Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods
BackgroundSmall-cell lung cancer (SCLC) usually presents as an extensive disease with a poor prognosis at the time of diagnosis. Exosomes are rich in biological information and have a powerful impact on tumor progression and metastasis. Therefore, this study aimed to screen for diagnostic markers of...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Frontiers Media S.A.
2022-12-01
|
| Series: | Frontiers in Oncology |
| Subjects: | |
| Online Access: | https://www.frontiersin.org/articles/10.3389/fonc.2022.1077118/full |
| _version_ | 1797979343210676224 |
|---|---|
| author | Kun Zhang Chaoguo Zhang Ke Wang Xiuli Teng Mingwei Chen |
| author_facet | Kun Zhang Chaoguo Zhang Ke Wang Xiuli Teng Mingwei Chen |
| author_sort | Kun Zhang |
| collection | DOAJ |
| description | BackgroundSmall-cell lung cancer (SCLC) usually presents as an extensive disease with a poor prognosis at the time of diagnosis. Exosomes are rich in biological information and have a powerful impact on tumor progression and metastasis. Therefore, this study aimed to screen for diagnostic markers of blood exosomes in SCLC patients and to build a prognostic model.MethodsWe identified blood exosome differentially expressed (DE) RNAs in the exoRBase cohort and identified feature RNAs by the LASSO, Random Forest, and SVM-REF three algorithms. Then, we identified DE genes (DEGs) between SCLC tissues and normal lung tissues in the GEO cohort and obtained exosome-associated DEGs (EDEGs) by intersection with exosomal DEmRNAs. Finally, we performed univariate Cox, LASSO, and multivariate Cox regression analyses on EDEGs to construct the model. We then compared the patients’ overall survival (OS) between the two risk groups and assessed the independent prognostic value of the model using receiver operating characteristic (ROC) curve analysis.ResultsWe identified 952 DEmRNAs, 210 DElncRNAs, and 190 DEcircRNAs in exosomes and identified 13 feature RNAs with good diagnostic value. Then, we obtained 274 EDEGs and constructed a risk model containing 7 genes (TBX21, ZFHX2, HIST2H2BE, LTBP1, SIAE, HIST1H2AL, and TSPAN9). Low-risk patients had a longer OS time than high-risk patients. The risk model can independently predict the prognosis of SCLC patients with the areas under the ROC curve (AUCs) of 0.820 at 1 year, 0.952 at 3 years, and 0.989 at 5 years.ConclusionsWe identified 13 valuable diagnostic markers in the exosomes of SCLC patients and constructed a new promising prognostic model for SCLC. |
| first_indexed | 2024-04-11T05:37:26Z |
| format | Article |
| id | doaj.art-d967a76f278943a5b5f25b532bf215bc |
| institution | Directory Open Access Journal |
| issn | 2234-943X |
| language | English |
| last_indexed | 2024-04-11T05:37:26Z |
| publishDate | 2022-12-01 |
| publisher | Frontiers Media S.A. |
| record_format | Article |
| series | Frontiers in Oncology |
| spelling | doaj.art-d967a76f278943a5b5f25b532bf215bc2022-12-22T11:45:06ZengFrontiers Media S.A.Frontiers in Oncology2234-943X2022-12-011210.3389/fonc.2022.10771181077118Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methodsKun ZhangChaoguo ZhangKe WangXiuli TengMingwei ChenBackgroundSmall-cell lung cancer (SCLC) usually presents as an extensive disease with a poor prognosis at the time of diagnosis. Exosomes are rich in biological information and have a powerful impact on tumor progression and metastasis. Therefore, this study aimed to screen for diagnostic markers of blood exosomes in SCLC patients and to build a prognostic model.MethodsWe identified blood exosome differentially expressed (DE) RNAs in the exoRBase cohort and identified feature RNAs by the LASSO, Random Forest, and SVM-REF three algorithms. Then, we identified DE genes (DEGs) between SCLC tissues and normal lung tissues in the GEO cohort and obtained exosome-associated DEGs (EDEGs) by intersection with exosomal DEmRNAs. Finally, we performed univariate Cox, LASSO, and multivariate Cox regression analyses on EDEGs to construct the model. We then compared the patients’ overall survival (OS) between the two risk groups and assessed the independent prognostic value of the model using receiver operating characteristic (ROC) curve analysis.ResultsWe identified 952 DEmRNAs, 210 DElncRNAs, and 190 DEcircRNAs in exosomes and identified 13 feature RNAs with good diagnostic value. Then, we obtained 274 EDEGs and constructed a risk model containing 7 genes (TBX21, ZFHX2, HIST2H2BE, LTBP1, SIAE, HIST1H2AL, and TSPAN9). Low-risk patients had a longer OS time than high-risk patients. The risk model can independently predict the prognosis of SCLC patients with the areas under the ROC curve (AUCs) of 0.820 at 1 year, 0.952 at 3 years, and 0.989 at 5 years.ConclusionsWe identified 13 valuable diagnostic markers in the exosomes of SCLC patients and constructed a new promising prognostic model for SCLC.https://www.frontiersin.org/articles/10.3389/fonc.2022.1077118/fullsmall-cell lung cancerexosomemachine learningdiagnostic markersprognostic model |
| spellingShingle | Kun Zhang Chaoguo Zhang Ke Wang Xiuli Teng Mingwei Chen Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods Frontiers in Oncology small-cell lung cancer exosome machine learning diagnostic markers prognostic model |
| title | Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods |
| title_full | Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods |
| title_fullStr | Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods |
| title_full_unstemmed | Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods |
| title_short | Identifying diagnostic markers and constructing a prognostic model for small-cell lung cancer based on blood exosome-related genes and machine-learning methods |
| title_sort | identifying diagnostic markers and constructing a prognostic model for small cell lung cancer based on blood exosome related genes and machine learning methods |
| topic | small-cell lung cancer exosome machine learning diagnostic markers prognostic model |
| url | https://www.frontiersin.org/articles/10.3389/fonc.2022.1077118/full |
| work_keys_str_mv | AT kunzhang identifyingdiagnosticmarkersandconstructingaprognosticmodelforsmallcelllungcancerbasedonbloodexosomerelatedgenesandmachinelearningmethods AT chaoguozhang identifyingdiagnosticmarkersandconstructingaprognosticmodelforsmallcelllungcancerbasedonbloodexosomerelatedgenesandmachinelearningmethods AT kewang identifyingdiagnosticmarkersandconstructingaprognosticmodelforsmallcelllungcancerbasedonbloodexosomerelatedgenesandmachinelearningmethods AT xiuliteng identifyingdiagnosticmarkersandconstructingaprognosticmodelforsmallcelllungcancerbasedonbloodexosomerelatedgenesandmachinelearningmethods AT mingweichen identifyingdiagnosticmarkersandconstructingaprognosticmodelforsmallcelllungcancerbasedonbloodexosomerelatedgenesandmachinelearningmethods |