Clonal Hematopoiesis Can Precede AML with Defining Cytogenetic Abnormalities - the Frequencies Vary Considerably

Background:

The 5th edition of the WHO Classification of Haematolymphoid Tumours (WHO 2022) highlights the importance of cytogenetic and molecular abnormalities in AML as disease-defining criteria. Mutations in genes frequently mutated in clonal hematopoiesis (CH) are also present in AML. However, it remains to be resolved in which order CH, other mutations, and AML-defining genetic abnormalities (DGA) occur. Here, we aimed to determine the clonal hierarchy in AML patients diagnosed according to WHO 2022 DGA.

Methods and patients:

Based on amplification free-WGS (median coverage >100x) or chromosome banding analysis followed by interphase FISH with corresponding probes, we classified 480 AML patients according to following DGA: CEBPA (n=61), NPM1 (n=173), PML::RARA (n=46), RUNX1::RUNX1T1 (n=40), CBFB::MYH11 (n=42), DEK::NUP214 (n=11), KMT2A-r (n=42) and MECOM-r (n=65). The mutational status of genes linked to myeloid neoplasms (MN) was analyzed for all samples (used variant caller: Strelka2, Pindel, HadoopCNV, and GATK4). Variant allele frequency (VAF) was corrected for copy number changes (including the presence of one X chromosome in males) and copy-neutral loss of heterozygosity (CN-LOH). For comparison, FISH VAF was calculated by dividing the cell count obtained from FISH analysis by two.

We postulated that the clone with the highest VAF within one sample emerged first. Therefore, samples were classified into three groups as follows: ‘WHO’: founder clone (FC) with DGA; ‘DTA’: FC with a mutation in DNTM3A, TET2, or ASXL1; and ‘other’: FC with a mutation in another gene with known relevance for MN. In the case of the same clone size within one sample (FISH vs. NGS VAF: +/- 10%; NGS vs. NGS VAF: +/- 5%), and no follow-up analysis was available to determine clonal hierarchy, no classification was performed (n=130/480).

Results:

In the first analysis, we observed that mutations in DTA genes were present in 152/480 (32%) samples. However, only 67/350 (19%) were assigned to the DTA origin group with varying frequencies between 0% and 35% depending on the AML-DGA (DEK::NUP214 (n=0), PML::RARA (n=1; 3%), KMT2A-r (n=3; 9%), CBFB::MYH11 (n=2; 8%), RUNX1::RUNX1T1 (n=2; 9%), CEBPA (n=8; 15%), MECOM-r (n=6; 17%), NPM1 (n=45; 35%)). For the majority of samples, the FC harbored a DGA (classification to ‘WHO’ group: PML::RARA (n=37; 95%), RUNX1::RUNX1T1 (n=20; 87%), CBFB::MYH11 (n=19; 79%), DEK::NUP214 (n=8; 89%), KMT2A-r (n=31; 89%), MECOM-r (n=5; 14%)). This observation was especially pronounced for PML::RARA-positive patients, highlighting the biological distinction of APL. In contrast, for MECOM-r AML cases, more than half of patients (69%) were classified into the group ‘other’. In this group, mutations in genes involved in splicing (32% with 6/8 in SF3B1), signaling (28%; NRAS, KRAS, and CBL), or related to transcription (20%; GATA2, ETV6, RUNX1, BCOR, and BCORL1) occurred prior the MECOM-r.

Comparing age between the WHO and DTA groups, overall, patients classified to the WHO group were significantly younger, irrespective of the underlying DGA (median age WHO vs. DTA: 53 y vs. 71 y p<0.001).

Conclusion:

Our data reveal that in most patients, the AML-defining chromosomal abnormality is indeed the early event driving the malignant process. However, for some, especially older patients, preceding CH needs to be considered as the origin of the disease. Also, in AML-MECOM-r and AML-NPM1, mutations in predominantly other genes (e. g. spliceosome) were present in founder clones. Further studies are needed to determine the clinical importance of clonal hierarchy regarding prognosis and impact on MRD monitoring of these patients to improve their therapeutic outcomes.