Background: Identification of genomic markers using NGS (next-generation sequencing)
technology would be valuable for guiding precision medicine treatments for pancreatic cancers.
Traditional somatic mutation methods require both tumor and matched non-tumor samples.
However, only tumor samples are available mostly, especially in retrospective studies. In this study,
we tried to analyze the associations between clinical features and oncogenic somatic mutations in
genome-wide tumor-only samples.
Methods: Fifty-four tumor-only samples derived from pancreatic cancer patients were used for
whole-exome sequencing. An approach involving SNP filtering of variants included in the
Catalogue of Somatic Mutations in Cancer (COSMIC) database was used to identify oncogenic
somatic mutations. The relationships between oncogenic mutations and clinical features were
analyzed and simultaneously compared with those from the TCGA database.
Results: By analyzing the mutations from tumor only samples, divergent mutation profiles were
observed in different locations (head vs. body/tail) of pancreatic tumors. The divergences between
pancreatic head and body/tail cancers were also confirmed by the TCGA data. Furthermore,
mutations of several genes were found to be significantly associated with clinical features, such as
pathological stage and the degree of tumor differentiation.
Conclusion: The results confirmed the efficiency of our approach in identifying oncogenic somatic
mutations from tumor only samples and revealed the associations between somatic mutations and
clinical features in pancreatic cancer.