Application of protein molecular marker in preparation of product for predicting effectiveness of neoadjuvant immunotherapy of lung squamous cell carcinoma patient

Abstract

The invention provides application of a protein molecular marker in preparation of a product for predicting the effectiveness of neoadjuvant immunotherapy of a lung squamous cell carcinoma patient, and belongs to the technical field of biomedicine. Proteomics is utilized to identify pathologic remission biomarkers of lung squamous cell carcinoma patients and establish a prediction model, 20 protein molecules related to pathologic remission are found, the 20 protein molecules comprise 16 neoadjuvant immunotherapy poor prognosis molecules KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2 and 4 treatment good prognosis molecules CD74, HLA-DPA1, ICAM1 and VAV1, the effectiveness of neoadjuvant immunotherapy of the lung squamous cell carcinoma patients can be evaluated, and the pathologic remission of the lung squamous cell carcinoma patients can be evaluated. And the predictive ability of the model may be expanded to immunotherapy of advanced lung cancer patients.

Description

Technical Field

[0001] The present invention belongs to the field of biomedical technology, and in particular relates to the application of protein molecular markers in the preparation of products for predicting the effectiveness of neoadjuvant immunotherapy for patients with lung squamous cell carcinoma. Background Art

[0002] Neoadjuvant immunotherapy is a promising treatment strategy in the field of solid tumors in recent years, and has brought a historic breakthrough for non-small cell lung cancer (NSCLC). A large number of clinical trials have shown that for early resectable NSCLC, a high pathological remission rate can be achieved after neoadjuvant immunotherapy, and the toxicity is controllable, which significantly prolongs the patient's survival. Studies have shown that the survival rate of patients with RVT less than or equal to 10% after neoadjuvant therapy is significantly improved. However, not all patients can get good treatment effects. In all clinical trials, up to 20% of patients did not undergo surgery, of which 5%-10% did not undergo surgery due to disease progression. Ineffective neoadjuvant therapy may lead to delayed surgery and increase the risk of immune-related adverse events. Therefore, accurate prediction of pathological remission can help guide patient management and clinical trial design of neoadjuvant immunotherapy, and can help improve individualized precision treatment. However, there is currently a lack of suitable markers to identify people who can potentially benefit from neoadjuvant therapy strategies. Summary of the invention

[0003] In view of this, the purpose of the present invention is to provide a protein molecular marker for the preparation of a product for predicting the effectiveness of neoadjuvant immunotherapy for patients with lung squamous cell carcinoma, wherein the protein molecular marker includes at least one of CD74, HLA-DPA1, ICAM1, VAV1, KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2. The present invention uses proteomics to identify biomarkers of pathological remission in patients with lung squamous cell carcinoma and establish a prediction model, and finds 20 protein molecules associated with pathological remission, including 16 neoadjuvant immunotherapy poor prognosis molecules and 4 treatment good prognosis molecules, which can evaluate the effectiveness of neoadjuvant immunotherapy for patients with lung squamous cell carcinoma, and the predictive ability of the model may be extended to immunotherapy for patients with advanced lung cancer.

[0004] In order to achieve the above-mentioned object of the invention, the present invention provides the following technical solutions:

[0005] The present invention provides an application of a protein molecular marker in the preparation of a product for predicting the effectiveness of neoadjuvant immunotherapy for patients with lung squamous cell carcinoma, wherein the protein molecular marker comprises at least one of CD74, HLA-DPA1, ICAM1, VAV1, KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2.

[0006] The present invention provides an application of a protein molecular marker in the preparation of a product for predicting a poor prognosis of a patient with squamous cell lung cancer undergoing neoadjuvant immunotherapy. The protein molecular marker comprises at least one of KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2.

[0007] The present invention provides the use of protein molecular markers in preparing a product for predicting a good prognosis of patients with lung squamous cell carcinoma after neoadjuvant immunotherapy, wherein the protein molecular markers include at least one of CD74, HLA-DPA1, ICAM1 and VAV1.

