A tumor microenvironment-related mRNA-ncRNA signature for prediction early relapse and chemotherapeutic sensitivity in early-stage lung adenocarcinoma.

OBJECTIVES

Postoperative early relapse of early-stage lung adenocarcinoma is implicated in poor prognosis. The purpose of our study was to develop an integrated mRNA and non-coding RNA (ncRNA) signature to identify patients at high risk of early relapse in stage I-II lung adenocarcinoma who underwent complete resection.

METHODS

Early-stage lung adenocarcinoma data from Gene Expression Omnibus database were divided into training set and testing set. Propensity score matching analysis was performed between patients in early relapse group and long-term nonrelapse group from training set. Transcriptome analysis, random survival forest and LASSO Cox regression model were used to build an early relapse-related multigene signature. The robustness of the signature was evaluated in testing set and RNA-Seq dataset from The Cancer Genome Atlas (TCGA). The chemotherapy sensitivity, tumor microenvironment and mutation landscape related to the signature were explored using bioinformatics analysis.

RESULTS

Twelve mRNAs and one ncRNA were selected. The multigene signature achieved a strong power for early relapse prediction in training set (HR 3.19, 95% CI 2.16-4.72, P < 0.001) and testing set (HR 2.91, 95% CI 1.63-5.20, P = 0.002). Decision curve analyses revealed that the signature had a good clinical usefulness. Groups divided by the signature exhibited different chemotherapy sensitivity, tumor microenvironment characteristics and mutation landscapes.

CONCLUSIONS

Our results indicated that the integrated mRNA-ncRNA signature may be an innovative biomarker to predict early relapse of early-stage lung adenocarcinoma, and may provide more effective treatment strategies.