Background: The classification of acute myeloid leukemia (AML) has long been overseen by the World Health Organization (WHO) and published into a series of “Blue Books.” These ledgers serve as the reference manual for AML classification and, in turn, classification-based treatment decisions. In 2022, two separate groups, each of which included hematologic oncologists, hematopathologists, and geneticists – developed and published two parallel classification systems for AML. One is from the WHO (WHO 5th edition), and a second is from an International Advisory Consortium (International Consortium Classification [ICC]). Summary: Both modern classification systems originated from the revised 4th edition of the WHO Blue Books and thus share many similarities. There are never-the-less several important differences with the potential to substantially alter disease classification, access to clinical trials, and treatment decision-making. In this manuscript, we review the organization of the WHO and ICC classification systems for AML with emphasis on their similarities and differences, followed by areas in which their application to clinical scenarios may present difficulties. Key Messages: (1) The ICC and WHO 5th edition are concordant for the majority of AMLs. (2) Key differences between AML classification by the ICC and WHO 5th edition include (a) the overall framework of classification, (b) AML-defining blast threshold, definition of myelodysplasia-related AML, and (c) strategy for assigning therapy-related or germline-associated AML modifiers.

Consensus classification of acute myeloid leukemia (AML) began in the 1970s with the French-American-British system [1], which organized AMLs by morphologic appearance (i.e., differentiation) of leukemic myeloblasts. In 2001, the FAB criteria were revised and incorporated into the 3rd edition of Blue Books on Classification of Tumors and Hematopoietic and Lymphoid Tissues, published by the World Health Organization (WHO) and the International Agency for the Research on Cancer (IARC) [2]. Development of the early WHO criteria for AML classification was the responsibility of a Clinical Advisory Committee (CAC) made up of leading hematopathologists, hematologic oncologists, and geneticists [3]. The aim of this endeavor was standardization of diagnostic criteria to permit accurate recording of cases entered on trial protocols and to provide a reference standard on which to apply new diagnostic techniques. The 4th (2008) [4] and revision to 4th (2016) [5‒7] edition of these Blue Books were subsequently published under the oversite of the WHO’s CAC and incorporated increasingly sophisticated genetic details about AML into their classification schemes.

More recently, classification of AML was revisited by two separate and parallel groups of leukemia experts who have published separate classification schemes for AML. One system is outlined in the 5th edition of the WHO and IARC Blue Book series [8], the other was published separately and titled the “International Consensus Classification” [9]. Importantly, both originate from the same revised 4th edition of the WHO Blue Book, and thus they share many similarities. For most clinical scenarios, the two classification systems are well aligned. However, the two systems also have multiple important distinctions in their nomenclature for AML – distinctions which may pose challenges to clinical researchers, treating physicians, and patients themselves [10].

In this manuscript, we review the organization of the WHO and ICC classification systems for AML with emphasis on their similarities and differences, followed by areas in which their application to clinical scenarios may present difficulties. Prior to the unification of these two classification schemes, which these authors hope to be forthcoming, we also present a strategy for working between both systems to accurately relay diagnostic information to patients and select short- and long-term treatments for AML in 2023 and beyond [10]. Paraphrased into the common vernacular: #it is complicated.

The framework for AML classification in the 5th edition of the WHO’s “Blue Book” divides AMLs into one of the following categories [8]: (1) “AML with defining genetic abnormalities,” (2) AML, defined by differentiation, and (3) myeloid sarcoma (Table 1) [11]. In general, genetic lesions are given increased significance, while blast percentage is de-emphasized. This limited hierarchy reflects the accumulation of data demonstrating that blast percentage has less prognostic significance than do the molecular characteristics of the blasts or overall functionality of the patient, and genetically defined AML is prioritized over differentiation [12‒14].

Table 1.

