Small cell lung cancer unsupervised subtyping system and immunotherapy prediction system thereof
Patent Information
- Application Number
- CN202610429188.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-04-02
- Publication Date
- 2026-06-26
Smart Images
Figure FT_1 
Figure SMS_1 
Figure SMS_2
Abstract
Description
Technical Field
[0001] This application belongs to the field of tumor analysis technology, specifically relating to an unsupervised subtyping system for small cell lung cancer and its immunotherapy prediction system. Background Technology
[0002] Small cell lung cancer (SCLC) is a highly malignant neuroendocrine tumor with a very poor prognosis, accounting for approximately 15% of all lung cancers. Although chemotherapy combined with immune checkpoint inhibitors (PD-(L)1 antibodies) can prolong median survival in patients with extensive-stage SCLC, the number of patients achieving long-term disease control remains limited. Currently, SCLC transcriptomic subtyping is mainly based on neurodevelopmental transcription factors, including ASCL1 (SCLC-A), NEUROD1 (SCLC-N), POU2F3 (SCLC-P), and SCLC-I, which exhibits inflammatory characteristics. Based on neuroendocrine characteristics, SCLC-I is further divided into two variants: SCLC-I-NE and SCLC-I-nNE. Among them, the inflammatory SCLC-I subtype (especially SCLC-I-NE) exhibits enhanced antigen presentation and CD8+. + Increased T-cell infiltration and decreased macrophage signaling, characteristic of immune activation, make patients more likely to benefit from chemotherapy combined with immunotherapy in clinical practice. However, these subtypes cannot explain why most SCLC patients do not respond sustainably to immunotherapy, nor do they reveal the upstream molecular mechanisms leading to immune heterogeneity.
[0003] Currently, the classification of SCLC is mainly based on neural transcription factors (such as ASCL1, NEUROD1, and POU2F3), but this method cannot effectively distinguish between immune-sensitive tumors and immune-rejecting tumors. The current classification method cannot accurately determine the effectiveness of immunotherapy strategies for SCLC patients. Summary of the Invention
[0004] The purpose of this invention is to provide an unsupervised classification system for small cell lung cancer. Based on integrating sex hormone signaling pathways (estrogen and androgen responses) and cholesterol metabolism pathways (cholesterol homeostasis) as sex hormone and cholesterol signaling features (Sex-Steroid & Cholesterol Signature, SSCS), and using the k-means clustering algorithm (k-means) unsupervised clustering algorithm to establish a tumor classification model at the transcriptome level, thereby stably obtaining three small cell lung cancer subtypes.
[0005] The present invention also aims to provide an unsupervised classification system for small cell lung cancer in predicting the prognostic efficacy of chemoimmunotherapy for patients with small cell lung cancer, thereby providing a basis for personalized treatment of patients with small cell lung cancer.
[0006] This invention provides an unsupervised typing system for small cell lung cancer, comprising the following connected functional modules: The data acquisition module is used to acquire gene expression data related to sex hormone and cholesterol signaling characteristics in user samples; The data analysis module is used to analyze the data acquired by the data acquisition module to obtain the small cell lung cancer classification results; The data acquisition method of the data acquisition module includes analysis based on a tumor subtyping model; the tumor subtyping model is constructed using the k-means unsupervised clustering algorithm. The pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway. The small cell lung cancer classification results include SSCS high subtype, SSCS medium subtype, and SSCS low subtype.
[0007] Preferably, the tumor subtyping model uses the Elbow method or Gap Statistic to determine the optimal number of clusters.
[0008] Preferably, the optimal number of clusters is 3.
[0009] Preferably, the classification of SSCS high subtype, SSCS medium subtype and SSCS low subtype is based on the relative position of the sample with respect to each cluster center in the enrichment score space of the four pathways; The small cell lung cancer classification results are entered into a pre-constructed tumor classification model based on the enrichment scores of the androgen response pathway, the early estrogen transcriptional response pathway, the late estrogen transcriptional response pathway, and the cholesterol homeostasis pathway, and then classified into the subtypes that are closest to the cluster centers in the training model. The tumor subtyping model is based on a four-dimensional pathway enrichment score matrix of training cohort samples and is established through k-means unsupervised clustering.
[0010] Preferably, the gene expression data related to sex hormone and cholesterol signaling features in the user sample are calculated using a gene variation analysis algorithm to determine the enrichment scores of each pathway in the sex hormone and cholesterol signaling features of the user sample.
