Abstract 1222P
Background
Currently, effective approaches for evaluating the prognosis of EGFRm advanced NSCLC treated with EGFR-TKI are lacking. In our study, we aimed to develop the prognosis-associated DNA methylation markers based on the differentially expressed methylation patterns between EGFRm and EGFR negative patients (pts).
Methods
Forty pts positive for EGFR Ex19del or L858R (EGFRm) and 21 pts negative for EGFR were included in this study. Blood was collected from pts prior to the treatment. Methylation levels from all samples were profiled by targeted bisulfite sequencing using a bespoke NSCLC methylation panel covering 80,672 CpG sites, spanning 1.05 mega bases of the human genome. To improve the linkage of methylation sites, we further analyzed the methylation profile as methylation blocks. Methylmean represents the average methylation values within the methylation block.
Results
We analyzed 61 blood samples from pts with advanced NSCLC. A total of 69 differentially methylated blocks (DMBs) in methylmean were common to Wilcoxon and limma methods,of which 13 upregulated and 56 downregulated by comparing EGFRm with wildtype. Univariate Cox hazard analysis for DMBs was used to identify potential prognostic blocks. By step-wise regression analysis, a four-DMB based prognostic RS model for EGFRm pts was developed, which involved 4 known cancer related genes. The RS stratified pts into high-risk (HR) and low-risk (LR) groups for PFS, and achieved AUC values of 0.818. Pts in the HR group (median RS as cutoff) showed significantly poorer PFS than those in the LR group in survival analysis (p = 0.0015; median PFS: 160 days vs 505 days). The prognostic DNA methylation model was observed to significantly associate with tumor size (p=0.037). Moreover, the co-occurring alterations TP53, RB1 and CDK4 were not associated with PFS(p>0.05). After adjusting confounding factors, the prognostic power of the four-block RS was independent of other clinical factors (p<0.001).
Conclusions
Our study constructed a DNA methylation-based RS that may potentially be used as the prognostic assay for predicting survival in EGFRm NSCLC pts after 1st line TKI treatment. Further validation is warranted.
Clinical trial identification
Editorial acknowledgement
Legal entity responsible for the study
The authors.
Funding
Has not received any funding.
Disclosure
All authors have declared no conflicts of interest.