Identification and validation of hypoxia-derived gene signatures to predict clinical outcomes and therapeutic responses in stage I lung adenocarcinoma patients

<b>Rationale:</b> The current tumour-node-metastasis (TNM) staging system is insufficient for precise treatment decision-making and accurate survival prediction for patients with stage I lung adenocarcinoma (LUAD). Therefore, more reliable biomarkers are urgently needed to identify the high-risk subset in stage I patients to guide adjuvant therapy. <b>Methods:</b> This study retrospectively analysed the transcriptome profiles and clinical parameters of 1,400 stage I LUAD patients from 14 public datasets, including 13 microarray datasets from different platforms and 1 RNA-Seq dataset from The Cancer Genome Atlas (TCGA). A series of bioinformatic and machine learning approaches were combined to establish hypoxia-derived signatures to predict overall survival (OS) and immune checkpoint blockade (ICB) therapy response in stage I patients. In addition, enriched pathways, genomic and copy number alterations were analysed in different risk subgroups and compared to each other. <b>Results:</b> Among various hallmarks of cancer, hypoxia was identified as a dominant risk factor for overall survival in stage I LUAD patients. The hypoxia-related prognostic risk score (HPRS) exhibited more powerful capacity of survival prediction compared to traditional clinicopathological features, and the hypoxia-related immunotherapeutic response score (HIRS) outperformed conventional biomarkers for ICB therapy. An integrated decision tree and nomogram were generated to optimize risk stratification and quantify risk assessment. <b>Conclusions:</b> In summary, the proposed hypoxia-derived signatures are promising biomarkers to predict clinical outcomes and therapeutic responses in stage I LUAD patients.