In prostate cancer, Black men face higher incidence rates, more advanced diagnoses and worse outcomes than White men. These disparities partially stem from environmental, socioeconomic, and healthcare access factors. Recent data highlight racial differences in prostate cancer molecular profiles, with tumors from Black men showing increased SPOP and decreased PIK3CA mutations and amplified androgen receptor signaling and inflammatory pathway activations. Thus, somatic mutation patterns may also contribute to these disparities. We sought to identify associations of mutational profiles in prostate adenocarcinoma (PRAD) with genetic ancestry instead of race and ethnicity categories, analyzing a de-identified cohort of 5,959 patients who underwent the Tempus xT 648-gene NGS tumor profiling test. We used 654 ancestry-informative markers to estimate genetic ancestry by calculating similarity to reference populations for five regions: Africa (AFR), Americas (AMR), East Asia (EAS), Europe (EUR), and South Asia (SAS). Associations between genetic ancestry proportions and somatic variants were assessed for 67 PRAD-related genes using logistic regression models. Analyses included copy number alterations (CNAs), oncogenic gene fusions, small splice and non-synonymous mutations (NS), OncoKB L1/2 & R1, and small missense driver somatic mutations predicted with boostDM. Likelihood ratio tests were employed to compare models with and without ancestry terms. The Benjamini-Hochberg method was utilized to adjust p-values and control the FDR at 5%. We confirmed reported associations in PRAD between AFR ancestry and SPOP NS mutations (OR=1.03 per doubling of AFR ancestry proportion, p=0.03) and decreased AFR with PIK3CA OncoKB and driver mutations (OR=0.94, p=0.02), PTEN CNAs (OR=0.95, p=0.0001), and TMPRSS2-ERG fusions (OR=0.95, p<0.0001). However, we also observed negative associations between AFR and AR CNAs (OR=0.95, p=0.004), EAS and PTEN CNAs (OR=0.96, p=0.01), and EAS and TMPRSS2-ERG fusions (OR=0.97, p=0.02). Additionally, a negative association was found between EAS and TP53 driver NS mutations (OR=0.96, p=0.04), while positive associations with driver mutations in FOXA1 (OR=1.06, p=0.03) and NS mutations in the PI3K/AKT/mTOR pathway genes (OR=1.04, p=0.01) were noted. Lastly, we identified novel positive associations between AMR and driver mutations in PIK3C3A (OR=1.12, p=0.0003) as well as small NS mutations or CNAs in the PI3K/AKT/mTOR pathway genes (OR=1.03, p=0.04). By analyzing a large, diverse real-world cohort and leveraging NGS-inferred genetic ancestry, our study confirms known associations between somatic alterations in PRAD cancer genes and race and ethnicity, while unveiling novel associations in understudied populations of potential significance for understanding disparities in disease outcomes.

VIEW THE PUBLICATION

VIEW THE POSTER