WHO 5th edition categories of acute myeloid leukemia (AML)

AML, defined by genetic abnormalities 
 Acute promyelocytic leukemia with PML::RARA fusion 
 AML with RUNX1::RUNX1T1 fusion 
 AML with CBFB::MYH11 fusion 
 AML with DEK::NUP214 fusion 
 AML with RBM15::MRTFA fusion 
 AML with BCR::ABL1 fusion** requires 20% myeloblasts 
 AML with KMT2A rearrangement 
 AML with MECOM rearrangement 
 AML with NUP98 rearrangement 
 AML with NPM1 mutation 
 AML with CEBPA mutation** requires 20% myeloblasts 
 AML, myelodysplasia-related** requires 20% myeloblasts 
  Prior history of MDS or MDS/MPN 
  Defining cytogenetic abnormalities 
   Complex karyotype (≥3 abnormalities) 
   Del(5q)/t(5q) 
   −7/del(7q)/t(7q) 
   Del(11q) 
   Del(12p)/t(12p) 
   −13/del(13q) 
   Del(17p)/t(17p)/iso(17q) 
   Idic(X)(q13) 
  Defining somatic mutations 
   ASXL1 
   BCOR 
   EZH2 
   SF3B1 
   SRSF2 
   STAG2 
   UAKF2 
   ZRSR2 
 AML with other defined genetic alterations 
AML, defined by differentiation** requires exclusion of genetic abnormalities listed above, prior MDS, MDS/MPN, or MPN, and myeloid neoplasm post-cytotoxic therapy (MN-pCT) 
 AML with minimal differentiation 
 AML without maturation 
 AML with maturation 
 Acute basophilic leukemia 
 Acute myelomonocytic leukemia 
 Acute monocytic leukemia 
 Acute erythroid leukemia 
 Acute megakaryoblastic leukemia 
Secondary myeloid neoplasm – separate group, nomenclature for MN above, with addition of the following descriptors 
 Post-cytotoxic therapy (pCT) 
 Associated with germline predisposition 
AML, defined by genetic abnormalities 
 Acute promyelocytic leukemia with PML::RARA fusion 
 AML with RUNX1::RUNX1T1 fusion 
 AML with CBFB::MYH11 fusion 
 AML with DEK::NUP214 fusion 
 AML with RBM15::MRTFA fusion 
 AML with BCR::ABL1 fusion** requires 20% myeloblasts 
 AML with KMT2A rearrangement 
 AML with MECOM rearrangement 
 AML with NUP98 rearrangement 
 AML with NPM1 mutation 
 AML with CEBPA mutation** requires 20% myeloblasts 
 AML, myelodysplasia-related** requires 20% myeloblasts 
  Prior history of MDS or MDS/MPN 
  Defining cytogenetic abnormalities 
   Complex karyotype (≥3 abnormalities) 
   Del(5q)/t(5q) 
   −7/del(7q)/t(7q) 
   Del(11q) 
   Del(12p)/t(12p) 
   −13/del(13q) 
   Del(17p)/t(17p)/iso(17q) 
   Idic(X)(q13) 
  Defining somatic mutations 
   ASXL1 
   BCOR 
   EZH2 
   SF3B1 
   SRSF2 
   STAG2 
   UAKF2 
   ZRSR2 
 AML with other defined genetic alterations 
AML, defined by differentiation** requires exclusion of genetic abnormalities listed above, prior MDS, MDS/MPN, or MPN, and myeloid neoplasm post-cytotoxic therapy (MN-pCT) 
 AML with minimal differentiation 
 AML without maturation 
 AML with maturation 
 Acute basophilic leukemia 
 Acute myelomonocytic leukemia 
 Acute monocytic leukemia 
 Acute erythroid leukemia 
 Acute megakaryoblastic leukemia 
Secondary myeloid neoplasm – separate group, nomenclature for MN above, with addition of the following descriptors 
 Post-cytotoxic therapy (pCT) 
 Associated with germline predisposition 

This table summarizes the categorization of AML according to the WHO 5th edition.

