Scoring system for predicting prognosis and immunotherapy response rate in melanoma patients

Abstract

The application discloses a scoring system which can predict the prognosis and immunotherapy response rate of melanoma patients. Based on the expression level of GPRs of melanoma patients and the infiltration level of immune cells in the TME, a GPRs-TME scoring system is constructed, which can better classify melanoma patients, accurately predict their prognosis and immunotherapy response rate. It makes up for the shortcomings of existing melanoma patient classification based on TMN staging, Breslow thickness and other clinical characteristics, and provides a certain reference for the precise treatment of melanoma. The application is simple and cost-effective, and only requires routine transcriptome sequencing of tumor tissue to complete the construction of the score. The application is stable and reliable, and has been tested on more than 1000 melanoma patients from five melanoma cohorts. The prediction performance of the application is good, and there are significant differences in the prognosis and immunotherapy efficacy of the three groups of patients after the construction of the score.

Description

Technical Field

[0001] This invention relates to a system capable of molecularly subtyping melanoma patients and predicting their prognosis and response rate to immunotherapy. Background Technology

[0002] Melanoma is the most malignant skin tumor, with its incidence and mortality rates increasing annually. Although treatment options include surgery, chemotherapy, and targeted therapies, efficacy varies among patients, resulting in generally poor outcomes. Immunotherapy, represented by immune checkpoint inhibitors, has significantly improved the prognosis of patients with advanced melanoma; however, response rates to immune checkpoint inhibitors differ among melanoma patients, and only a small percentage of patients benefit from it. Currently, traditional clinical characteristics such as TMN staging and Breslow thickness cannot predict the response rate to immunotherapy in melanoma patients. Therefore, scientists are currently dedicated to exploring novel biomarkers that can predict the prognosis and efficacy of immunotherapy in melanoma patients.

[0003] G protein-coupled receptors (GPRs) are a large class of membrane protein receptors that play a crucial role in tumorigenesis and metastasis by binding to chemicals around the cell and activating a series of intracellular signaling pathways, ultimately leading to alterations in cell state. Due to their widespread distribution in the body, GPRs have become important targets for drug therapy; 34% of drugs approved by the U.S. Food and Drug Administration (FDA) target GPRs. GPRs can also influence the tumor microenvironment (TME) by recruiting immune cells and stromal cells, ultimately affecting the prognosis of cancer patients. However, systematic and mature studies are currently lacking to explore how GPRs affect the TME and immunotherapy response in melanoma.

[0004] The advent of high-throughput sequencing technologies, especially single-cell RNA sequencing (scRNA-seq), has ushered in an era of precision in tumor research. The combination of scRNA-seq and traditional transcriptome sequencing (bulk-seq) makes it possible to jointly explore the role of individual genes in tumorigenesis and progression at both the tissue and single-cell levels. Summary of the Invention

[0005] The purpose of this invention is to improve the current classification method for melanoma patients; to accurately predict the prognosis of melanoma patients; and to accurately predict the response of melanoma patients to immune checkpoint inhibitors.

[0006] To achieve the above objectives, the technical solution of the present invention provides a scoring system that can predict the prognosis and immunotherapy response rate of melanoma patients, characterized in that it includes:

[0007] Prognostic gene identification module: Used to identify key prognostic genes affecting melanoma patients based on scRNA-seq, bulk-seq, genomic data and clinical prognostic data, through machine learning algorithms and multi-cohort validation;

[0008] Prognostic immune cell identification module: Based on scRNA-seq, bulk-seq, genomic data and clinical prognostic data, this module uses machine learning algorithms and multi-cohort validation to identify key prognostic immune cells affecting melanoma patients.

[0009] GPR scoring module: used to calculate the GPR score for each patient. The GPR score for each patient is the sum of the products of the expression levels of all GPRs in the current patient as determined by the prognostic gene determination module and their Bootstrap coefficients.

[0010] TME scoring module: used to calculate the TME score for each patient. The TME score for each patient is the sum of the products of the infiltration level of all immune cells in the current patient, as determined by the prognostic immune cell determination module, and their Bootstrap coefficients.

[0011] The combined grouping module, based on the cut-off values ​​of the GPR and TME scores obtained from the training set, assigns the current patient to a high GPR / low TME group, a low GPR / high TME group, or a mixed group, according to the current patient's GPR score obtained through the GPR scoring module and the current patient's TME score obtained through the TME scoring module. Specifically: if the current patient is assigned to the high GPR / low TME group, it indicates the worst prognosis and the lowest response rate to immunotherapy; if the current patient is assigned to the low GPR / high TME group, it indicates the best prognosis and the highest response rate to immunotherapy.

[0012] Preferably, the prognostic gene determination module identified 12 GPRs as key prognostic genes affecting melanoma patients, including ADGRE5, ADGRG5, FZD6, GAL, GPR143, GPR171, GPR84, GPR85, NLRP6, SSTR2, TAPT1, and TSHR.

[0013] Preferably, the prognostic immune cell determination module identifies five types of immune cells as key prognostic immune cells affecting melanoma patients, including M1 Macrophages, activated NK cells, activated CD4 memory T cells, CD8 T cells, and gamma delta T cells.

