Cancer Essential Genes Stratified Lung Adenocarcinoma Patients with Distinct Survival Outcomes and Identified a Subgroup from the Terminal Respiratory Unit Type with Different Proliferative Signatures in Multiple Cohorts.
Kuo-Hao HoTzu-Wen HuangAnn-Jeng LiuChwen-Ming ShihKu-Chung ChenPublished in: Cancers (2021)
Background: Heterogeneous features of lung adenocarcinoma (LUAD) are used to stratify patients into terminal respiratory unit (TRU), proximal-proliferative (PP), and proximal-inflammatory (PI) subtypes. A more-accurate subtype classification would be helpful for future personalized medicine. However, these stratifications are based on genes with variant expression levels without considering their tumor-promoting roles. We attempted to identify cancer essential genes for LUAD stratification and their clinical and biological differences. Methods: Essential genes in LUAD were identified using genome-scale CRIPSR screening of RNA sequencing data from Project Achilles and The Cancer Genome Atlas (TCGA). Patients were stratified using consensus clustering. Survival outcomes, genomic alterations, signaling activities, and immune profiles within clusters were investigated using other independent cohorts. Findings: Thirty-six genes were identified as essential to LUAD, and there were used for stratification. Essential gene-classified clusters exhibited distinct survival rates and proliferation signatures across six cohorts. The cluster with the worst prognosis exhibited TP53 mutations, high E2F target activities, and high tumor mutation burdens, and harbored tumors vulnerable to topoisomerase I and poly(ADP ribose) polymerase inhibitors. TRU-type patients could be divided into clinically and molecularly different subgroups based on these essential genes. Conclusions: Our study showed that essential genes to LUAD not only defined patients with different survival rates, but also refined preexisting subtypes.
Keyphrases
- genome wide
- end stage renal disease
- ejection fraction
- genome wide identification
- chronic kidney disease
- newly diagnosed
- bioinformatics analysis
- prognostic factors
- papillary thyroid
- dna methylation
- machine learning
- squamous cell carcinoma
- squamous cell
- long non coding rna
- deep learning
- poor prognosis
- high resolution
- binding protein
- single molecule
- respiratory tract
- big data