Acute myeloid leukemia (AML) is a cancer of the blood in which immature myeloid cells multiply uncontrollably in the bone marrow. When AML is diagnosed, one of the first steps is examining a patient's chromosomes under a microscope - a process called cytogenetic analysis. The pattern of chromosomal changes found in the leukemia cells is one of the most powerful predictors of how the disease will behave and how well a patient will respond to treatment.
The chromosomal changes in AML fall into recognizable patterns. Cytogenetically normal AML (CN-AML) has no detectable chromosome changes. Core-binding factor AML (CBF-AML) has specific chromosome rearrangements involving blood cell development genes. Complex karyotype AML (CK-AML) has three or more chromosome abnormalities and tends to be particularly aggressive. Each of these groups carries a different prognosis.
Beyond chromosomes, AML cells also carry gene mutations - smaller changes in the DNA sequence of specific genes that can activate cancer-promoting pathways, block normal cell maturation, or disable tumor suppressor mechanisms. Key mutations include those in FLT3, NPM1, DNMT3A, IDH1, IDH2, and many others. In recent years, it has become clear that chromosome abnormalities and gene mutations are not independent - specific mutations tend to cluster in specific cytogenetic groups.
Despite this knowledge, there had been no large study systematically mapping which gene mutations occur within each of the major chromosomal subgroups. The goal of this study was to fill that gap by creating a comprehensive oncoprint - a visual and statistical map of how 80 cancer-related gene mutations distribute across 34 cytogenetic subgroups in 1,603 AML patients, one of the largest such analyses ever performed.
The study analyzed 1,603 adult patients with newly diagnosed de novo AML - meaning AML that arose without a prior blood disorder or prior cancer treatment. Patients were enrolled in Cancer and Leukemia Group B (CALGB) and Alliance clinical trials over many years, providing a uniformly treated and well-characterized cohort. Of these, 1,080 were younger than 60 years of age and 523 were 60 or older.
Each patient's leukemia cells underwent cytogenetic analysis by standard methods, and the results were confirmed by central expert review. Patients were classified into 34 specific cytogenetic subgroups, which were then consolidated into five major categories: cytogenetically normal (CN-AML), complex karyotype (CK-AML), core-binding factor (CBF-AML), balanced rearrangements other than CBF, and unbalanced chromosomal abnormalities in non-complex karyotypes.
Genomic DNA from each patient's leukemia cells was analyzed using targeted amplicon sequencing on the MiSeq Illumina platform - a next-generation sequencing approach that reads specific pre-selected regions of the genome with high accuracy. This method covered 79 protein-coding genes known to be relevant in cancer and leukemia. A separate Sanger sequencing test was used for the CEBPA gene, bringing the total to 80 genes analyzed per patient.
The 80 genes were organized into nine functional groups based on their biological roles: chromatin remodeling, cohesin complex, kinases, methylation-related, NPM1, RAS pathway, spliceosome, transcription factors, and tumor suppressors. This grouping allowed statistical comparisons of entire biological pathways rather than individual genes, increasing the power to detect meaningful patterns.
Across all 1,603 patients, the study detected a total of 4,390 gene mutations, with a median of three mutations per patient (ranging from 0 to 9). The most commonly affected functional groups were methylation-related genes (DNMT3A, IDH1, IDH2, or TET2 - mutated in 45% of all patients), followed by kinases (36%), and NPM1 (31%). The least frequently affected were the cohesin complex genes (12%) and tumor suppressors (14%).
An important finding concerned the timing of mutations during cancer evolution. By comparing variant allele fractions - a measure of what proportion of cells carry a mutation - the researchers inferred which mutations tended to arise early versus late in the development of leukemia. Tumor suppressor gene mutations, cohesin complex mutations, and spliceosome mutations were most often the first mutations to occur. Kinase mutations and RAS pathway mutations tended to appear later, suggesting they represent secondary events that accelerate disease after earlier mutations have already established an abnormal clone.
The mutation patterns differed strikingly between cytogenetic groups. CN-AML had the broadest mutation spectrum, with methylation-related genes (61% of CN-AML patients), NPM1 (57%), and kinases (46%) being most common. CBF-AML had very few additional mutations - a median of just one - with NRAS (23%) and KIT (20%) being most frequent, and almost no NPM1 or methylation-related mutations. CK-AML was dominated by TP53 mutations (38%), reflecting its fundamentally different biology.
The data confirmed and extended several known associations while also revealing novel ones. For example, patients with the rare chromosomal rearrangement t(6;9)(p23;q34) had a remarkably high rate of FLT3-ITD mutations (71% of these patients), with very few other co-occurring mutations - suggesting a particularly clean two-hit mechanism. Patients with t(9;22)/BCR-ABL frequently had RUNX1 and ZRSR2 mutations, a finding with implications for therapy since BCR-ABL is targetable with tyrosine kinase inhibitors.
Within complex karyotype AML (CK-AML), the study revealed an important biological division between 'typical' CK-AML (with chromosome losses from 5q, 7q, and 17p) and 'atypical' CK-AML (without these canonical losses). TP53 mutations were present in 52% of typical CK-AML patients but only 5% of atypical CK-AML patients. Conversely, atypical CK-AML showed a broader mutation spectrum with nine different mutations each occurring in at least 10% of patients, including FLT3-TKD, DNMT3A, IDH2, NPM1, NRAS, PHF6, RUNX1, TET2, and ZRSR2. This distinction matters because typical and atypical CK-AML may warrant different treatment approaches.
Among patients with CBF-AML, an interesting difference emerged between the two major subtypes - t(8;21) and inv(16). Cohesin complex mutations were found in 15% of t(8;21) patients but zero inv(16) patients. Similarly, CCND2 mutations were found in 12% of t(8;21) patients, a recently discovered feature of this subtype. These differences suggest that t(8;21) and inv(16), despite both being CBF-AML, have distinct secondary mutation profiles and may benefit from subtype-specific targeted therapy approaches.