[0008] The present invention provides the use of KRT5 / 6 combined with PD-L1 in the preparation of a product for predicting a poor prognosis of neoadjuvant immunotherapy in patients with squamous cell lung cancer.

[0009] Preferably, the product comprises a kit or a biochip.

[0010] The present invention also provides a method for constructing a neoadjuvant immunotherapy efficacy prediction model for patients with lung squamous cell carcinoma, comprising the following steps:

[0011] 1) Data collection: The residual tumor percentage (RVT) of the tumors resected after neoadjuvant therapy for squamous cell lung carcinoma was measured. CPR was the absence of viable tumor cells in both the primary lung lesion and lymph nodes.

[0012] 2) Identify potential molecular subtypes in patients with squamous cell lung carcinoma: Identify potential molecular subtypes in patients with squamous cell lung carcinoma by performing consensus clustering on the protein expression matrix of the squamous cell lung carcinoma cohort; Identify potential molecular subtypes in patients with squamous cell lung carcinoma by clustering analysis of 346 differentially expressed proteins between CPR and non-CPR;

[0013] 3) Weighted gene co-expression network analysis: All identified proteins are taken as input and soft thresholds are calculated to establish the basis of the co-expression network. On this basis, the co-expression network is constructed and genes are classified into different modules, and then the correlation between genes and phenotypes in the modules of interest is explored;

[0014] 4) Mfuzz cluster analysis: The differentially expressed proteins among the three groups of RVT 0%, 1-10%, and 11-100% were clustered using the Mfuzz algorithm;

[0015] 5) Correlation analysis: Correlation analysis between protein molecule expression level and RVT;

[0016] 6) Build the model.

[0017] Preferably, the potential molecular subtypes in patients with lung squamous cell carcinoma in step 2) are three proteomic subtypes: CCP_I, CCP_II and CCP_III.

[0018] Preferably, in the module of interest in step 3), the light cyan module is significantly positively correlated with the CCP_I subtype and negatively correlated with CPR; the cyan module and the green module are positively correlated with the CCP_II subtype and CPR.

[0019] Preferably, in step 4), cluster analysis obtains 6 clusters with similar trajectories, including cluster 1, cluster 4 and cluster 6.

[0020] Preferably, the model building in step 6) includes: taking the intersection of proteins whose expression levels are negatively correlated with RVT percentage, differentially expressed proteins (DEPs) significantly increased in CPR patients, proteins in green modules and cyan modules, proteins in cluster 1 in Mfuzz analysis, and significantly highly expressed DEPs in CCP_II subtype (fold change>2), and identifying four biomarkers indicating good prognosis CD74, HLA-DPA1, ICAM1 and VAV1; at the same time, proteins whose expression levels are positively correlated with RVT, DEPs significantly decreased in CPR patients The intersection of proteins in the light cyan module, proteins in clusters 4 and 6 identified by Mfuzz, and DEPs highly expressed in CCP_I subtype (fold change > 2) was taken, and 16 biomarkers indicating poor prognosis were obtained, including KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2; According to the patient's disease-free survival (DFS), a prognostic model was constructed and the risk score was calculated.

[0021] The present invention utilizes proteomics to identify biomarkers of pathological remission of patients with squamous cell lung cancer and establish a prediction model, and finds 20 protein molecules associated with pathological remission, including 16 neoadjuvant immunotherapy poor prognosis molecules KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3, TACSTD2 and 4 treatment good prognosis molecules CD74, HLA-DPA1, ICAM1, VAV1, which can evaluate the effectiveness of neoadjuvant immunotherapy for patients with squamous cell lung cancer, and the predictive ability of the model may be extended to immunotherapy for patients with advanced lung cancer. Among them, high keratin expression indicates a poor prognosis, so we further evaluated the immunohistochemistry level of keratin molecules CK5 / 6, which have been used in clinical practice, and verified the previous analysis results. It was found that CK5 / 6 combined with PD-L1 can better predict the degree of pathological remission. It is intended to be directly applied in routine clinical diagnosis and treatment to identify patients who can benefit from neoadjuvant immunotherapy. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1 Flow chart for proteomics analysis;

