Whole Genome DNA Methylation Profiling Identifies Neuroendocrine Tumor Origin

#2846

Introduction: Determining the origin of a neuroendocrine tumor (NET) of unknown primary can be challenging. Liver metastases can originate from any organ in the body, while pulmonary NETs can be metastases but also primary tumors. This especially holds true for Multiple Endocrine Neoplasia Type 1 patients, who often have multiple primary pancreatic and intestinal NETs. It is important to know the origin of the primary tumor since resection or ablation is crucial in case of treatment with curative intent. Furthermore, the site of origin determines prognosis, treatment options and eligibility for clinical trials.

Aim(s): DNA methylation profiles are highly tissue specific and can be used to determine tumor origin. The use of whole genome methylation profiling to determine NET origin is explored.

Materials and methods: A training cohort was built with publicly available Illumina 450k data of 67 pulmonary, pancreatic and ileal NETs. As test cohort, formalin-fixed and paraffin-embedded (FFPE) tissues of 8 primary NETs (4 pancreas, 3 lung, 1 ileum), 3 liver metastases (2 pancreas, 1 ileum) and 1 lymph node NET of unknown primary were analyzed with the Illumina 850k array. A random forest model built on the 5000 most variable CpG sites in the training cohort was used to predict tumor origin in the test cohort.

Conference: 17th Annual ENETSConcerence (2020)

Presenting Author:

Authors: Hackeng W, Geisenberger C, de Leng W, Morsink F, Vriens M,

Keywords: Methylation profiling, Neuroendocrine tumor, Machine Learning, Unknown Primary,