Clinical application of molecular genetics in acute myeloid leukemia

INTRODUCTION

Acute myeloid leukemia (AML) is one of the most common adult leukemias in adults and accounts for 1% of all cancer cases.^[1] The disease is characterized by a proliferation of myeloid blasts but there is vast heterogeneity in the molecular signature of the malignant clone. For several decades, diagnosis of AML was dependent on morphology and immunophenotyping, risk stratification was largely determined by conventional karyotyping and treatment comprised of induction with “3+7” followed by consolidation with high dose cytarabine or allogeneic transplant. Molecular genetics has provided us with a better understanding of pathophysiology, refined risk stratification, and led to the development of novel targeted treatment modalities. The latest recommendations on diagnosis and treatment of AML reflect this changing paradigm and have been covered in this review.

EMERGING ROLE OF MOLECULAR MARKERS IN THE DIAGNOSIS OF AML

AML is a myeloid disorder characterized by genetic heterogeneity. Over the past few years, access to standardized molecular techniques has led to a better understanding of this disease. Next generation sequencing (NGS) has identified several important genetic mutations, grouping different patients of AML based on their mutated gene showing similar clinical profile, morphology, disease course, targeted therapy, and/or prognosis. Major scientific groups in hematology (World Health Organisation, International Consensus Classification and European LeukemiaNet) have proposed a new classification of AML mostly based on underlying genetic mutations with certain differences identified [Table 1].^[2,3]

The cutoff for blast percentage for the diagnosis of AML has always been a point of debate. As a rule of thumb, a blast count of ≥20% was established as an acceptable cutoff except for AML in the presence of defined or recurrent genetic abnormalities. No lower cutoff for blast count has been given by the WHO for this category; however, according to ELN and ICC, a minimum blast count of 10% is required. AML with BCR: ABL1 fusion and AML with CEBPA mutation still require at least 20% of blast count for diagnosis.^[2,4]

AML with recurrent/defined genetic abnormalities has now been expanded with new genetic mutations such as BCR: ABL1 fusion, KMT2A rearrangement, DEK: NUP214 fusion, RBM15:MRTFA fusion, and MECOM rearrangement impacting the patient’s management and therapeutic decision. The WHO has replaced AML with t(9;11) fusions with KMT2A rearrangements because of several partner genes identified for KMT2A; however, ELN does not recognize these rearrangements and retains the fusions.^[5]

TP53 mutations in AML are an entity representing a part of a spectrum of myeloid neoplasms with mutated TP53. The pathogenic TP53 mutation should have a variant allele frequency (VAF) of at least 10% with or without the loss of the wild allele.^[5] The presence of complex cytogenetics with 17p deletion in the absence of TP53 mutation does not qualify for this entity. Clinically, these myeloid neoplasms have a very poor prognosis and are frequently associated with complex chromosomal abnormalities.^[6,7] Diagnosis of AML with TP53 mutations takes precedence over AML with myelodysplasia-related gene mutations.^[5]

A previously known entity of AML-MRC has been removed based on the recognition of more objective criteria of myelodysplasia-related gene mutations and myelodysplasia-related cytogenetic abnormalities. Detection of genetic mutations such as ASXL1, BCOR, EZH2, RUNX1, SF3B1, SRSF2, STAG2, U2AF1, and/or ZRSR2 in AML has led to the identification of a new AML category – “AML with myelodysplasia-related gene mutations” with poor prognosis. These genes have been commonly associated with AML arising from MDS or MDS/MPN.^[8,9]

The list of mutations causing germline predisposition in AML is also expanding due to better molecular techniques. It is important to identify these mutations to identify other affected organs and also have an important therapeutic implication in matched sibling hematopoietic stem cell transplant.

Hierarchical classification of AML has been given by ICC where AML with recurrent genetic abnormalities takes precedence over all other categories. Among the remaining categories, AML with mutated TP53 supersedes AML with myelodysplasia-related gene mutations which supersedes AML with myelodysplasia-related cytogenetic abnormalities.

