Background Africa bears 7.8% of global cancer deaths from only 5.9% of cases yet remains critically underrepresented in cancer genomics (<3% of GWAS). African patients are systematically excluded due to limited capacity and tools that ignore African epidemiology, genetic diversity, and resource realities, causing commercial panels to miss population-relevant variants. This study aims to design an optimized, Africa-specific pan-cancer gene panel balancing clinical actionability, population relevance, and economic feasibility. Methods We integrated TCGA somatic mutations (12 cancers), GLOBOCAN 2022 incidence, OncoKB, CIViC, and African cancer genomics literature to design a targeted panel. A composite framework weighted pan-cancer frequency (30%), African burden (30%), literature support (20%), and expert curation (20%) to rank 22,197 genes. Optimal panel sizes used convergence of elbow analysis, coverage thresholds, actionable gene saturation, and cost-effectiveness modeling. Sensitivity analyses used TCGA Black and AACR GENIE v13 Non-Hispanic Black subsets. Findings Three panels were developed: 30-gene (resource-limited), 70-gene (standard clinical), and 130-gene (comprehensive). The 70-gene panel covered 55/67 (82.1%) actionable FDA biomarkers with 87.1% concordance vs commercial platforms. Notably, it identified African-burden-prioritized genes ( FANCL in 70-gene; nine additional genes in 130-gene panel) absent from MSK-IMPACT and FoundationOne CDx. Sensitivity analysis showed gene ranking stability; alternate weighting showed low Jaccard similarity but retained high-priority genes. Interpretation This African-optimized panel balances actionability, population relevance, and feasibility for resource-limited settings. African-ancestry sensitivity analyses support robustness, though prospective validation in continental cohorts is essential. Implementation could advance equitable cancer care while capturing variation missed by European-centric designs. Funding No specific funding was received for this study.
Abubakar et al. (Mon,) studied this question.