Identification of Four Novel Prognostic Biomarkers and Construction of Two Nomograms in Adrenocortical Carcinoma: A Multi-Omics Data Study via Bioinformatics and Machine Learning Methods

Xiaochun Yi, Yueming Wan, Weiwei Cao, Keliang Peng, Xin Li, Wangchun LiaoFrontiers in Molecular Biosciences 2022 · DOI

TL;DR

Researchers have identified four key genes (ASPM, BIRC5, CCNB2, and CDK1) that can accurately predict the survival and prognosis of patients with adrenocortical carcinoma (ACC), a rare and aggressive cancer. They've also developed two nomograms, which are visual tools, to help clinicians quickly and accurately assess a patient's overall survival and disease-free survival.

What Problem Does This Paper Address?

Adrenocortical carcinoma (ACC) is a rare cancer with a generally poor prognosis. Early and accurate prediction of patient survival is crucial for effective treatment planning. However, current methods for predicting ACC prognosis are not always precise enough, highlighting the need for novel biomarkers and more sophisticated predictive tools. This study aims to address this by leveraging multi-omics data and machine learning to identify reliable prognostic markers and develop user-friendly predictive models for ACC.

Methods

The researchers employed a comprehensive bioinformatics approach, integrating several analytical techniques:

  • Differential Gene Expression (DEG) Analysis: To identify genes that are expressed at significantly different levels between ACC tumors and normal tissues.
  • Weighted Gene Co-expression Network Analysis (WGCNA): To identify modules of highly correlated genes that might be functionally related and associated with clinical traits like survival.
  • Protein-Protein Interaction (PPI) Network Construction: To understand how identified genes interact with each other, helping to pinpoint key "hub" genes.
  • Survival Analysis: To assess the relationship between gene expression levels and patient outcomes (overall survival and disease-free survival).
  • Machine Learning Classifiers (LDA, KNN, SVM): To rigorously screen candidate genes for their accuracy in predicting survival.
  • Receiver Operating Characteristic (ROC) and Decision Curve Analysis (DCA): To evaluate the diagnostic and prognostic performance of the identified biomarkers.
  • Cox Proportional Hazards Regression and Nomogram Construction: To identify independent prognostic factors and build visual predictive models for clinical use.

Datasets

The study utilized several publicly available datasets from the Gene Expression Omnibus (GEO) database and The Cancer Genome Atlas (TCGA):

  • GEO Datasets: GSE75415, GSE12368 (for DEG analysis), GSE76021, GSE19750, GSE10927, GSE76019 (for various analyses including WGCNA, survival, and validation).
  • TCGA-ACC Data: Used for RNA sequencing data, clinical information, and validation of findings.

Key Insights and Ideas

  • Cell Cycle Genes as Key Players: The identified prognostic biomarkers are heavily involved in the cell cycle. This suggests that dysregulation of cell cycle progression is a fundamental driver of ACC development and progression.
  • Beyond Gene Expression: The study highlights that not only gene expression but also genetic alterations like mutations and copy number variations (CNVs) in these biomarkers are significantly associated with patient survival. This provides a more holistic view of their impact.
  • Immune Microenvironment Connection: The expression levels of these biomarkers are linked to the tumor's immune microenvironment. This opens avenues for exploring immunotherapies or understanding how immune cells interact with ACC progression.
  • Clinical Utility of Nomograms: The development of user-friendly nomograms offers a practical tool for clinicians, enabling quick and visual assessment of patient prognosis, which can aid in treatment decisions.

Key Results

  • Identification of Nine Hub Genes: Through integrated analysis, nine genes were identified as being significantly correlated with ACC prognosis.
  • Four Novel Prognostic Biomarkers (MPBs): From the nine hub genes, four (ASPM, BIRC5, CCNB2, and CDK1) were selected as meaningful prognostic biomarkers due to their high accuracy in survival prediction.
  • High Predictive Accuracy: These four MPBs demonstrated high accuracy in predicting overall survival (OS) and disease-free survival (DFS) across multiple independent datasets. For example, ASPM showed an average accuracy of 0.8228 in machine learning classification.
  • Association with Genetic Alterations: Mutations and CNVs in these MPBs were found to be associated with ACC patient survival. For instance, ASPM shallow deletions were linked to better OS.
  • Two Predictive Nomograms: An OS-nomogram and a DFS-nomogram were successfully constructed, showing good predictive performance (e.g., OS-nomogram C-index of 0.875 in TCGA-ACC data).
  • Potential Drug Targets: The study identified five small molecule drugs (chlorpromazine, trifluoperazine, alpha-estradiol, 15-delta prostaglandin J2, and vorinostat) that might be useful for treating ACC, with some of these drugs showing enrichment in cell cycle pathways.

Limitations and Future Work

  • Lack of In Vitro/In Vivo Validation: The authors acknowledge that the study relied on computational analyses and lacked experimental validation using cell lines or animal models. They propose conducting histology or animal experiments in the future.
  • Small Dataset Sizes and Clinical Trial Gaps: While multiple independent datasets were used, they were noted as being of small size. The authors suggest the need for verification using larger patient cohorts and clinical trials.
  • Further In-Depth Validation: Despite promising results, the authors emphasize that the identified biomarkers and predictive models require more in-depth validation to solidify their clinical utility.
  • Exploring Tumor Progression: While some genes showed associations with tumor progression in certain datasets, this correlation was not consistent across all datasets, indicating a need for further investigation into the relationship between these genes and tumor progression.

Summary updated

Abstract

<b>Background:</b> Adrenocortical carcinoma (ACC) is an orphan tumor which has poor prognoses. Therefore, it is of urgent need for us to find candidate prognostic biomarkers and provide clinicians with an accurate method for survival prediction of ACC via bioinformatics and machine learning methods. <b>Methods:</b> Eight different methods including differentially expressed gene (DEG) analysis, weighted correlation network analysis (WGCNA), protein-protein interaction (PPI) network construction, survival analysis, expression level comparison, receiver operating characteristic (ROC) analysis, and decision curve analysis (DCA) were used to identify potential prognostic biomarkers for ACC via seven independent datasets. Linear discriminant analysis (LDA), K-nearest neighbor (KNN), support vector machine (SVM), and time-dependent ROC were performed to further identify meaningful prognostic biomarkers (MPBs). Cox regression analyses were performed to screen factors for nomogram construction. <b>Results:</b> We identified nine hub genes correlated to prognosis of patients with ACC. Furthermore, four MPBs (ASPM, BIRC5, CCNB2, and CDK1) with high accuracy of survival prediction were screened out, which were enriched in the cell cycle. We also found that mutations and copy number variants of these MPBs were associated with overall survival (OS) of ACC patients. Moreover, MPB expressions were associated with immune infiltration level. Two nomograms [OS-nomogram and disease-free survival (DFS)-nomogram] were established, which could provide clinicians with an accurate, quick, and visualized method for survival prediction. <b>Conclusion:</b> Four novel MPBs were identified and two nomograms were constructed, which might constitute a breakthrough in treatment and prognosis prediction of patients with ACC.
BibTeX
@inproceedings{yi2022identification,
  author = {Xiaochun Yi and Yueming Wan and Weiwei Cao and Keliang Peng and Xin Li and Wangchun Liao},
  title = {{Identification of Four Novel Prognostic Biomarkers and Construction of Two Nomograms in Adrenocortical Carcinoma: A Multi-Omics Data Study via Bioinformatics and Machine Learning Methods}},
  booktitle = {Frontiers in Molecular Biosciences},
  year = {2022},
  doi = {10.3389/fmolb.2022.878073},
}
Figure