RISK STRATIFICATION

The ELN 2022 guidelines have made several changes in the risk stratification of AML at diagnosis.^[5] A major change is that measuring the allelic ratio of FLT3-ITD mutations is no longer necessary. This change was made due to the lack of standardization of the assays for measuring the FLT3 allelic ratio. Furthermore, all FLT3-ITD-mutated AML now come under the intermediate risk category regardless of the coexistence of NPM1 mutations. This is because targeted treatment with FLT3 inhibitors has led to improved outcomes in all FLT3-mutated cases in the RATIFY trial.^[10]

AML with myelodysplasia-related gene mutations is observed to carry a worse prognosis. In the 2017 ELN guidelines, the single gene mutations leading to inferior outcomes were TP53, RUNX1, and ASXL1. This list has now been expanded to include SRSF2, STAG2, BCOR, U2AF1, EZH2, SF3B1, and ZRSR2. The patients in ELN 2017 intermediate risk group were observed to carry a worse prognosis if carrying any of these myelodysplasia-related mutations and hence are now included in the adverse risk category.^[11]

NPM1-mutated AML without FLT3-ITD carries a favorable outcome and those with FLT3-ITD fall under the intermediate risk group. A change in the recent risk stratification is that the presence of adverse risk cytogenetics overrides the significance of NPM1 mutation as these group of patients have been observed to have lower remission rates, inferior 5-year OS, and EFS with a higher 5-year cumulative incidence of relapse.^[12]

With regards to CEBPA mutations, the previous understanding was that biallelic mutations carried a favorable prognosis and this was irrespective of the type of mutation seen. The understanding has now evolved as now it is observed that it is the in-frame mutations in the basic leucine zipper (bZIP) region of CEBPA that actually confer a favorable prognosis, regardless of whether they are monoallelic or biallelic.^[13] CEBPA mutations in the transactivating domain or non-in-frame bZIP mutations do not confer this advantage.^[14]

Cytogenetic risk groups in the adverse risk group have a few additions. t(3q26.2;v) involving rearrangements in the MECOM gene and t(8;16)(p11;p13) associated with KAT6A: CREBBP gene fusion are added to this group. Complex karyotype in AML is defined as having ≥3 unrelated chromosome abnormalities in the absence of other class-defining recurring genetic abnormalities. Patients with hyperdiploid karyotype in the absence of other adverse cytogenetics have been observed not to carry a worse outcome and these are now excluded from the definition of complex karyotype.^[15] [Table 2] provides a side-by-side comparison of the 2017 and 2022 ELN risk stratification recommendations.^[5,16]

MEASURABLE RESIDUAL DISEASE (MRD)

Assessment of MRD after initial treatment is one of the most powerful tools for the determination of long-term prognosis and planning of further treatment. The 2021 consensus document on MRD measurement by the ELN sought to standardize the methodology and cutoffs for MRD in AML.^[17] The measurement of MRD in AML may be done by flow cytometry or polymerase chain reaction (PCR) and the baseline immunophenotype and molecular profile of the disease determines the method to be employed.

Molecular MRD by PCR is the preferred method when there is a presence of a quantifiable AML-related mutation. This is employed for mutated NPM1; CBFB: MYH11, RUNX1:RUNX1T1, KMT2A: MLLT3, DEK: NUP214, BCR: ABL1 gene fusions; and WT1. The most robust data are available for NPM1 mutated and core-binding factor AML. The assessment is to be done from blood after two cycles of induction, from bone marrow at the end of treatment, and subsequently for follow-up-every 3 months from bone marrow or every 4–6 weeks from blood.^[17] PCR-based MRD is considered negative if undetectable with Ct ≥40 in at least two of three replicates with at least ≥10,000 (preferably ≥30,000) copies of the housekeeping gene measured. Specifically for NPM1-mutated AML, there may be MRD persistence at low levels (<2%) without portending an unfavorable prognosis and this is labeled MRD-LL.

Flow cytometry-based MRD assessment is recommended for all cases without the abovementioned molecular. LAIP (leukemia-associated immunophenotype) and Dfn (different from normal immunophenotype) are the markers employed to identify the malignant clone. MRD by flow cytometry is considered negative if <0.1%.

NGS-based methods for MRD quantification are not recommended to be used as a standalone method at present. Their current utility is to be used as an adjunct to the abovementioned methods and for use as a research tool. Mutations associated with clonal hematopoiesis of indeterminate potential should not be used as MRD markers – these include DNMT3A, TET2, and ASXL1 also to be excluded from the analysis are germline mutations that may be present with a VAF>50%. Error corrected panel-based sequencing is to be employed.^[18]

INDIVIDUALIZED TREATMENTS FOR SPECIFIC MOLECULAR TARGETS

A complete discussion on the treatment of AML is beyond the scope of this review. We shall briefly discuss the currently approved drugs targeting disease-specific mutations in AML.

CONCLUSION

The management of AML is slowly pivoting in the direction defined by molecular genetics. With the increasing availability of PCR-based assays and next generation sequencing techniques, we can refine the way we diagnose, prognosticate, and treat this disease. A lot of new insights have been gained regarding the pathophysiology of AML and many more changes in diagnostic and prognostic criteria are expected in the future. The heterogeneous nature of the disease means that there is a wide range of mutations that may become amenable to targeted therapy in the future.