[0014] Preferably, it also includes a GPR grouping module: based on the cut-off value of the GPR score obtained from the training set, the current patient is divided into a high GPR score group or a low GPR score group according to the current patient's GPR score obtained through the GPR scoring module.

[0015] Preferably, it also includes a TME grouping module: based on the cut-off value of the TME score obtained from the training set, the current patient is divided into a high TME score group or a low TME score group according to the current patient's TME score obtained through the TME scoring module.

[0016] Preferably, the hybrid group includes a high GPR high TME group and a low GPR low TME group.

[0017] This invention constructs a GPRs-TME scoring system based on the expression levels of GPRs in melanoma patients and the infiltration levels of immune cells in their TME (tumor vena cava). This system can better classify melanoma patients molecularly and accurately predict their prognosis and immunotherapy response rate. It overcomes the shortcomings of existing classification methods for melanoma patients based on clinical characteristics such as TMN staging and Breslow thickness, providing a certain reference for precision treatment of melanoma.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] 1) This invention is simple and cost-effective, requiring only routine transcriptome sequencing of the patient's tumor tissue to construct the scoring system;

[0020] 2) This invention is stable and reliable, having been tested on more than 1,000 melanoma patients in five major melanoma cohorts (TCGA-SKCM, GSE65904, GSE91061, GSE145996 and GSE35640).

[0021] 3) The present invention has good predictive efficacy, and there are significant differences in the prognosis and immunotherapy efficacy among the three groups of patients after the scoring is constructed. Detailed Implementation

[0022] The present invention will be further illustrated below with reference to specific embodiments. It should be understood that these embodiments are for illustrative purposes only and are not intended to limit the scope of the invention. Furthermore, it should be understood that after reading the teachings of this invention, those skilled in the art can make various alterations or modifications to the invention, and these equivalent forms also fall within the scope defined by the appended claims.

[0023] The GPR family comprises approximately 800 proteins, and previous studies have confirmed their crucial roles in tumorigenesis and the tumor immune microenvironment. However, the role of GPRs in melanoma development and how they influence the tumor microenvironment (TME) and consequently, the efficacy of immunotherapy, remains unknown. This invention utilizes multiple machine learning algorithms, including LASSO, COX, and Bootstrap, to identify the most critical GPRs and immune cells influencing the prognosis and immunotherapy response rate of melanoma patients. By analyzing each patient's GPR expression levels and immune cell infiltration levels, combined with the Bootstrap coefficient, a GPR and TME score can be derived for each melanoma patient, categorizing them into high GPR / low TME, low GPR / high TME, and mixed groups. Based on the analysis of over 1000 melanoma patients in five major melanoma cohorts (TCGA-SKCM, GSE65904, GSE91061, GSE145996, and GSE35640), the low GPR and high TME group showed the best prognosis and the highest response rate to immunotherapy, while the high GPR and low TME group showed the worst prognosis and the lowest response rate to immunotherapy. The scoring system disclosed in this invention can serve as a highly efficient predictor of the prognosis and efficacy of immunotherapy for melanoma patients.

[0024] Specifically, the technical solutions disclosed in this invention include the following:

[0025] Whole-genome expression matrices and prognostic data of melanoma patients were downloaded from the TCGA-SKCM and GSE65904 cohorts. The data were preprocessed, and expression matrices of all GPRs were extracted. Based on scRNA-seq, bulk-seq, genomic data, and clinical prognostic data, univariate Cox regression combined with bootstrap analysis, LASSO regression analysis, and multivariate Cox regression combined with bootstrap analysis were used to identify GPRs closely related to the prognosis of melanoma patients. Twelve GPRs—ADGRE5, ADGRG5, FZD6, GAL, GPR143, GPR171, GPR84, GPR85, NLRP6, SSTR2, TAPT1, and TSHR—were identified as key prognostic genes influencing melanoma patients, and their corresponding bootstrap coefficients were obtained.

[0026] Constructing the GPR score: Calculate the GPR score for each patient. The GPR score is the sum of the products of the expression level of each included GPR and its Bootstrap coefficient. Based on the median GPR score, patients are divided into high / low GPR score groups.

[0027] Based on the patient's whole genome expression matrix, the infiltration levels of 22 immune cells in the patient's total melanoma genome (TME) were calculated using the CIBERSORT deconvolution algorithm. Survival analysis was then used to identify the immune cell types associated with prognosis. In this invention, five immune cell types—M1 Macrophages, activated NK cells, activated CD4 memory T cells, CD8 T cells, and gammadelta T cells—were identified as key prognostic immune cells affecting melanoma patients. Their corresponding bootstrap coefficients were calculated using multivariate Cox regression combined with bootstrap analysis.

[0028] TME score construction: The TME score for each patient was calculated as the sum of the products of the infiltration level of each included immune cell and its Bootstrap coefficient. Patients were then assigned to high / low TME score groups based on the median TME score.

