on HLA-A HED (A), HLA-B HED (B), HLA-C HED (C), and mean HED (D).

[0017] Figure 3 Survival analysis of patients with advanced hepatocellular carcinoma based on HLA-A HED (A), HLA-B HED (B), HLA-C HED (C), and mean HED (D).

[0018] Figure 4 Correlation analysis of mean HED and predicted number of tumor-associated antigens (TAAs) in patients with early-stage hepatocellular carcinoma (A) and patients with advanced-stage hepatocellular carcinoma (B).

[0019] Figure 5 Correlation analysis (AC) of HLA-C HED and tumor-associated antigen (TAA)-specific T-cell response after ablation therapy in patients with early-stage hepatocellular carcinoma.

[0020] Figure 6 This study aimed to analyze the correlation between HLA-A (A), HLA-B (B), and mean (C) HED and tumor-associated antigen (TAA)-specific T cell responses in early-stage hepatocellular carcinoma patients who had not experienced recurrence after ablation therapy.

[0021] Figure 7 Correlation analysis of HLA-A (A), HLA-B (B), and mean (C) HED with tumor-associated antigen (TAA)-specific T cell responses in patients with early-stage hepatocellular carcinoma who relapsed after ablation therapy.

[0022] Figure 8 Correlation analysis (AD) between HED and tumor-associated antigen (TAA)-specific T cell response in patients with advanced hepatocellular carcinoma receiving immune checkpoint inhibitor therapy.

[0023] Figure 9 Correlation analysis of HED and tumor-associated antigen (TAA)-specific T cell responses in patients with advanced hepatocellular carcinoma receiving immune checkpoint inhibitor therapy.

[0024] Figure 10 HED is used as a prognostic indicator for immune checkpoint therapy in hepatocellular carcinoma. Detailed Implementation

[0025] To facilitate understanding of the present invention, a more comprehensive description will be given below. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to provide a thorough and complete understanding of the disclosure of the present invention. Example 1

[0026] 1. Experimental Methods Patients and Clinical Treatment: This study included 168 patients with hepatocellular carcinoma. Among them, 81 patients with early-stage liver cancer were admitted to Beijing You'an Hospital of Capital Medical University between October 19, 2012 and August 25, 2022.

[0027] The detailed procedure for transarterial chemoembolization (TACE) combined with ablation is as follows: The TACE procedure is performed by two interventional radiologists with more than five years of experience. Under local anesthesia, a catheter is inserted percutaneously into the right femoral artery. Using a super-slippery guidewire, the hepatic artery cannula is inserted into the hepatic artery, and under DSA guidance, a high-pressure injector is connected to inject 16 mL of contrast agent at a flow rate of 4 mL / s to visualize the proper hepatic artery, the left and right hepatic arteries, and their branches. Subsequently, using selective / superselective techniques, a highly flexible coaxial microcatheter is inserted into the tumor-feeding artery, and a mixture of doxorubicin and iodized oil is injected. The microcatheter is also connected to a high-pressure injector for angiography to confirm the contrast. Finally, embolization is performed using embolic materials such as gelatin sponge or polyvinyl alcohol particles until blood flow is completely stopped. The drug dosage is determined based on a comprehensive assessment of the patient's white blood cell count, platelet count, and liver function. The embolization endpoint is considered to be when angiography shows tumor vessel occlusion, adequate filling with embolic agent, and disappearance of tumor staining. Ablation therapy is performed within 2 weeks after TACE under CT or MRI guidance. The procedure is as follows: (1) Select the optimal puncture path and formulate an ablation plan using CT or MRI; (2) Insert the ablation needle after disinfection, draping, and local anesthesia; (3) Perform multi-point, overlapping ablation according to the size and number of tumors, and monitor the ablation range in real time; (4) After ablation, remove the needle and ablate the needle tract to prevent bleeding and needle tract seeding metastasis. Regardless of whether single or multi-stage ablation is chosen, a safe ablation boundary of 0.5-1.0 cm must be maintained to ensure complete tumor coverage, thereby achieving complete ablation.

[0028] The inclusion / exclusion criteria are: Inclusion criteria included: (1) age between 18 and 75 years; (2) diagnosis of early-stage hepatocellular carcinoma and treatment with radical ablation; and (3) presence of at least one measurable lesion according to the modified response criteria for solid tumors (mRECIST).

[0029] Exclusion criteria include: (1) Child-Pugh C liver function; (2) ECOG-PS>1; (3) history of other malignant tumors; (4) active autoimmune diseases or immunodeficiency requiring treatment; (5) incomplete or missing key clinical data.

[0030] The cohort of patients with intermediate and advanced hepatocellular carcinoma included 87 patients who were admitted between June 12, 2020 and March 28, 2025.