[0023] Figure 2 Risk factor linkage diagram: risk score scatter plot, patient survival (DFS) scatter plot and protein expression heat map;

[0024] Figure 3 This is the survival curve of the prognostic model for high-risk and low-risk groups in the neoadjuvant immunotherapy cohort;

[0025] Figure 4 The area under the curve (AUC) for the prognostic model to predict 12-, 18-, and 24-month DFS in the neoadjuvant cohort;

[0026] Figure 5 Generate curves for the prognostic model of high- and low-risk groups in the GSE135222 immunotherapy cohort;

[0027] Figure 6 The area under the curve (AUC) of the prognostic model for predicting 200-, 300-, and 400-day PFS in the GSE135222 cohort;

[0028] Figure 7 This is a schematic diagram of the strength of CK5 / 6 immunohistochemistry (IHC);

[0029] Figure 8 is the correlation diagram between CK5 / 6 expression and residual viable tumor percentage (RVT);

[0030] Fig. 9The expression level of CK5 / 6 in the CPR group and the non-CPR group;

[0031] Fig.10 The expression level of CK5 / 6 in the MPR group and the non-MPR group;

[0032] Fig.11 The area under the receiver operating characteristic (ROC) curve (AUC) plot for the combined prediction of CPR by PD-L1 and CK5 / 6;

[0033] Fig.12 This is the area under the receiver operating characteristic (ROC) curve (AUC) graph for the combined prediction of MPR by PD-L1 and CK5 / 6. DETAILED DESCRIPTION

[0034] The technical solutions provided by the present invention are described in detail below in conjunction with the embodiments, but they should not be construed as limiting the protection scope of the present invention.

[0035] Example 1

[0036] 1. Data Collection

[0037] We conducted a multicenter retrospective study in which subjects were collected from three tertiary hospitals, including the Second Affiliated Hospital of Zhejiang University School of Medicine, the Fourth Affiliated Hospital of Zhejiang University School of Medicine, and the Taizhou Hospital of Zhejiang Province. Figure 1 ), each patient received neoadjuvant immunotherapy combined with chemotherapy for 2-4 cycles before surgical resection of the tumor. We retrospectively analyzed the patients' clinical data, including clinical information, treatment modality, radiology, pathological response level, and disease-free survival (DFS). MPR was defined as no more than 10% of viable tumor cells in the primary tumor bed by measuring the residual tumor percentage (RVT) in the tumor resected after neoadjuvant therapy. The absence of viable tumor in the primary lung lesion but metastatic cancer in the lymph nodes was also defined as MPR. CPR was defined as the absence of viable tumor cells in both the primary lung lesion and the lymph nodes.

[0038] The specific inclusion and exclusion criteria for patients are as follows: inclusion criteria: age between 18 and 85 years old; stage IB-IIIB (according to the 8th edition of the AJCC / UICCTNM staging system); pathologically confirmed as LUSC; no anti-tumor treatment; and radical surgical resection after 2-4 cycles of neoadjuvant immunotherapy combined with chemotherapy; exclusion criteria: combined with other malignant tumors; combined with autoimmune diseases; positive driver genes; distant metastatic lesions; no available data.

[0039] Validation dataset: Transcriptome sequencing data and corresponding clinical data (GSE135222) of the immunotherapy cohort of NSCLC patients were collected to evaluate the predictive ability of immunotherapy efficacy. The dataset GSE135222 (with chip expression data and available PFS survival data) was downloaded from the NCBI-GEO database (https: / / www.ncbi.nlm.nih.gov / gds). The gene expression profiles and clinical information of LUSC samples were extracted from The Cancer Genome Atlas (TCGA) database (https: / / portal.gdc.cancer.gov / ). We also collected single-cell sequencing from 4 patients with squamous cell lung cancer receiving neoadjuvant immunotherapy, and FFPE biopsy tissue blocks from another 81 patients to evaluate the CK5 / 6 immunohistochemical expression levels to verify the prediction of keratin to predict the efficacy of neoadjuvant immunotherapy.