[0011] Preferably, the enrichment scores of each pathway in the sex hormone and cholesterol signaling features of the user sample are obtained by Z-score normalization.
[0012] Preferably, it also includes a data output module for outputting the small cell lung cancer subtyping results obtained by the data analysis module to a display terminal.
[0013] This invention provides a system for predicting the efficacy of chemotherapy combined with immunotherapy for small cell lung cancer, including the unsupervised subtyping system for small cell lung cancer and a therapy analysis module; The therapy analysis module is used to predict the treatment effect of small cell lung cancer subtype obtained by the data analysis module: when the subtype is SSCS high subtype, it is determined that the sample belongs to immune rejection type tumor and has low benefit from chemotherapy combined with immunotherapy. When the classification result is SSCS intermediate subtype, the sample is judged to belong to the intermediate transitional type of tumor, and its immunotherapy benefit is between that of SSCS high subtype and SSCS low subtype. When the classification result is SSCS low subtype, the sample is judged to belong to immune-sensitive tumors, which are highly beneficial for chemotherapy combined with immunotherapy.
[0014] Preferably, the immunotherapy includes PD-1 inhibitors and / or PD-L1 inhibitors.
[0015] This invention provides a reagent for detecting gene expression data related to sex hormone and cholesterol signaling characteristics in the preparation of a small cell lung cancer subtyping test kit or a kit for predicting the prognosis and efficacy of immunotherapy for small cell lung cancer. The pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway.
[0016] This invention provides an unsupervised typing system for small cell lung cancer, which performs clustering typing based on gene expression data related to sex hormone and cholesterol signaling features in user samples. The pathways involved in these sex hormone and cholesterol signaling features include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway. Compared with traditional typing methods, the typing system provided by this invention has better immune prediction accuracy and can significantly distinguish between immunosensitive and immune-rejecting tumors, providing an effective predictive tool for immunotherapy of small cell lung cancer. Furthermore, the typing system shows consistent performance across multiple cohort datasets, indicating that the results are accurate and reliable. Attached Figure Description
[0017] Figure 1The results of hormone-related biological tests in small cell lung cancer are as follows: A shows the correlation of four MSigDB Hallmark pathways in the IMpower133 cohort; B shows the optimized result of the elbow analysis for the optimal number of clusters; C shows the enrichment pattern heatmap of different samples on the four pathways; D shows the PFS and OS results of different subtypes in patients receiving chemoimmunotherapy; E shows the PFS and OS results of different subtypes in patients receiving chemotherapy alone; F shows the optimized result of the elbow analysis for the optimal number of clusters in the PH cohort cohort; G shows the enrichment pattern heatmap of different samples on the four pathways in the PH cohort cohort; and H shows the OS results of different subtypes in patients in the PH cohort cohort. Detailed Implementation
[0018] This invention provides an unsupervised typing system for small cell lung cancer, comprising the following connected functional modules: The data acquisition module is used to acquire gene expression data related to sex hormone and cholesterol signaling characteristics in user samples; The data analysis module is used to analyze the data acquired by the data acquisition module to obtain the small cell lung cancer classification results; The data acquisition method of the data acquisition module includes analysis based on a tumor subtyping model; the tumor subtyping model is constructed using the k-means unsupervised clustering algorithm. The pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway. The small cell lung cancer classification results include SSCS high subtype, SSCS medium subtype, and SSCS low subtype.
[0019] The unsupervised typing system for small cell lung cancer provided by this invention includes a data acquisition module for acquiring gene expression data related to sex hormone and cholesterol signaling characteristics in user samples.
[0020] In this invention, the pathways involved in the sex hormone and cholesterol signaling characteristics include four MSigDB Hallmark pathways: the androgen response pathway (ANDROGEN_RESPONSE), the early estrogen transcriptional response pathway (ESTROGEN_RESPONSE_EARLY), the late estrogen transcriptional response pathway (ESTROGEN_RESPONSE_LATE), and the cholesterol homeostasis pathway (CHOLESTEROL_HOMEOSTASIS). The gene types related to the androgen response pathway preferably include the Hallmark gene set from the MSigDB database. Specifically, the early estrogen transcriptional response pathway refers to the "early response to estrogen" as defined in the MSigDB Hallmark gene set. The late estrogen transcriptional response pathway refers to the "early response to estrogen" as defined in the MSigDB Hallmark gene set.
