Machine learning-random forest model was used to construct gene signature associated with cuproptosis to predict the prognosis of gastric cancer

Xiaolong Liu ^1,2, Pengxian Tao ^3,4,5, He Su ⁶

¹The First School of Clinical Medical, Lanzhou University, 222 Tianshui South Road, Lanzhou, 730000, Gansu, People's Republic of China.
²Department of Science and Education, The Third People's Hospital of Gansu Province, Lanzhou, 730000, Gansu, People's Republic of China.
³Cadre Ward of General Surgery Department, Gansu Provincial Hospital, 204 Donggang West Road, Chengguan, Lanzhou, 730000, Gansu, People's Republic of China.
⁴Key Laboratory of Molecular Diagnostics and Precision Medicine for Surgical Oncology in Gansu Province, Gansu Provincial Hospital, Gansu, 730000, People's Republic of China.
⁵NHC Key Laboratory of Diagnosis and Therapy of Gastrointestinal Tumor, Gansu Provincial Hospital, Lanzhou, 730000, Gansu, People's Republic of China.
⁶Cadre Ward of General Surgery Department, Gansu Provincial Hospital, 204 Donggang West Road, Chengguan, Lanzhou, 730000, Gansu, People's Republic of China. suhedoctor@163.com.
⁷The First School of Clinical Medical, Lanzhou University, 222 Tianshui South Road, Lanzhou, 730000, Gansu, People's Republic of China. liyul@lzu.edu.cn.
⁸Department of Anesthesiology, The First Hospital of Lanzhou University, Lanzhou, 730000, Gansu, People's Republic of China. liyul@lzu.edu.cn.

⁰The First School of Clinical Medical, Lanzhou University, 222 Tianshui South Road, Lanzhou, 730000, Gansu, People's Republic of China.

Scientific Reports +

|

February 5, 2025

View abstract on PubMed

Summary

This study developed a prognostic model using 7 cuproptosis-related genes (CRGs) to predict gastric cancer (GC) patient outcomes. High EFNA4 expression correlated with better prognosis, offering new insights for GC diagnosis and treatment.

Area Of Science

Oncology
Molecular Biology
Genetics

Background

Gastric cancer (GC) is a prevalent malignancy with poor prognosis, often linked to metastasis driven by resistance to cell death.
Cuproptosis, a novel cell death pathway, and its related genes (CRGs) have limited investigation in GC.
Understanding CRGs' role is crucial for improving GC diagnosis and treatment strategies.

Purpose Of The Study

To establish a reliable prognostic model based on cuproptosis-related genes (CRGs) for gastric cancer (GC).
To identify key CRGs and pathways associated with GC patient prognosis.
To explore the potential of EFNA4 as a prognostic biomarker in GC.

Main Methods

Utilized transcriptome and clinical data from The Cancer Genome Atlas and Gene Expression Omnibus datasets for GC patients.
Employed Single Sample Gene Set Enrichment Analysis (GSEA) and the randomized forest method to construct a prognostic model.
Validated the model using Kaplan-Meier survival curves, ROC diagrams, and nomograms; analyzed pathway enrichment with GSEA and GSVA; assessed EFNA4 expression via immunohistochemistry.

Main Results

A 7-gene prognostic model (RTKN2, INO80B, EFNA4, ELF2, MUSTN, KRTAP4, ARHGEF40) was established as an independent predictor for GC prognosis.
High-risk GC patients showed enrichment in angiogenesis and TGF-beta signaling pathways.
EFNA4 expression was significantly elevated in GC tissues, and high EFNA4 expression correlated with improved patient prognosis.

Conclusions

The CRG-based prognostic model offers a valuable tool for predicting GC patient outcomes.
The findings provide new insights into the molecular mechanisms underlying GC progression and metastasis.
EFNA4 emerges as a promising biomarker for predicting favorable prognosis in gastric cancer patients.

Keywords:

Cuproptosis Gastric cancer Machine learning Prognosis Random forest