[0040] 2. Proteomic subtypes and their biological functions

[0041] We identified potential molecular subtypes of LUSC patients by performing consensus clustering on the protein expression matrix of the LUSC cohort. Potential molecular subtypes in LUSC patients were determined by clustering analysis of 346 differentially expressed proteins (DEPs) between CPR and non-CPR (P < 0.05). Consensus clustering was calculated using the R package ConsensusClusterPlus v.3.8

[117] . Clustering analysis was performed using the following settings: maxK = 6, reps = 1000, pltem = 0.8, pFeature = 1, clusterAlg = "hc", distance = "speason". Based on this, we identified three potential protein molecular subtypes: CCP_I, CCP_II, and CCP_III. The three proteomic subtypes were significantly correlated with pathological response. The CCP_I subtype had the worst treatment effect, the CCP_II subtype had the best pathological response, and the CCP_III subtype had an intermediate treatment effect.

[0042] 3. Weighted gene co-expression network analysis

[0043] All identified proteins were used as input and soft thresholds were calculated to establish the basis of the co-expression network. On this basis, the co-expression network was constructed and genes were classified into different modules, and then the correlation between the genes and phenotypes of the modules of interest was explored. The light cyan module was significantly positively correlated with the CCP_I subtype and negatively correlated with CPR. This result indicates that the co-expressed proteins in the light cyan module can predict the poor efficacy of immunotherapy response. The cyan module and the green module were positively correlated with the CCP_II subtype and CPR, indicating that the protein molecules associated with better immunotherapy effects are enriched in the cyan module and the green module.

[0044] 4. Mfuzz Clustering

[0045] In order to systematically analyze the dynamic changes of proteins during the increase of RVT, we clustered the differentially expressed proteins among the three groups of RVT 0%, 1-10%, and 11-100% using the Mfuzz algorithm, and obtained 6 clusters with similar trajectories (Kruskal-Wallis test, P < 0.05). Among them, the protein expression levels in clusters 4 and 6 showed an increasing trend, while the proteins in cluster 1 showed a decreasing trend.

[0046] 5. Correlation Analysis

[0047] The protein expression levels were correlated with RVT (Spearman correlation, P<0.05). Proteins positively correlated with RVT were significantly correlated with keratinization and metabolism, while proteins negatively correlated with RVT were significantly correlated with immune response.

[0048] VI. Construction of a prediction model for the efficacy of neoadjuvant immunotherapy

[0049] Four biomarkers suggesting a good prognosis (CD74, HLA-DPA1, ICAM1, and VAV1) were identified by intersecting proteins whose expression levels were negatively correlated with RVT percentage, DEPs significantly increased in CPR patients, proteins in the green and cyan modules, proteins in cluster 1 in Mfuzz analysis, and DEPs significantly highly expressed in CCP_II subtype (fold change > 2); The intersection of EPs, proteins in the light cyan module, proteins in clusters 4 and 6 identified by Mfuzz, and highly expressed DEPs in CCP_I subtype (fold change > 2) resulted in 16 biomarkers indicating poor prognosis (KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3, and TACSTD2). Based on the patient's DFS, a prognostic model was constructed (the model was constructed using cph, and the risk factor linkage map was drawn using the ggrisk package, Figure 2 ), calculated the risk score, and divided all patients into high-risk group and low-risk group according to the median risk score. Compared with the low-risk group, the DFS of high-risk patients was significantly reduced ( Figure 3 ), in order to evaluate the robustness of the prognostic model, we used the public immunotherapy lung cancer cohort GSE135222 as an independent validation set for validation, and also proved that patients with high risk scores had shorter survival ( Figure 5 ). ROC curve ( Figure 4 , Figure 6 ) to determine the predictive value of the immune efficacy of the model.