[0031] The specific approach to TKI combined with ICI therapy: The choice of targeted drug (tyrosine kinase inhibitor TKI, including lenvatinib or sorafenib) is determined by the attending physician based on patient tolerability and current clinical guidelines. Lenvatinib is administered orally once daily: 8 mg for patients weighing <60 kg and 12 mg for patients weighing ≥60 kg; the standard dose of sorafenib is 400 mg orally twice daily. Immune checkpoint inhibitor ICIs (including camrelizumab 200 mg, tislelizumab 200 mg, or sintilimab 200 mg) are administered intravenously every 3 weeks.

[0032] This study followed the Declaration of Helsinki and was approved by the Ethics Committee of Beijing You'an Hospital, Capital Medical University (approval number:

[2025] 103). All patients signed written informed consent forms.

[0033] Sample Collection and Processing: In the intermediate-to-advanced hepatocellular carcinoma (HCC) cohort, 13 patients receiving ICI combined with TKI therapy provided 10 mL peripheral blood samples at five time points: before treatment (T0) and after the first (T1), second (T2), third (T3), and fourth (T4) doses of anti-PD-1 therapy. Peripheral blood cells (PBMCs) were separated within 6 hours of blood collection using Ficoll density gradient centrifugation. The obtained PBMCs were resuspended, aliquoted, and immediately stored in liquid nitrogen in a cryopreservation buffer containing 90% fetal bovine serum and 10% DMSO. In the early-stage HCC cohort, 52 patients received 10 mL peripheral blood samples after TACE ablation therapy. The PBMC separation and cryopreservation methods were identical to those in the advanced-stage group.

[0034] Follow-up and efficacy assessment: All patients were followed up every 4–8 weeks. Tumor burden was assessed by imaging or physical examination according to the Modified RECIST criteria for evaluating the efficacy of treatment in solid tumors (mRECIST).

[0035] Treatment response is categorized as follows: Continuing clinical benefit (DCB): complete response (CR), partial response (PR), or stable disease (SD) lasting ≥24 weeks; No continuing benefit (NDB): progression (PD) or SD lasting <24 weeks. The primary endpoint for early-stage liver cancer is recurrence-free survival, while the primary endpoint for intermediate-to-advanced-stage liver cancer is progression-free survival.

[0036] HLA genotyping and HED calculation: Genomic DNA was extracted from fresh peripheral blood for HLA-I genotyping. HLA-A, HLA-B, and HLA-C alleles were detected using sequencing-based genotyping (PCR-SBT). HLA evolutionary divergence (HED) was calculated using existing methods, based on the Gratham distance between two alleles within the peptide-binding region. HED was calculated using the HLA-HED software (https: / / g4hithub.com / sunhuaibo / HLA-HED).

[0037] Tumor-associated antigen (TAA) peptide prediction: The full-length amino acid sequences of AFP, MAGE-A1, MAGE-A3, NY-ESO-1, SALL4, and SSX-2 are input into MHCflurry (v2.0.0) to predict the 9-mer peptide that binds to the individual patient's HLA-I allele.

[0038] The selection criteria for candidate TAA peptides were: high predictive affinity for at least one patient-specific HLA-I allele (IC50 < 500 nmol / L); and a ranking score within the top percentile of the software-recommended strong-binding peptides.

[0039] Synthetic peptides: 334 18-mer (10-amino acid overlap) peptides covering the full-length sequences of AFP, SALL4, MAGE-A1, MAGE-A3, NY-ESO-1, and SSX2 were synthesized. All peptides were confirmed to have a purity >90% by HPLC (GenScript Biotech, Jiangsu). They were first dissolved in DMSO, then diluted with RPMI-1640, and divided into 9 peptide libraries (23-45 peptides per library).

[0040] IFN-γ ELISPOT assay: The ELISPOT assay was performed on PBMC samples from 52 patients with early-stage liver cancer and 13 patients with advanced-stage liver cancer at five time points (T0-T4) to quantify antigen-specific T-cell responses. The specific method was as follows: 2.5 × 10⁻⁶ cells were added to each well. 5 PBMCs were prepared and a peptide library (8 μg / mL) was added, followed by stimulation for 16–18 hours; then biotinylated antibody (2 h) and streptavidin-alkaline phosphatase (1 h) were added, and NBT / BCIP was used for colorimetric development. Results were expressed as SFUs / 10 after background subtraction. 6 PBMC defines a positive result as: >3 times that of the negative control and >25 SFUs / 10 6 Background > 25 SFUs / 10 6 Experiments that fail to produce a positive control should be excluded.