[0021] In this embodiment of the invention, to analyze the correlation between the four pathways, Spearman correlation analysis was performed on the gene expression data of the four MSigDB Hallmark pathways, and pairwise correlation analysis was conducted on the ssGSEA scores of the four pathways. The results showed that in the IMpower133 dataset, the ssGSEA scores of the four pathways were strongly positively correlated, suggesting that they constitute a closely related functional module.
[0022] In this invention, the gene expression data related to sex hormones and cholesterol signaling characteristics includes transcriptome sequencing data (RNA-seq) from small cell lung cancer patients and preprocessed data. The preprocessed data is obtained by calculating the enrichment scores of user samples on the four pathways mentioned above using the GSVA algorithm based on the transcriptome sequencing data of small cell lung cancer patients, and then standardizing the enrichment scores of each pathway using Z-scores. Preferably, the preprocessed data is analyzed using the GSVA R package (v1.48.0) and the single-sample gene set enrichment analysis (ssGSEA) method to obtain ssGSEA values.
[0023] The unsupervised typing system for small cell lung cancer provided by the present invention includes a data analysis module, which is used to analyze the data acquired by the data acquisition module to obtain the typing results of small cell lung cancer.
[0024] In this invention, the method for acquiring data by the data acquisition module preferably includes analysis based on a tumor subtyping model. The data used to construct the tumor subtyping model is the patient's transcriptome sequencing data, preferably from a public database or a hospital-built clinical sample cohort. The public database preferably includes at least one of the following: TCGA, EGA, GEO, and cBioPortal. The data used to construct the tumor subtyping model is preferably preprocessed data, processed using the same method as the user sample data, to obtain enrichment scores for four pathways in different samples, forming an n×4 dimension feature matrix. During tumor subtyping model construction, the preprocessed data is preferably input in TPM or FPKM format.
[0025] In this invention, the tumor subtyping model is preferably constructed using k-means unsupervised clustering algorithm on the data acquired by the data acquisition module. The optimal number of clusters in the tumor subtyping model is preferably determined using Elbow's method or Gap Statistic. The optimal number of clusters is preferably 3. The small cell lung cancer subtyping results preferably include the following three types: high SSCS subtype, medium SSCS subtype, and low SSCS subtype. The small cell lung cancer subtyping results are categorized into the subtype closest to the cluster center in the training model after being input into the pre-constructed tumor subtyping model based on the enrichment scores of the androgen response pathway, early estrogen transcriptional response pathway, late estrogen transcriptional response pathway, and cholesterol homeostasis pathway. The tumor subtyping model is established based on the four-dimensional pathway enrichment score matrix of the training cohort samples through k-means unsupervised clustering. The high SSCS subtype shows significantly enhanced hormone and cholesterol signals, immune rejection characteristics, and low PD-L1 expression. The low SSCS subtype shows increased MHC-I and CD8 expression. + High T-cell infiltration indicates effective immunotherapy. For the test sample, the enrichment scores of the four pathways mentioned above are first calculated, and the sample is then classified into the closest subtype based on its distance from the cluster centers in the training model. Therefore, the classification of SSCS high subtype, SSCS medium subtype, and SSCS low subtype is based on "the relative position of the sample to each cluster center in the enrichment score space of the four pathways".
[0026] The unsupervised typing system for small cell lung cancer provided by the present invention preferably further includes a data output module for outputting the small cell lung cancer typing results obtained by the data analysis module to a display terminal.
[0027] The present invention does not impose any special restrictions on the type of display terminal; any terminal device well known in the art can be used, such as a mobile phone, a computer, or a test report printer.
[0028] This invention provides a system for predicting the efficacy of chemotherapy combined with immunotherapy for small cell lung cancer, including the unsupervised subtyping system for small cell lung cancer and a therapy analysis module; The therapy analysis module is used to predict the treatment effect of small cell lung cancer subtype obtained by the data analysis module: when the subtype is SSCS high subtype, it is determined that the sample belongs to immune rejection type tumor and has low benefit from chemotherapy combined with immunotherapy. When the classification result is SSCS intermediate subtype, the sample is judged to belong to the intermediate transitional type of tumor, and its immunotherapy benefit is between that of SSCS high subtype and SSCS low subtype. When the classification result is SSCS low subtype, the sample is judged to belong to immune-sensitive tumors, which are highly beneficial for chemotherapy combined with immunotherapy.
