Disparities in cancer genomics by ancestry in the 100,000 Genomes Project
Nguyen, T.; Tallman, S.; Cho, Y.; Sosinsky, A.; Ambrose, J.; Thorn, S.; Mackintosh, M.; Brown, M. A.; Moutsianas, L.; Silver, M. J.; Kuchenbaecker, K.
Show abstract
PurposeMost research on genetic screening and precision oncology is based on participants of European ancestry, making it vital to evaluate the performance of these approaches in diverse populations. We analysed data from the 100,000 Genomes Project (100kGP) to assess ancestry-related differences in cancer variant prioritisation. Patients and MethodsTo assess the representativeness of the 14,775 participants with cancer from the 100kGP, we compared recruitment ratios for self-reported ethnicities to those in England. For genetic ancestry groups we analysed differences in detection rates for potential pathogenic variants (PVs) in the germline and somatic mutations in genes with treatment implications and investigated possible causes of observed disparities. ResultsRecruitment rates for Black and Asian ethnicities compared with White ethnicity in the 100kGP were consistent with rates in England, except for bladder and prostate (Black and Asian) and breast (Asian only) where Black and Asian ethnicities were recruited at higher rates than expected compared to White ethnicity. Patients with non-European genetic ancestry were more likely to carry variants classified as potential pathogenic compared to European ancestry (p=0.006). PVs were identified in 4.6% of South Asian (adjusted model: odds ratio=1.88, 95%CI=1.21-2.93) and 5.3% of African ancestry patients (odds ratio=2.24, 95%CI=1.44-3.48) compared with 2.2% in European. Fewer non-synonymous somatic mutations in actionable genes were identified in patients of non-European ancestry (p=0.004). WGS failed to identify treatment-relevant findings for 26% of patients of South Asian ancestry compared with 16% of European ancestry. ConclusionThe excess germline variants classified as PVs in patients with non-European ancestry may impede the diagnostic process. Our analysis demonstrates the need for better variant classification across diverse ancestries to ensure equitable implementation of genomics in cancer care.
Matching journals
The top 8 journals account for 50% of the predicted probability mass.