Case study of a neuroendocrine tumor of uncertain origin: single-cell transcriptomics unravels potential primary location

Abstract

Purpose: Determining the primary origin of non-organ-confined neuroendocrine tumors (NETs) for accurate diagnosis and management. Neuroendocrine tumors are rare neoplasms with diverse clinical behaviors. Determining their primary origin remains challenging in cases of non-organ-confined NETs. This study explores the histogenesis of a retroperitoneal, non-functional NET localized between the duodenum and pancreatic head, utilizing advanced molecular diagnostics to elucidate its probable primary source.

Methods: Initial diagnostic methods, including imaging and histopathology, failed to resolve the tumor's origin. The tumor was subjected to single-cell RNA sequencing (scRNA-seq) and whole exome sequencing (WES). Publicly available transcriptomic datasets from pancreatic and small intestine NETs were used to develop and validate a molecular gene signature for tissue-of-origin identification.

Results: The gene signature distinguished pancreatic and small intestine NETs with high accuracy. The tumor cells presented a molecular profile consistent with a pancreatic origin, likely derived from ectopic pancreatic tissue.

Conclusions: This case demonstrates the value of integrating scRNA-seq and WES for the molecular characterization of complex NETs. Identifying the tumor's pancreatic origin informed a targeted management approach, avoiding unnecessary systemic treatment and underscoring the potential of single-cell approaches in personalized oncology.

Keywords: Duodenal NET; Ectopic pancreatic tissue; Neuroendocrine tumor; Non-organ-confined NET; Pancreatic NET; Single-cell RNA‐seq.

Fig. 1

(a-c) Computed tomography scans with intravenous contrast enhancement. Tumor size measures are illustrated in the images (in cm). Arrows indicate: 1 - tumor; 2 – duodenal lumen; 3 – biopsy needle; 4 – pancreatic tissue. (d) Intraoperative view. Arrows indicate: 1 – tumor; 2 – pancreatic tissue adjacent to the tumor; 3 – vascular pedicle of the tumor; 4 – posterior surface of the descending part of the duodenum. (e) Endosonographic data. Arrows indicate: 1 - tumor; 2 – duodenal lumen; 3 – biopsy needle; 4 – pancreatic tissue

Fig. 2

(a) Histological examination of the surgical specimen of the pancreas (haematoxylin-eosin) of patient A. (x400). (b) Immunohistochemical examination of the surgical specimen of the pancreas of patient A, chromogranin A staining (x400). (c) Immunohistochemical examination of the surgical specimen of the pancreas of patient A, synaptophysin staining (x400). (d) Immunohistochemical examination of the surgical specimen of the pancreas of patient A, ki67 staining (x400)

Fig. 3

ScRNA-seq analysis of the tumor sample. (a) Annotated cells from the tumor sample on UMAP coordinates. (b) Dot plot with cell type markers used to annotate cell clusters. (c) Tumor cells on UMAP coordinates, black - normal neuroendocrine cells without CNV, light blue - tumor cells from clone 1, blue - tumor cells from clone 2. (d) Plot with CNV patterns in the two identified clones (above), scheme of clonal evolution from normal cells to clone 1 and clone 2 (below)

Fig. 4

Comparison to published data and tumor origin identification. (a) Bar plot with cellular compositions of different tumor samples; pNET and siNET - pancreatic and small intestine NET samples from published data, respectively, tumor - NET sample from this study. (b) Brief scheme of score construction to identify tumor origin. (c) Box plots with the score, colors represent samples, box plot represent distribution of score either for tumor cells, or for microenvironment cells from different samples

Fig. 5

Results of WES data analysis. (a) Germline and somatic mutation oncoplot. (b) Somatic CNA profile