In:
Journal of Clinical Oncology, American Society of Clinical Oncology (ASCO), Vol. 39, No. 15_suppl ( 2021-05-20), p. e16216-e16216
Abstract:
e16216 Background: Pancreatic cancer is an extremely malignant tumor that is associated with low survival rates. Currently, TNM staging, serum CA19–9, CA125 and CEA are used to assess the risk level and estimate prognosis. However, it is still a challenge to predict survival of pancreatic cancer patients (pts) with due to the impact of wide variability of outcomes and genetic heterogeneity. In our study, we aimed to develop a model based on multiple prognostic-related methylation markers and clinical parameters to predict the overall survival (OS) of pancreatic cancer pts. Methods: A total of 50 pts with early-stage resectable pancreatic cancer were included in this study. Preoperative blood, tumor and tumor-distant normal tissue samples were obtained from the pts. Methylation levels from all samples were profiled using targeted bisulfite sequencing using bespoke pancreatic cancer 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. Results: A total of 1162 tumor-specific methylation blocks, including 737 hypermethylated and 425 hypomethylated blocks were found to be differentially methylated in tumor tissues as compared to tumor-distant normal tissues (P 〈 0.05). Genes in hypermethylated blocks were significantly enriched in neuroactive ligand−receptor interaction and ca+ signaling pathways, whereas those in hypomethylated blocks were focused in E. coli infection and leukocyte transendothelial migration pathways. All these involved pathways are pancreatic cancer related. Moreover, 7 differentially methylated blocks were identified to significantly associate with OS, including 5 hyper- and 2 hypomethylated blocks. Therefore, we constructed a model based on prognostic-related blocks to predict OS of pancreatic cancer pts. A risk score was derived for each patient based on the model. Pts in the high-risk (HR) group (median risk score as cutoff) showed significantly poorer OS than those in the low-risk (LR) group in survival analysis (p = 0.0016, HR = 0.30). When clinical parameters were also considered, the risk score was found to be the only independent prognostic parameter (p 〈 0.001) by Cox regression analysis. The prognostic effect of the risk score remained significant in patient groups separated by CA125 levels (p 〈 0.01). In pre-operative blood samples, risk score was also significantly associated with OS (p = 0.031). Pts in the LR group showed significantly longer OS than those in the HR group. Conclusions: Our data revealed distinct methylation patterns between tumor tissues and tumor-distant normal tissues, suggesting tumor-specific methylation patterns could potentially be developed as diagnostic biomarkers. Importantly, our study also identified 7 blocks-based risk score that could potentially be used as prognostic markers for pancreatic cancer.
Type of Medium:
Online Resource
ISSN:
0732-183X
,
1527-7755
DOI:
10.1200/JCO.2021.39.15_suppl.e16216
Language:
English
Publisher:
American Society of Clinical Oncology (ASCO)
Publication Date:
2021
detail.hit.zdb_id:
2005181-5
Permalink