Patients with unbalanced chromosomal abnormalities - gains or losses of chromosomes in non-complex karyotypes - showed high rates of spliceosome mutations (32%), particularly in patients with chromosomal gains (trisomies). Among patients with sole trisomy 13, half carried SRSF2 spliceosome mutations. This pattern suggests that chromosomal gains and spliceosome mutations may cooperate during leukemia development, and raises the possibility that spliceosome inhibitor drugs - which are in clinical development - might be particularly relevant for this cytogenetic subgroup.
Patients with monosomal karyotype - defined by two or more complete chromosome losses - had patterns resembling typical complex karyotype, with TP53 mutations in 39% and only NRAS and TET2 also reaching 10%. Monosomal karyotype carries an extremely poor prognosis, and the dominance of TP53 mutations in this group suggests that targeting TP53 pathways may be relevant, though this remains challenging as TP53 itself is difficult to drug directly.
The study analyzed 1,080 patients younger than 60 years and 523 patients aged 60 or older. Even the distribution of cytogenetic subgroups differed markedly by age. CBF-AML was more than five times more common in younger than older patients (15.1% vs. 2.7%), while CK-AML and unbalanced chromosomal abnormalities were much more frequent in older patients. This difference in cytogenetic distribution partly explains why older AML patients tend to have worse outcomes - they more often have high-risk chromosomal subtypes.
Beyond cytogenetics, the mutational profiles also differed with age. Younger patients had a median of two mutations per patient; older patients had a median of three. Kinase mutations (FLT3, KIT) were more common in younger patients (39% vs. 29%), while methylation-related mutations (61% vs. 36%), spliceosome mutations (37% vs. 12%), transcription factor mutations (29% vs. 18%), and chromatin remodeling mutations (23% vs. 14%) were all significantly more common in older patients.
The higher rates of spliceosome and methylation-related mutations in older AML patients are likely related to clonal hematopoiesis - a well-documented phenomenon in which blood stem cells accumulate mutations with age, and some of these age-related mutations predispose to leukemia. Spliceosome mutations (particularly SRSF2) and DNMT3A, TET2, ASXL1 mutations are among the most common clonal hematopoiesis mutations, and they appear to serve as early initiating events that then acquire further mutations leading to overt leukemia in older individuals.
These age-related differences have practical implications. The higher prevalence of FLT3-ITD mutations in younger patients means that targeted FLT3 inhibitors (such as midostaurin, which was FDA-approved based on clinical trial data in this period) may be particularly relevant in younger patients. The biology of AML in older patients appears fundamentally different - more driven by acquired age-related changes in the epigenome and splicing machinery than by kinase mutations, which may explain why standard chemotherapy works better in younger patients.
The oncoprint data has direct practical implications for how patients should be evaluated and treated. The findings support a more targeted mutation testing strategy: rather than blindly testing all 80 genes in every patient, clinicians could use the cytogenetic findings to focus on the mutations most likely to be present. For example, knowing that a patient has CBF-AML suggests prioritizing testing for KIT mutations, which are found in 20% of CBF-AML patients and have prognostic relevance for whether to intensify treatment.
The finding that certain cytogenetic subgroups have very specific mutation co-occurrences enables better identification of therapeutic targets. Patients with sole trisomies and spliceosome mutations represent a group that might benefit from spliceosome inhibitor therapies in development. Patients with t(9;22)/BCR-ABL and RUNX1 mutations represent a newly characterized group where combining BCR-ABL-targeted tyrosine kinase inhibitors with other targeted agents might be considered.
The data also refines our understanding of prognosis within cytogenetic subgroups. Two patients with the same chromosomal abnormality but different co-occurring mutations may have very different disease outcomes. For example, within the balanced rearrangements group, those with TP53 mutations within a complex karyotype context have dramatically worse prognosis than those without - suggesting these patients should be reclassified into the high-risk group for treatment planning purposes.
The oncoprint concept itself - creating a visual summary of genomic alterations across a large patient cohort - is a contribution to how oncology presents and communicates complex genomic data. The resulting maps are incorporated into genomic data portals like cBioPortal, making them available to researchers worldwide as a reference for understanding AML biology and identifying opportunities for targeted therapy development.
This study provides the most comprehensive mapping to date of the co-occurrence patterns between cytogenetic abnormalities and gene mutations in AML. The scale of the cohort - 1,603 patients with systematic analysis of 80 genes and 34 cytogenetic subgroups - makes this a reference dataset for the field. The key message is that AML is not a single disease but a collection of biologically distinct subtypes defined by the combination of chromosomal and mutational features.
The findings reveal important differences between subgroups that were previously thought to be similar - such as typical versus atypical complex karyotype AML, or t(8;21) versus inv(16) CBF-AML. These distinctions may translate into different treatment approaches and different outcomes within groups that are currently treated the same way. Prospective clinical trials in defined subgroups, guided by the mutation data presented here, are a natural next step.
The study also demonstrates how age fundamentally changes AML biology. The different distribution of cytogenetic groups and the different mutation profiles between younger and older patients suggest that AML in older patients is a biologically distinct disease process - one driven more by age-related epigenetic and splicing pathway alterations than by the kinase and transcription factor mutations that dominate in younger patients. This difference likely contributes to the inferior outcomes seen in older AML patients.
As comprehensive genomic profiling becomes standard in AML diagnosis - facilitated by next-generation sequencing panels now available in clinical laboratories - the oncoprint provides a reference framework for interpreting what any given combination of cytogenetic and molecular findings means. This has direct value for treatment decision-making at the individual patient level.