[0041] Statistical analysis: Continuous variables are expressed as mean ± standard deviation or median (IQR). Cutoff values ​​for quantitative variables were determined using the `surv_cutpoint` function of the `survminer` package. Risk factors were selected using univariate and multivariate Cox regressions to construct the final nomogram. Model discrimination was assessed using the AUC of time-dependent ROC curves, and calibration was assessed using calibration curves and Brier scores. Patients were divided into high-risk and low-risk groups based on their total model score. R software (v4.2.2) was used with packages such as "rms", "survival", "riskRegression", "pec", "plotROC", and "timeROC"; Python was used with "hla_hed" and "mhcflurry". A two-sided p < 0.05 was considered statistically significant.

[0042] 2. Experimental Results: 1. Basic Patient Characteristics: A total of 168 patients with hepatocellular carcinoma (HCC) were included, of whom 81 were early-stage patients who received TACE combined with ablation therapy and 87 were advanced-stage patients who received ICI combined with TKI therapy. The overall cohort was predominantly male (82.1%), with a mean age of 57.6 years, and there was no significant age difference between the two groups. Among the early-stage HCC patients, 21.1% were stage 0 BCLC and 78.9% were stage A. The distribution of advanced-stage patients was as follows: stage B 40.3% and stage C 59.7%. The early-stage group had better liver function (Child-Pugh A 77.2%), compared to 44.0% in the advanced-stage group (Table 1).

[0043] Table 1 Baseline characteristics of patients with early and late hepatocellular carcinoma

[0044] 2. HLA-I HED characteristics in patients with early-stage and intermediate-stage hepatocellular carcinoma: Comparing the HLA-I distribution between patients with early-stage and intermediate-stage hepatocellular carcinoma, there was no significant difference in overall HLA-I levels (p>0.05). Figure 1 A). The HLA-C genotypes differed significantly between the two groups. Figure 1 B), but there was no significant difference in HLA-A / B. There were no significant differences in HLA-A / B / C loci and mean HED between the two groups (p>0.05). Hierarchical clustering after Z-score normalization showed that HED could form high / low divergence allele clusters, consistent with the relationship between HLA-A / B / C. Early-stage liver cancer showed lower cluster separation ( Figure 1 C), the cluster separation of mid-to-late stage liver cancer is relatively high ( Figure 1 D). These results suggest that tumor immune selection pressure may promote the evolution of HLA alleles towards a more conserved direction in order to evade immune surveillance or maintain immune recognition stability.

[0045] 3. High HLA-A HED is associated with a better prognosis in patients with hepatocellular carcinoma: In early-stage hepatocellular carcinoma, patients with high HLA-A HED had significantly longer recurrence-free survival than those with low HED (HR = 0.59, p = 0.037). Figure 2 A). HLA-B and HLA-C HED were not significantly associated with relapse-free survival ( Figure 2 B, 2C). The high mean HED group also showed a trend toward longer recurrence-free survival (HR = 0.52, p = 0.05). Figure 2 D).

[0046] In intermediate and advanced hepatocellular carcinoma, both high HLA-A (HR = 0.54, p = 0.046) and high HLA-C HED (HR = 0.52, p = 0.011) were significantly associated with longer progression-free survival. Figure 3 A, 3C), no significant difference in HLA-B ( Figure 3 B). The high-mean HED group also had better progression-free survival (HR = 0.59, p = 0.033). Figure 3 D).

[0047] Therefore, high HED is associated with a good prognosis for liver cancer, and this effect is more pronounced in patients with advanced stages.

[0048] 4. Correlation between HED and the predicted number of TAA peptides In early and intermediate-to-late-stage hepatocellular carcinoma, HLA-A / B / C ratios and mean HED were significantly positively correlated with the number of TAA (AFP, SALL4, MAGE-A1, MAGE-A3, NY-ESO-1) peptides (r = 0.28–0.47, all p < 0.01). SSX2 showed only a non-significant trend in early-stage patients; NY-ESO-1 showed no significant trend in the late-stage group. Figure 4 Overall, high HED corresponds to more TAA peptides that can bind to HLA-I.

[0049] Correlation between HED and TAA-specific T-cell response in early-stage hepatocellular carcinoma: To validate the predictive results, we stratified early-stage hepatocellular carcinoma patients by recurrence status and performed ELISPOT assays to determine TAA-specific T-cell responses. The recurrence-free group ( Figure 5 AC: AFP-specific response showed a positive trend with HLA-C HED (R = 0.64, p = 0.12), while SMNMS showed no significant correlation. Total response also showed a positive trend (R = 0.56, p = 0.11). HLA-A / B / mean HED showed consistent results ( Figure 6 Relapse group () Figure 5DF): AFP response was significantly negatively correlated with HLA-C HED (R = -0.48, p = 0.043), and the overall response was also significantly negatively correlated (R = -0.54, p = 0.021). HLA-A / B / mean HED showed a similar trend ( Figure 7 These results suggest that in relapse-free patients, high HED supports a stronger anti-tumor T-cell response; in relapsed patients, a negative correlation may reflect tumor immune escape or impaired / exhausted T-cell function.

