Transcriptomic Diversity of Pediatric Acute Myeloid Leukemia Genetic Drivers Correlates With Clinical Outcome and Expression of Stemness-Related Genes.

BACKGROUND: Pediatric acute myeloid leukemia (pAML) is comprised of a diverse set of oncogenic drivers (ODs) that have been risk-stratified to inform prognosis and therapeutic decision-making. Despite proteomic, transcriptomic, genetic, and epigenetic characterization of the pAML landscape, questions still remain about why certain ODs have poorer prognoses than others. METHODS: We analyze a large pAML bulk-RNA dataset (n = 435) and organize ODs along an axis of transcriptomic diversity by calculating the Simpson Diversity Index (SDI) of individual ODs. RESULTS: When comparing patients with low diversity ODs to patients with high diversity ODs, we observe poorer overall survival (HR = 1.877, 95% CI: 1.377-2.558, p = 0.0002) among patients harboring high diversity ODs in addition to an enrichment of stemness-related genes. We observe poorer survival of patients with high diversity ODs even when comparing patients with similar transcriptomic profiles (HR = 3.443, 95% CI: 1.817-6.525, p = 0.0028). CONCLUSION: We identify a link between transcriptomic diversity, expression of stemness-related genes, and clinical outcome. Higher transcriptomic heterogeneity exhibited by high diversity ODs warrants further attention when identifying patients who can benefit from novel or high-intensity therapy.