3833 - Gene expression signature of DNA damage response to predict the prognosis of early stage lung adenocarcinoma

Background

For early-stage lung cancer, a clinically reliable prognostic biomarker is still in unmet needs. DNA damage response (DDR) system is necessary for genomic stability, whose alterations may affect prognosis via modulating immune response. We aimed to develop and validate a DDR expression signature to optimize prognostic stratification in stage I-II lung adenocarcinoma (LUAD).

Methods

A 254-patient training cohort including 4 Gene Expression Omnibus (GEO) data sets was used to develop prognostic algorithm. A 628-patient cohort from another 3 GEO data sets and a 387-patient cohort from The Cancer Genome Atlas (TCGA) data set were defined as validation cohort 1 and 2 respectively. Only resected LUAD of stage I-II with mRNA and survival data were included. Furthermore, we analyzed the associations of DDR genes signature and tumor mutation burden (TMB), copy number variations (CNVs) and tumor infiltrating lymphocytes (TILs).

Results

An 8-gene signature score was developed and stratified patients into high- and low-risk groups, including FNACA, NUDT1, CHEK1, RAD51, RAD51B, RAD54B, FAN1 and MBD4. In validation cohort 1, low-risk group showed longer disease-free survival (DFS, 102 vs 41.4 months, P = 0.008, Hazard ratio (HR) = 1.51(1.11-2.06)) and overall survival (OS, 128.8 vs 105.4 months, P = 0.003, HR = 1.53(1.15-2.02)) compared with high-risk group. For low-risk stage II patients, no significant difference was observed between patients with and without adjuvant therapy. For high-risk stage II patients, adjuvant therapy tended to improve DFS (31.0 vs 20.4months) and OS (95.0 vs 61.4months) compared with observation, but P values were not significant with limited sample size. In validation cohort 2, patients had similar gene expression pattern to that of training cohort and high-risk group showed worse survival outcomes. Analyses of TCGA cohort revealed that high-risk group had remarkably higher TMB and CNVs, and lower TILs (all P < 0.001).

Conclusions

The 8-gene DDR signature is a promising biomarker to optimize prognostic staging and personalize adjuvant therapy of early-stage LUAD. An independent multi-center study is underway to further validate the predictive value and clinical feasibility of this model.

Clinical trial identification

Legal entity responsible for the study

Jie Wang.

Funding

This work was supported by the National Natural Sciences Foundation Key Program (81630071, 81330062), National Key R&D Program of China (2016YFC0902300), National High Technology Research and Development Program 863 (SS2015AA020403), CAMS Innovation Fund for Medical Sciences (CIFMS 2016-I2M-3-008), China National Natural Sciences Foundation (81472206), Beijing Novel Program Grants (Z141107001814051).

Editorial Acknowledgement

Disclosure

All authors have declared no conflicts of interest.