Molecular and immune landscape of early-onset versus average-onset well-differentiated enteropancreatic neuroendocrine tumors.

Abstract

643 Background: The incidence of neuroendocrine tumors (NETs) has increased over the past few decades, especially among younger patients. While data from other gastrointestinal malignancies suggest early-onset may be associated with biological differences, this is yet to be investigated in NETs. We sought to characterize the molecular and immune landscape of early (EO)- versus average-onset (AO) pancreatic (pNETs) and small intestinal NETs (siNETs). Methods: Cases with metastatic pNETs and siNETs sequenced using the Tempus xT assay were included. DNA sequencing was performed to identify somatic alterations and whole transcriptome RNA-seq data were normalized to log2(TPM+1) with assay correction. Immune profiling analysis included tumor mutational burden (TMB), microsatellite instability (MSI), PD-L1 status, and immune infiltration estimated via quanTIseq. Differential expression between EO (<50 years of age at diagnosis) and AO-NETs (>50 years) was assessed using Wilcoxon rank sum tests with Benjamini–Hochberg correction, and pathway enrichment was assessed via GSEA (Gene Set Enrichment Analysis). Statistical significance was defined as p<0.05 and q<0.10 for GSEA. Results: A total of 502 patients were included, of which, 134 had EO-NETs (EO-pNETs=94, EO-siNETs=40). Compared to AO, EO-pNETs had a significantly lower prevalence of KRAS (2.1% vs 14%, p=0.001), TP53 (14% vs 25%, p=0.004), SMAD4 (3.2% vs 12%, p=0.01) and RB1 (6.4% vs 14%, p=0.04) and a higher prevalence of LRP1B (8.5% vs 1.3%, p=0.003) alterations. Among siNETs, EO cohort had a significantly higher prevalence of PAX5 (5% vs 0%, p=0.04) and HDAC2 (5% vs 0%, p=0.04) alterations. On GSEA, EO-pNETs had a significantly higher expression in the VEGF, apical junction, hedgehog signaling, and myogenesis pathways while AO-pNETs had higher expression in MYC, E2F, DNA repair and G2M checkpoint pathways. No significant differences in transcriptomic expression were noted between EO- and AO-siNETs. On immune profiling, EO-pNETs cases were enriched for M2 macrophages (p=0.004) while siNETs displayed no significant age-based differences. TMB-H, MSI-H, and PD-L1 expression did not differ by age in either group. Conclusions: This is the largest molecular analysis comparing EO and AO-enteropancreatic NETs. More age-based differences were seen in pNETs with some changes in EO cases predicting better response to immunotherapy while AO cases attained more alterations found in grade 3 NETs or NECs. Our results suggest that age at diagnosis may be an important determinant of tumor biology and clinical management options.