Comprehensive Clinical Genomic Profiling Defines Age-Associated Molecular Targets in Pediatric and Adult Acute Myeloid Leukemia

Genomic profiling in AML has led to increased understanding of oncogenic mutations, refined risk stratification, and enhanced identification of alterations that can serve as targets for therapeutic intervention. Comprehensive genomic profiling (CGP) identifies a variety of alterations, including base pair substitutions, insertions, deletions, copy number alterations, and fusions. As distinct age-dependent alterations in AML are being increasingly recognized, we sought to use CGP to gain insight into age-associate variants in pediatric and adult AML. Identification of the significant age-associated biologic differences is critical to improving understanding of the age-dependent drivers of leukemogenesis. This can also advance therapeutic efforts with enhanced risk stratification, disease monitoring, and drug development.

Diagnostic and relapse specimens from a total of 934 patients, comprised of 179 pediatric (age 0-18 years) and 755 adult (age 19-87 years) specimens underwent clinical comprehensive sequencing. DNA and RNA integrated next-generation sequencing was performed in a CLIA-certified, CAP-accredited, NYS-approved laboratory for FoundationOne Heme. All captured libraries were sequenced to high depth averaging 569X for DNA (405 genes) and >3M unique pairs for RNA 9265 genes). Somatic variants identified included short variants (single nucleotides variants (SNVs) and short insertions, indels), and structural variants (fusions, amplifications, loss of whole genes, or truncations).

The total variant prevalence was 571 in 131 genes in the pediatric cohort vs. 3020 variants in 219 genes in the adult cohort. The average number of variants in the pediatric cohort was 3.2 vs. 4.0 in the adults (p<0.001). Overall, genomic alterations were less frequent in the youngest cohort of patients < 5 years of age compared to older patients (p<0.001). We subsequently analyzed the differences between the pediatric vs. adult cohorts based on the presence of short and structural variants. There were 404 somatic short variants detected in 146 samples (82%) of the pediatric cohort vs. 2645 in 718 (95%) of adult specimens. The average number of the short variants/patient was lower in the pediatric cohort at 2.3 vs. 3.5 in the adult cohort (p<0.001). Although there was overlap between several of the commonly mutated genes such as N/KRAS and FLT3, there were some significant differences between the cohorts. Many of genes involved in epigenetic modification are highly prevalent in the adult cohort and absent or at a very low prevalence in the pediatric cohort (DNMT3A: 21% vs. 2%;p<0.001), IDH1/2 21% vs. 3%;p<0.001), TET2 16% vs. 3%;p<0.01).

A total of 167 structural variants in 108 (60%) pediatric specimens were identified, compared to 375 variants in 292 (38%) adult specimens (p<0.001). Partial tandem duplications of KMT2A were more common in adults vs. pediatrics (9% vs. 2%;p<0.001), while KMT2A translocations were more common in the pediatric cohort (12% vs. 4%;p<0.001). Structural alterations involving the NUP family of genes were more common in the pediatric vs. adult cohort (15% vs. 2%; p<0.001). Alterations involving the CDKN2A/B genes were also more common in pediatric vs. adult AML (13% vs. 2%;p<0.001) (Figure 1). Genomic profiling also identified cryptic fusions across all ages, including novel fusions involving transcription factors, such as the ETS family, CREBBP, and NPM1. Within the pediatric cohort, we identified several cryptic fusions that are becoming increasingly recognized as poor prognostic features (e.g. CBF2T3A-GLIS).

AML has distinct age-associated biologic traits, with pediatric AML characterized by a lower prevalence of genomic alterations. We found that structural variants were very common in pediatric AML, occurring in 60% of patients. Importantly, a number of these variants may be important as prognostic markers, targets for therapeutic intervention, or used for disease monitoring. Genomic profiling identified significant biologic differences in pediatric vs. adult AML and advances understanding of age-dependent drivers of leukemogenesis and contribute to advancing therapeutic development that is based on strong biologic rationale. Patient specific CGP can further precision medicine efforts in AML across all age groups.

Figure 1

View largeDownload slide

Age-Associated differences in the some of the most common structural variants in the pediatric compared to adult cohort.

Figure 1

View largeDownload slide

Age-Associated differences in the some of the most common structural variants in the pediatric compared to adult cohort.

Close modal