[0029] In this invention, the immunotherapy preferably includes a PD-1 inhibitor and / or a PD-L1 inhibitor. The chemotherapy method preferably includes etoposide in combination with platinum-based drugs.
[0030] In this invention, in the small cell lung cancer chemotherapy combined with immunotherapy efficacy prediction system, the data output module of the small cell lung cancer unsupervised subtyping system is used to output the prognostic assessment results obtained by the prognostic analysis module to the terminal.
[0031] In this invention, the small cell lung cancer chemotherapy combined with immunotherapy efficacy prediction system is used to predict whether small cell lung cancer patients are suitable for immunotherapy based on the subtype results, providing a basis and reference for personalized treatment of clinical patients.
[0032] This invention provides a reagent for detecting gene expression data related to sex hormone and cholesterol signaling characteristics in the preparation of a small cell lung cancer subtyping test kit or a kit for predicting the prognosis and efficacy of immunotherapy for small cell lung cancer. The pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway.
[0033] In this invention, the reagents for detecting gene expression data related to sex hormone and cholesterol signaling characteristics preferably include transcriptome sequencing reagents or RT-qPCR reagents for detecting gene expression levels related to sex hormone and cholesterol signaling characteristics. The kit is used to detect gene expression levels related to sex hormone and cholesterol signaling characteristics. Based on the gene expression levels, four pathway scores are calculated using the GSVA algorithm. These scores are then Z-score standardized to obtain the enrichment scores for the four pathways in the user sample, i.e., the single-sample gene set enrichment analysis score (ssGSEA). A tumor subtyping model constructed based on the optimal number of clusters determined by the k-means clustering algorithm and Elbow method is used to subtype the ssGSEA scores. When the subtyping result is a high SSCS subtype, it indicates that the user sample belongs to an immune-rejecting tumor, and concurrent immunotherapy is not recommended. When the subtyping result is a low SSCS subtype, it indicates that the user sample belongs to an immune-sensitive tumor, and concurrent immunotherapy is recommended.
[0034] The following detailed description of the unsupervised typing system for small cell lung cancer and its immunotherapy prediction system provided by the present invention, with reference to the embodiments, should not be construed as limiting the scope of protection of the present invention.
[0035] Example 1 Predicting the efficacy of chemotherapy combined with immunotherapy in patients with small cell lung cancer (SCLC) Somatic RNA-seq gene expression analysis 1.1 Dataset Source Description RNA-seq and matched clinical data from the IMpower133 cohort were obtained from the European Genome Phenotypic Archive (EGA), including SCLC patients receiving chemotherapy combined with immunotherapy and SCLC patients receiving chemotherapy alone. Data from the PH cohort were generated from a Phase II clinical trial conducted at Shanghai Pulmonary Hospital (Yu, J., et al., Camrelizumab, an Anti-PD-1 Monoclonal Antibody, Plus Carboplatin and Nab-Paclitaxel as First-Line Setting for Extensive-Stage Small-Cell Lung Cancer: A Phase 2 Trial and Biomarker Analysis. MedComm, 2025. 6(8): p. e70300.).
[0036] 1.2 Analytical Methods Progression-free survival (PFS) and overall survival (OS) were estimated using the Kaplan-Meier method, and differences between clusters were assessed using the log-rank test. Cox proportional hazards regression models were used to calculate hazard ratios (HRs) for pairwise cluster comparisons with 95% confidence intervals. Forest plots were generated to visualize HRs and CIs. All analyses were performed using the survival and survminer R packages.
[0037] 1.3 Results To capture hormone-related biological features in SCLC, we focused on four MSigDB Hallmark pathways directly related to sex hormone signaling and sterol metabolism: androgen response (ANDROGEN_RESPONSE), early estrogen response (ESTROGEN_RESPONSE_EARLY), late estrogen response (ESTROGEN_RESPONSE_LATE), and cholesterol homeostasis (CHOLESTEROL_HOMEOSTASIS).
