Dataset Information

A transcription factor signature predicts the survival of patients with adrenocortical carcinoma.

ABSTRACT:

Background

Adrenocortical carcinoma (ACC) is a rare endocrine cancer that manifests as abdominal masses and excessive steroid hormone levels and is associated with poor clinical outcomes. Transcription factors (TFs) deregulation is found to be involved in adrenocortical tumorigenesis and cancer progression. This study aimed to construct a TF-based prognostic signature for the prediction of survival of ACC patients.

Methods

The gene expression profile and clinical information for ACC patients were downloaded from The Cancer Genome Atlas (TCGA, training set) and Gene Expression Omnibus (GEO, validation set) datasets after obtained 1,639 human TFs from a previously published study. The univariate Cox regression analysis was applied to identify the survival-related TFs and the LASSO Cox regression was conducted to construct the TF signature based on these survival-associated TFs candidates. Then, multivariate analysis was used to reveal the independent prognostic factors. Furthermore, Gene Set Enrichment Analysis (GSEA) was performed to analyze the significance of the TFs constituting the prognostic signature.

Results

LASSO Cox regression and multivariate Cox regression identified a 13-TF prognostic signature comprised of CREB3L3, NR0B1, CENPA, FOXM1, E2F2, MYBL2, HOXC11, ZIC2, ZNF282, DNMT1, TCF3, ELK4, and KLF6. The risk score based on the TF signature could classify patients into low- and high-risk groups. Kaplan-Meier analyses showed that patients in the high-risk group had significantly shorter overall survival (OS) compared to the low-risk patients. Receiver operating characteristic (ROC) curves showed that the prognostic signature predicted the OS of ACC patients with good sensitivity and specificity both in the training set (AUC > 0.9) and the validation set (AUC > 0.7). Furthermore, the TF-risk score was an independent prognostic factor.

Conclusions

Taken together, we identified a 13-TF prognostic marker to predict OS in ACC patients.

SUBMITTER: Zhao J

PROVIDER: S-EPMC8667743 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:Hepatocellular carcinoma (HCC) is one of the most common public health challenges, worldwide. Because of molecular complexity and tumor heterogeneity, there are no effective predictive models for prognosis of HCC. This underlines the unmet need for accurate prognostic models for HCC. Analysis of GSE14520 data from gene omnibus (GEO) database identified multiple differentially expressed mRNAs (DEMs) between HCC and normal tissues. After randomly stratifying the patients into the training and testing groups, we performed univariate, lasso, and multivariable Cox regression analyses to delineate the prognostic gene signature in training set. We then used Kaplan-Meier plot, time-dependent receiver operating characteristic (ROC), multivariable Cox regression analysis of clinical information, nomogram, and decision curve analysis (DCA) to evaluate the predictive and overall survival value of a novel five-gene signature (CNIH4, SOX4, SPP1, SORBS2, and CCL19) within and across sets, separately and combined. We also validated the prognostic value of the five-gene signature using The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC), GSE54236 and International Cancer Genome Consortium (ICGC) sets. Multivariable Cox regression analysis revealed that the five-gene signature and tumor node metastasis (TNM) stage were independent prognostic factors for overall survival of HCC patients in GSE14520 and TCGA-LIHC. Combining TNM stage clinical pathological parameters and nomogram greatly improved the prognosis prediction of HCC. Further gene set enrichment analysis (GSEA) revealed enrichment of KEGG pathways related to cell cycle in the high-risk group and histidine metabolism in the low-risk group. Finally, all these five mRNAs are overexpressed between 12 pairs of HCC and adjacent normal tissues by quantitative real-time PCR validation. In brief, a five-gene prognostic signature and a nomogram were identified and constructed, respectively, and further validated for their HCC prognostic value. The five-gene risk score together with TNM stage models could aid in rationalizing customized therapies in HCC patients.