[0050] 7. Keratin immunohistochemistry levels for predicting the efficacy of neoadjuvant immunotherapy

[0051] Through single-cell sequencing analysis, we found that the expression of KRT6A, KRT6B, KRT5, KRT17, MAL2, SPRR1B, PKP1, and JUP molecules in non-CPR patients was significantly higher than that in CPR patients in epithelial cells. Among them, KRT6A, KRT6B, and KRT5 molecules were the most significant. Given that cytokeratin 5 / 6 (CK5 / 6, also called KRT5 / 6) is one of the most commonly used LUSC pathological diagnostic markers in clinical practice, we retrospectively collected an independent validation cohort of 81 patients with lung squamous cell carcinoma who received neoadjuvant immunotherapy combined with chemotherapy and found that the IHC intensity of CK5 / 6 ( Figure 7) is significantly correlated with pathological remission, and patients with low CK5 / 6 expression have a higher CPR rate in neoadjuvant therapy. At the same time, because PD-L1 is currently recognized as a predictive marker for immunotherapy, we used logistic regression analysis to find that the combination of PD-L1 and CK5 / 6 is more effective than CK5 / 6 or PD-L1 alone in predicting CPR and MPR. Figures 8 to 12 As shown in the results, the correlation between CK5 / 6 expression (IHC intensity 3+ ratio) and the residual viable tumor percentage (RVT) was significantly positively correlated; the expression levels of CK5 / 6 in the CPR group and the non-CPR group were significantly increased in the non-CPR group; the expression levels of CK5 / 6 in the MPR group and the non-MPR group were significantly increased in the non-MPR group; the area under the receiver operating characteristic curve (ROC) curve (AUC) of the combined prediction of PD-L1 and CK5 / 6 for CPR was significantly higher than that of PD-L1; the area under the receiver operating characteristic curve (ROC) curve (AUC) of PD-L1 and CK5 / 6 for MPR was significantly higher than that of PD-L1.

[0052] The tissue samples obtained from patients with lung squamous cell carcinoma at the time of diagnosis were subjected to mass spectrometry-based quantitative proteome sequencing. A prognostic model based on 20 key molecules was obtained through analysis, and the reliability and robustness of the model were verified by an external cohort. Finally, combined with clinical practice, keratin 5 / 6 immunohistochemistry was widely used in the molecular diagnosis of lung squamous cell carcinoma in clinical practice to evaluate the correlation between keratin 5 / 6 expression levels and pathological remission.

[0053] A prognostic model for immunotherapy of squamous cell lung carcinoma was constructed based on 20 genes (CD74, HLA-DPA1, ICAM1, VAV1, KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3, and TACSTD2), which can be used for the subsequent development of detection kits; based on the expression of CK5 / 6 and PD-L1 immunohistochemistry, these two molecules have been widely used in clinical practice. High expression of CK5 / 6 + low expression of PD-L1 indicates a poor prognosis for neoadjuvant immunotherapy in patients with squamous cell lung carcinoma, and can be directly used to predict the efficacy of neoadjuvant immunotherapy for squamous cell lung carcinoma.

[0054] Based on the above 20 protein molecules, a patient survival prognostic model was constructed, and the risk score was calculated. According to the median risk score, the patients were divided into a high-risk group and a low-risk group. Among them, 16 molecules indicating poor prognosis (KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3, and TACSTD2) were highly expressed in the high-risk group, while 4 molecules indicating a good prognosis (CD74, HLA-DPA1, ICAM1, and VAV1) were highly expressed in the low-risk group. The survival prognosis of the high-risk group was significantly shorter than that of the low-risk group. Through time-dependent ROC analysis, the areas under the curve (AUC) of the 12-, 18-, and 24-month DFS of the prognostic model we constructed were 0.765, 0.785, and 0.711, respectively. Figure 4 shown.

[0055] The public dataset of the lung cancer immunotherapy cohort GSE135222 also proved that the survival rate of patients in the high-risk group was significantly lower than that in the low-risk group. The AUC of the model used to predict the 200-, 300-, and 400-day PFS in the GSE135222 cohort was as high as 0.971, 0.984, and 0.977, respectively. Figure 6 shown.

[0056] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. The use of protein molecular markers in the preparation of a product for predicting the effectiveness of neoadjuvant immunotherapy for patients with lung squamous cell carcinoma, characterized in that: The protein molecule markers include: at least one of CD74, HLA-DPA1, ICAM1, VAV1, KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2.