[0038] To assess the consistency of these pathways within the cohort, the four pathways were integrated into a "Sex Hormone-Cholesterol Characteristic Profile (SSCS)," and the GSVA algorithm was used to calculate the enrichment score of each sample on the four pathways in the IMpower-133 tumor transcriptome data. The specific method is as follows: RNA-seq expression matrices and matched clinical information of small cell lung cancer patients from the IMpower133 and PH cohorts were collected. After standardizing the expression matrices, gene sets corresponding to the four Hallmark pathways ANDROGEN_RESPONSE, ESTROGEN_RESPONSE_EARLY, ESTROGEN_RESPONSE_LATE, and CHOLESTEROL_HOMEOSTASIS were extracted from the MSigDB database. Single-sample gene set enrichment analysis was performed on each sample using the GSVA R package to obtain the ssGSEA scores for the four pathways. Z-score standardization was further applied to each pathway score to construct a sample × 4-dimensional pathway score matrix. Based on the 4-dimensional pathway score matrix, k-means unsupervised clustering was used for genotyping, and the optimal number of clusters k was determined using the elbow method. The sample was divided into three subtypes: SSCS-high, SSCS-intermediate, and SSCS-low. Further, combining patient clinical follow-up data, the Kaplan-Meier method was used to assess the differences in progression-free survival and overall survival among the subtypes, and the log-rank test was used to evaluate statistical significance.
[0039] Next, GSVA was performed to quantify pathway activity at both global and single-cell resolution. For the large RNA-seq dataset, normalized expression matrices were analyzed using the GSVA R package (v1.48.0) and the single-sample gene set enrichment analysis (ssGSEA) method. Four hormone-related Hallmark pathways—androgen response, early estrogen response, late estrogen response, and cholesterol homeostasis—from the MSigDB database (v2025.1) were used as predefined gene sets. A score for each pathway was calculated for each patient sample, Z-score normalized for each pathway score, and visualized as a heatmap.
[0040] Pairwise correlation analysis (typically Spearman correlation) was performed on the ssGSEA scores of the four pathways, and the correlation structure among the pathways was displayed using a correlation heatmap. The results showed that the ssGSEA scores of these pathways were strongly positively correlated in the IMpower133 dataset. Figure 1 (A) indicates that it constitutes a closely related functional module.
[0041] 2. Unsupervised clustering and subtype definition Unsupervised clustering was performed on both queues based on the 4-dimensional ssGSEA scores. To determine the optimal number of clusters k, the total within-sum of squares was calculated for different k values, and the elbow method was used to select the k value at the inflection point.
[0042] Elbow analysis supports the classification of patients into three categories ( Figure 1 (Medium B). Subsequently, the 4-dimensional SSGSEA score matrix (sample × 4 pathways) of each sample is used as the "SSCS feature space", and a heatmap is used to show the enrichment patterns of different samples on the 4 pathways, namely SSCS-low, SSCS-medium, and SSCS-high. Figure 1 In the middle (C), the color bar is ssGSEA score).
[0043] 3. Survival analysis and the relationship between subtyping To test the clinical relevance of SSCS subtyping, the IMpower-133 cohort was stratified by treatment regimen: Chemotherapy combined with immunotherapy subgroup: PFS and OS were compared among the three subtypes of SSCS-low, SSCS-intermediate and SSCS-high. Kaplan-Meier curves were used to illustrate the differences between groups, and the log-rank test was used to assess the differences between groups. The p-value / HR for pairwise comparisons between groups was also given.
[0044] Chemotherapy-only subgroup: The PFS and OS of the above three subtypes were compared using the same method to determine whether the subtype is more biased towards "immunotherapy-related stratification" rather than simple prognostic stratification.
[0045] In clinical association analysis, among patients receiving chemotherapy combined with immunotherapy, the SSCS-low, SSCS-intermediate, and SSCS-high subtypes showed significantly shortened progression-free survival (PFS) and overall survival (OS). Figure 1 In the middle D), there was no such difference in the chemotherapy-only cohort ( ). Figure 1 The subtype SSCS (as indicated by the Chinese E-type subtype) suggests that the subtype characteristics of the Sanya type are immune-related biological features rather than chemotherapy sensitivity.
[0046] The classification results for some patients are shown in Table 1.
[0047] Table 1. Classification results of some patients
[0048] PFS_months: Progress-free survival, in months; PFS_censor: Censored data for Progress-Free Survival (PFS); OS_months: Total lifetime; OS_censor: Censored / truncated data for the entire operating system (OS).
