Characterizing the molecular landscape of early onset and average onset prostate cancer

5059 Background: Prostate cancer (PC) has a median onset age of 66, however, recent evidence suggests an increase in incidence of PC diagnosis in males <55 years of age. Family history and increased mutational burden has been associated with early onset PC (EOPC), nonetheless, comprehensive molecular and immune signatures that cluster in EOPC and average onset PC (AOPC) is poorly understood. Here, we characterized EOPC and AOPC, and their association with molecular and immune signature. Methods: 5,305 PC samples (<55 yrs, n=575; ≥65 yrs, n=4730) were tested by NGS (592, NextSeq; WES, NovaSeq), WTS (NovaSeq) (Caris Life Sciences, Phoenix, AZ). PC Patients with age <55 and ≥65 was classified as EOPC and AOPC, respectively. Microsatellite-instability (MSI) was tested by IHC and NGS. Tumor mutational burden (TMB) totaled somatic mutations per tumor (high≥10 mt/MB). Androgen receptor (AR) signature and Neuroendocrine Prostate Cancer (NEPC) score was calculated based on expression of a list of previously defined genes (Hieronymus et al. 2006, Beltran et al. 2016). Pathway enrichment determined by Gene Set Enrichment Analysis (Broad Inst). Immune cell fractions were calculated by deconvolution of WTS: Quantiseq. Statistical significance was determined using chi-square and Mann-Whitney U test (p <0.05) along with FDR-corrected p-values (q)<0.05. Results: EOPC had higher frequency of TMPRSS2 (35.8% vs 29.7%), ETV1 (5.2% vs 2.8%), ETV4 (3.4% vs 1.6%), and BRAF (1.9 vs 0.7%) fusions, but lower frequency of APC (4% vs 8%), CTNNB1 (2.4% vs 4.6%), RB1 (2.3% vs 4.7%) and AR (1% vs 4.9%) mutation, TMB high (2.1% vs 4.6%) and dMMR/MSI-H (2.1% vs 4%) compared to AOPC (all p<0.05). EOPC had pathway enrichment of myogenesis (NES: 1.5, q<0.01), but no difference in hedgehog signaling (NES:1.3, q=0.09) and epithelial mesenchymal transition (NES: 1.3, q=0.09) pathways compared to AOPC. EOPC had higher median NEPC score (0.359 vs 0.353, q=0.04), higher MAPK pathway activity score (MPAS) (3-fold, q=0.02) but no difference in median AR signature (q=0.39) compared to AOPC. In addition, EOPC had higher expression of PSA (1.2-fold, q<0.01) and immunomodulatory genes ( IL12A, CTLA4, FC:1.2, q<0.05) with reduced AR expression (1.3-fold, q<0.01), however there was no difference in IHC-AR (q=1) and PD-L1 (q=0.57). Analysis of inferred immune cell infiltrates showed that EOPC had increased infiltration of NK cells (4.5% vs 4.1%, q<0.01) and dendritic cells (1.9% vs 1.8% q<0.05) compared to AOPC. Conclusions: Our data suggest that EOPC is enriched in fusion events including TMPRSS2, ETV1, ETV4 and BRAF. Distinct transcriptomic features seen in EOPC included neuroendocrine differentiation, MAPK activations, immunomodulatory gene expression, and increased infiltration of NK cells and dendritic cells, suggesting inherent molecular differences and differential tumor immune microenvironment in EOPC and AOPC.