2. The use of protein molecular markers in the preparation of products for predicting poor prognosis of patients with lung squamous cell carcinoma after neoadjuvant immunotherapy, characterized in that: The protein molecule markers include: at least one of KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3 and TACSTD2.

3. The use of protein molecular markers in the preparation of products for predicting a good prognosis of patients with lung squamous cell carcinoma after neoadjuvant immunotherapy, characterized in that: The protein molecule markers include: at least one of CD74, HLA-DPA1, ICAM1 and VAV1.

4. Application of KRT5 / 6 combined with PD-L1 in the preparation of products for predicting poor prognosis of patients with squamous cell lung carcinoma after neoadjuvant immunotherapy.

5. The use according to any one of claims 1 to 4, characterized in that: The product includes a kit or a biochip.

6. A method for constructing a neoadjuvant immunotherapy efficacy prediction model for patients with lung squamous cell carcinoma, characterized in that: The following steps are involved: 1) Data collection: The residual tumor percentage (RVT) of the tumors resected after neoadjuvant therapy for squamous cell lung carcinoma was measured. CPR was the absence of viable tumor cells in both the primary lung lesion and lymph nodes. 2) Identify potential molecular subtypes in patients with squamous cell lung carcinoma: Identify potential molecular subtypes in patients with squamous cell lung carcinoma by performing consensus clustering on the protein expression matrix of the squamous cell lung carcinoma cohort; Identify potential molecular subtypes in patients with squamous cell lung carcinoma by clustering analysis of 346 differentially expressed proteins between CPR and non-CPR; 3) Weighted gene co-expression network analysis: All identified proteins are taken as input and soft thresholds are calculated to establish the basis of the co-expression network. On this basis, the co-expression network is constructed and genes are classified into different modules, and then the correlation between genes and phenotypes in the modules of interest is explored; 4) Mfuzz cluster analysis: The differentially expressed proteins among the three groups of RVT 0%, 1-10%, and 11-100% were clustered using the Mfuzz algorithm; 5) Correlation analysis: Correlation analysis between protein molecule expression level and RVT; 6) Build the model.

7. The construction method according to claim 6, characterized in that: In the step 2), the potential molecular subtypes in patients with lung squamous cell carcinoma are three proteomic subtypes: CCP_I, CCP_II and CCP_III.

8. The construction method according to claim 6 or 7, characterized in that: Among the modules of interest in step 3), the light cyan module is significantly positively correlated with the CCP_I subtype and negatively correlated with CPR; the cyan module and the green module are positively correlated with the CCP_II subtype and CPR.

9. The construction method according to claim 6, characterized in that: In step 4), cluster analysis obtains 6 clusters with similar trajectories, including cluster 1, cluster 4, and cluster 6.

10. The construction method according to claim 6, characterized in that: The model constructed in step 6) includes: taking the intersection of proteins whose expression levels are negatively correlated with RVT percentage, differentially expressed proteins (DEPs) significantly increased in CPR patients, proteins in green modules and cyan modules, proteins in cluster 1 in Mfuzz analysis, and significantly highly expressed DEPs in CCP_II subtype (fold change>2), and identifying four biomarkers indicating good prognosis, CD74, HLA-DPA1, ICAM1, and VAV1; at the same time, proteins whose expression levels are positively correlated with RVT, DEPs significantly decreased in CPR patients, shallow The proteins in the cyan module, the proteins in the 4 clusters and 6 clusters identified by Mfuzz, and the highly expressed DEPs in the CCP_I subtype (fold change > 2) were intersected to obtain 16 biomarkers indicating poor prognosis, including KRT3, KRT5, KRT6A, KRT6B, KRT6C, KRT16, KRT17, KRT20, KRT75, JUP, MAL2, PKP1, SPRR1A, SPRR1B, SPRR3, and TACSTD2. According to the disease-free survival (DFS) of the patients, a prognostic model was constructed and the risk score was calculated.

Images