[0050] Dynamic correlation between HED and TAA-specific T cell response in intermediate and advanced hepatocellular carcinoma (ICI treatment): Further analysis of the dynamic relationship between HED and TAA-specific response during ICI+TKI treatment (T0-T4): Taking HLA-A as an example ( Figure 8 A): T0: Strong negative correlation (R = −0.93, p<0.01); T1: Negative correlation still exists, but its strength weakens (R = -0.66, p = 0.019); T2-T3: Negative correlation gradually weakens (T2: R = -0.32; T3: R = -0.45); T4: Turns into positive correlation (R = 0.31). HLA-B, HLA-C, and average HED all show similar dynamic transitions. Figure 8 BD, Figure 9 Explanation: Before treatment (T0), immune activation is insufficient → overall T cell response is low. With continued ICI treatment (T1-T4), immune activation is enhanced. After suppression is relieved, the antigen presentation advantage of high HED gradually becomes apparent → stronger specific immune response.

[0051] High HED as a predictor of progression-free survival for immunotherapy in intermediate and advanced liver cancer: Univariate and multivariate Cox analyses showed that HLA-A HED, HLA-C HED, GGT, Fib, and DBIL / TBIL ratio were independent predictors of progression-free survival. Figure 10 A, Table 2). A Nomogram was constructed based on these variables (…). Figure 10 (B) was used to predict progression-free survival at 6 / 12 / 18 months. The model performed well: time-dependent AUCs were 0.731, 0.848, and 0.927, respectively. Figure 10 C), the calibration curve fits well ( Figure 10 D). Brier scores were 0.202, 0.107, and 0.094, respectively. Risk stratification showed that progression-free survival was significantly shorter in the high-risk group than in the low-risk group (p = 0.003). Figure 10 E). The risk table shows that no patients in the high-risk group remained progression-free after 15 months, while some patients in the low-risk group remained progression-free at the end of the 25-month follow-up.

[0052] Table 2 Univariate Cox regression analysis of prognostic factors

[0053] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A system for predicting the prognosis of liver cancer patients based on HED, characterized in that, The system includes a sample data acquisition module, an HED calculation module, a tumor-associated antigen peptide prediction module, and a prediction result output module.

2. The system for predicting the prognosis of liver cancer patients based on HED according to claim 1, characterized in that, The sample data includes early-stage liver cancer patients who received TACE combined with ablation therapy and intermediate-to-advanced liver cancer patients who received TKI combined with ICI therapy.

3. The system for predicting the prognosis of liver cancer patients based on HED according to claim 1, characterized in that, The HED calculation module includes the calculation of HLA evolutionary divergence, based on the Gratham distance of the amino acid difference between two alleles within the peptide-binding region. The specific calculation method is: Gratham distance = d i,j =ρ((α(c i −c j ) 2 +β(p i −p j ) 2 +γ(v i −v j ) 2 ) 1 / 2 , where d is the Grantham distance after pairwise sequence alignment, i and j are amino acids at homologous positions; c, p and v represent the biochemical composition, polarity and volume of the amino acid, respectively, and α, β and γ are constants.

4. The system for predicting the prognosis of liver cancer patients based on HED according to claim 1, characterized in that, The tumor-associated antigen peptide prediction module predicts the 9-mer peptides that bind to AFP, MAGE-A1, MAGE-A3, NY-ESO-1, SALL4, and SSX-2 and the individual patient's HLA-I allele; the HLA-I allele includes HLA-A, HLA-B, and HLA-C alleles.

5. The system for predicting the prognosis of liver cancer patients based on HED according to claim 1, characterized in that, The prediction result output module outputs the HED value and indicates whether the prognosis risk is high or low; among them, the HED-C HED cutoff value is 5.88, the HED-A HED cutoff value is 1.16, and the Mean HED cutoff value is 5.

35. Values ​​higher than these cutoff values ​​indicate a low prognosis risk.

6. The system for predicting the prognosis of liver cancer patients based on HED according to claim 5, characterized in that, High HED is associated with a good prognosis for liver cancer.

7. The system for predicting the prognosis of liver cancer patients based on HED according to claim 5, characterized in that, High HED corresponds to more TAA peptides that can bind to HLA-I.

8. The system for predicting the prognosis of liver cancer patients based on HED according to claim 5, characterized in that, High HED supports a stronger anti-tumor T cell response.

Images