Targeted deep sequencing reveals the genetic heterogeneity in well-differentiated pancreatic neuroendocrine tumors with liver metastasis

BACKGROUND: Pancreatic neuroendocrine tumor is a rare and heterogeneous entity, and approximately half of the patients harbored liver metastasis when initially diagnosed, whose prognosis is dismal. High-throughput sequencing has largely uncovered the genomic features of pancreatic neuroendocrine tumor, but the genetic alterations in the metastatic cases remain relatively unclear, which we aimed to study. METHODS: Pathologically confirmed well-differentiated pancreatic neuroendocrine tumor samples resected in our hospital from 2000 to 2019 were collected. We performed deep sequencing on the exome of 341 tumor-related genes, and compared the differences of genetic alterations between the metastatic and the non-metastatic cases, as well as between the primary and the paired liver metastatic tumors. RESULTS: Sequencing data of 79 samples from 29 pancreatic neuroendocrine tumor patients were included into analysis. A total of 2,471 somatic variants were identified, 75.5% of which were considered as low-abundance. NOTCH1 was the most frequently mutated gene, altered in 26 (53.1%) pancreatic neuroendocrine tumor samples from 18 (62.1%) patients. Compared with the non-metastatic pancreatic neuroendocrine tumors, the metastatic cases were discovered with more single nucleotide variants and copy number variations, indicating the increased genomic instability. In addition, among the paired metastatic cases, the primary and the metastatic lesions shared limited mutated genes. CONCLUSIONS: Through the targeted deep sequencing, we identified the intratumor, intraindividual, and interindividual heterogeneity in the pancreatic neuroendocrine tumor patients, particularly in the metastatic cases, bringing potential challenges for the current biopsy strategies in guiding clinical treatments.