[0049] 4. Verification within the independent team In an independent PH cohort (n=41), the Kaplan–Meier and log-rank analyses were repeated for patients receiving chemotherapy combined with immunotherapy to verify the stratification ability of SSCS classification on OS.
[0050] The three subtypes mentioned above were successfully reproduced in the independent validation queue (PH queue). Figure 1 Patients with high SSCS (F~G) levels still showed the worst overall survival (OS) even when receiving combination therapy. Figure 1 The H subtype supports the predictive value of this classification for the efficacy of immunotherapy.
[0051] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.
Claims
1. An unsupervised subtyping system for small cell lung cancer, the system comprising: Includes the following connected functional modules: The data acquisition module is used to acquire gene expression data related to sex hormone and cholesterol signaling characteristics in user samples; The data analysis module is used to analyze the data acquired by the data acquisition module to obtain the small cell lung cancer classification results; The data acquisition method of the data acquisition module includes analysis based on a tumor subtyping model; the tumor subtyping model is constructed using the k-means unsupervised clustering algorithm. The pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway. The small cell lung cancer classification results include SSCS high subtype, SSCS medium subtype, and SSCS low subtype.
2. The unsupervised subtyping system of claim 1, wherein, The tumor subtyping model uses the Elbow method or Gap Statistic to determine the optimal number of clusters.
3. The unsupervised subtyping system of claim 2, wherein, The optimal number of clusters is 3.
4. The unsupervised typing system for small cell lung cancer according to claim 1, characterized in that, The classification of SSCS high subtype, SSCS medium subtype and SSCS low subtype is based on the relative position of the sample with respect to each cluster center in the enrichment score space of the four pathways; The small cell lung cancer classification results are entered into a pre-constructed tumor classification model based on the enrichment scores of the androgen response pathway, the early estrogen transcriptional response pathway, the late estrogen transcriptional response pathway, and the cholesterol homeostasis pathway, and then classified into the subtypes that are closest to the cluster centers in the training model. The tumor subtyping model is based on the four-dimensional pathway enrichment score matrix of the training cohort samples and is established through k-means unsupervised clustering.
5. The unsupervised typing system for small cell lung cancer according to claim 1, characterized in that, The gene expression data related to sex hormone and cholesterol signaling features in the user samples were calculated using a gene variation analysis algorithm to determine the enrichment scores of each pathway in the sex hormone and cholesterol signaling features of the user samples.
6. The unsupervised typing system for small cell lung cancer according to claim 5, characterized in that, The enrichment scores of each pathway in the sex hormone and cholesterol signaling features of the user samples were obtained by Z-score normalization.
7. The unsupervised typing system for small cell lung cancer according to any one of claims 1 to 6, characterized in that, It also includes a data output module, which is used to output the small cell lung cancer subtyping results obtained by the data analysis module to the display terminal.
8. A system for predicting the efficacy of chemotherapy combined with immunotherapy in small cell lung cancer, characterized in that, Includes the unsupervised subtyping system and therapy analysis module for small cell lung cancer as described in any one of claims 1 to 7; The therapy analysis module is used to predict the treatment effect of small cell lung cancer subtype obtained by the data analysis module: when the subtype is SSCS high subtype, it is determined that the sample belongs to immune rejection type tumor and has low benefit from chemotherapy combined with immunotherapy. When the classification result is SSCS intermediate subtype, the sample is judged to belong to the intermediate transitional type of tumor, and its immunotherapy benefit is between that of SSCS high subtype and SSCS low subtype. When the subtype result is SSCS low subtype, the sample is judged to be an immunosensitive tumor, which is more beneficial to chemotherapy combined with immunotherapy.
9. The small cell lung cancer chemotherapy combined with immunotherapy efficacy prediction system according to claim 8, characterized in that, The immunotherapy includes PD-1 inhibitors and / or PD-L1 inhibitors.
10. The application of a reagent for detecting gene expression data related to sex hormone and cholesterol signaling characteristics in the preparation of a small cell lung cancer subtyping test kit or a kit for predicting the prognostic efficacy of immunotherapy for small cell lung cancer, wherein the pathways involved in the sex hormone and cholesterol signaling characteristics include the androgen response pathway, the early estrogen transcriptional response pathway, the cholesterol homeostasis pathway, and the late estrogen transcriptional response pathway.