[0029] By combining GPR and TME scores and using the optimal cut-off values ​​for GPR and TME scores, melanoma patients were divided into a high GPR-low TME group, a low GPR-high TME group, and a mixed group (high GPR-high TME, low GPR-low TME). Survival analysis revealed that the low GPR-high TME group had the best prognosis and the highest response rate to immunotherapy, while the high GPR-low TME group had the worst prognosis and the lowest response rate to immunotherapy.

[0030] Whole-genome expression matrices and immunotherapy response data of melanoma patients were downloaded from the GSE91061, GSE145996, and GSE35640 cohorts. GPR and TME scores were constructed as described above, and patients were divided into three subgroups. The differences in immunotherapy response rates among the three subgroups were compared. It was found that the low GPR / high TME group had the highest immunotherapy response rate, while the high GPR / low TME group had the lowest. Therefore, the combined GPR and TME scores can be used to predict the prognosis of melanoma patients and their response rate to immunotherapy.

[0031] One implementation of the above technical solution includes the following steps:

[0032] Step 1: Obtain tumor tissue from the melanoma patient and perform traditional transcriptome sequencing to obtain the patient's whole-genome expression matrix.

[0033] Step 2: Extract the expression level matrix of the patient's 12 GPRs (ADGRE5, ADGRG5, FZD6, GAL, GPR143, GPR171, GPR84, GPR85, NLRP6, SSTR2, TAPT1, and TSHR), and calculate the patient's GPR score based on the Bootstrap coefficient obtained from the training set.

[0034] Step 3: Based on the patient's whole gene expression matrix, the infiltration levels of 22 immune cells in the patient's TME are calculated using the CIBERSORT deconvolution algorithm. The infiltration levels of five immune cells, namely M1 Macrophages, activated NK cells, activated CD4 memory T cells, CD8 T cells, and gamma delta T cells, are extracted. The patient's TME score is calculated based on the Bootstrap coefficient obtained from the training set.

[0035] Step 4: Based on the cut-off value of the GPR score obtained from the training set, patients are divided into high / low GPR score groups; based on the cut-off value of the TME score obtained from the training set, patients are divided into high / low TME score groups. Further, patients are divided into high GPR / low TME groups, low GPR / high TME groups, and mixed groups.

[0036] Step 5: Based on the results of the training set, predict the patient's prognosis and immunotherapy response rate.

Claims

1. A scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients, characterized in that, include: Prognostic gene identification module: Used to identify key prognostic genes affecting melanoma patients based on scRNA-seq, bulk-seq, genomic data and clinical prognostic data, through machine learning algorithms and multi-cohort validation using melanoma cohorts; Prognostic immune cell identification module: Based on scRNA-seq, bulk-seq, genomic data and clinical prognostic data, this module uses machine learning algorithms and multi-cohort validation to identify key prognostic immune cells affecting melanoma patients. GPR scoring module: used to calculate the GPR score for each patient. The GPR score for each patient is the sum of the products of the expression levels of all GPRs in the current patient as determined by the prognostic gene determination module and their corresponding Bootstrap coefficients. TME scoring module: used to calculate the TME score for each patient. The TME score for each patient is the sum of the products of the infiltration level of all immune cells in the current patient as determined by the prognostic immune cell determination module and their corresponding Bootstrap coefficients. Joint grouping module: Based on the median GPR score and median TME score obtained from the training set, the current patient is divided into a high GPR-low TME group, a low GPR-high TME group, or a mixed group according to the current patient's GPR score obtained through the GPR scoring module and the current patient's TME score obtained through the TME scoring module. Where: if the current patient is divided into the high GPR-low TME group, it means that the current patient has the worst prognosis and the lowest response rate to immunotherapy; if the current patient is divided into the low GPR-high TME group, it means that the current patient has the best prognosis and the highest response rate to immunotherapy.

2. The scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients as described in claim 1, characterized in that, The prognostic gene determination module identified 12 GPRs as key prognostic genes affecting melanoma patients, including ADGRE5, ADGRG5, FZD6, GAL, GPR143, GPR171, GPR84, GPR85, NLRP6, SSTR2, TAPT1, and TSHR.

3. The scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients as described in claim 1, characterized in that, The prognostic immune cell identification module identified five types of immune cells as key prognostic immune cells affecting melanoma patients, including M1 Macrophages, activated NK cells, activated CD4 memory T cells, CD8 T cells, and gamma delta T cells.

4. The scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients as described in claim 1, characterized in that, It also includes a GPR grouping module: based on the median GPR score obtained from the training set, the current patient is divided into a high GPR score group or a low GPR score group according to the current patient's GPR score obtained through the GPR scoring module.

5. The scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients as described in claim 1, characterized in that, It also includes a TME grouping module: based on the median TME score obtained from the training set, the current patient is divided into a high TME score group or a low TME score group according to the current patient's TME score obtained through the TME scoring module.

6. The scoring system for predicting the prognosis and immunotherapy response rate of melanoma patients as described in claim 1, characterized in that, The hybrid groups include a high GPR high TME group and a low GPR low TME group.