The category of AML with defining genetic abnormalities can be subdivided into those with gene fusions, rearrangements, mutations, myelodysplasia-related, and “other defined” genetic alterations (Table 1). Gene fusions specifically recognized include acute promyelocytic leukemia with PML::RARA fusion and AML with either RUNX1::RUNX1T1, CBFB::MYH11, DEK::NUP214, RBM15::MRTFA (formerly RBM15::MKL1), or BCR::ABL1 fusions. Genetic rearrangements with discrete subcategories include AML with KMT2A, MECOM, or NUP98 rearrangements. Genetic mutations identifying distinct subcategories of AML include NPM1 and CEBPA. Detection of the indicated chromosomal changes is sufficient to define AML as opposed to MDS or other myeloid malignancy, with the exception of the BCR::ABL1 fusion and CEBPA mutations – both of which require the presence of at least 20% myeloblasts for the diagnosis of AML with defined genetic abnormalities. The subcategory “AML with other defined genetic alterations” was developed as a placeholder for nascent genetic lesions, which do not yet merit their own subcategory, and replaces AML not otherwise specified (NOS).

The category of AML, myelodysplasia-related (AML-MR), was significantly revised. Diagnosis of AML-MR requires ≥20% myeloblasts and the presence of specific myelodysplasia-associated genetic abnormalities (Table 1). AML-MR may be diagnosed in myeloid malignancies that meet either criteria, whether de novo or as progression of previously diagnosed MDS or MDS/MPN. The cytogenetic abnormalities that define AML-MR include complex karyotype, 5q deletion or loss of 5q due to unbalanced translocation, monosomy 7, 7q deletion, or loss of 7q due to unbalanced translocation, 11q deletion, 12p deletion or loss of 12p due to unbalanced translocation, monosomy 13 or 13q deletion, 17p deletion or loss of 17p due to unbalanced translocation, isochromosome 17q, idic(X)(q13). This listing should be helpful for cytogeneticists who are designing fluorescence in situ hybridization (FISH) panels for urgent diagnostic testing. Defining somatic mutations include those affecting ASXL1, BCOR, EZH2, SF3B1, SRSF2, STAG2, UAKF2, and ZRSR2 [11].

The category of “AML, defined by differentiation” is the second tier of AML classification in the WHO 5th edition and includes those AMLs that lack defining genetic lesions or prior diagnosis of MDS (Table 1). Subcategories include AML with minimal differentiation, without maturation, and with maturation, acute basophilic leukemia, acute myelomonocytic leukemia, acute monocytic leukemia, acute erythroid leukemia, and acute megakaryoblastic leukemia. The definitions for the specific criteria for these differentiation-defined AMLs can be reviewed in the publication by Khoury et al. [8].

Myeloid sarcomas are given their own category, as was the case in the revised 4th edition of the WHO’s Blue Book [5, 7]. Genetic characterization of these tumors is strongly encouraged, with concurrent comparison to bone marrow sampling. For those patients with isolated chloromas and without overt bone marrow leukemia, clonal hematopoietic lesions or other genetic abnormalities may nevertheless still be identified in the bone marrow [15, 16].

The 5th edition of WHO revised their definition of secondary myeloid neoplasms, which now includes [1] myeloid neoplasm post-cytotoxic therapy and [2] myeloid neoplasm with germline predisposition [8]. The category of myeloid neoplasm with germline predisposition is further subclassified into myeloid neoplasm with germline predisposition without preexisting platelet or organ dysfunction, myeloid neoplasm with germline predisposition and preexisting platelet disorder, and myeloid neoplasm with germline predisposition and potential organ dysfunction. Specific entities within each subcategory are defined by both clinical phenotype and genetic lesions. Excluded from the category of “secondary myeloid neoplasms” are leukemic transformation of MPN (included in the MPN rubric) and AML following MDS or MDS/MPN (classified as AML-MR as previously discussed). Myeloid neoplasms that occur in relationship to prior cytotoxic therapy (-pCT) or with a germline predisposition retain their primary classification but are contained within a separate group of secondary myeloid neoplasms. For these malignancies, “-pCT” or “with germline predisposition” is added to the diagnosis to indicate their inclusion within the group of secondary myeloid neoplasms. The WHO specifies the medication class that could trigger application of the modifier -pCT and includes PARP1 inhibitors and excludes methotrexate.

The 2022 ICC of AMLs [9] applies a structured and hierarchical classification structure that divides AML and related neoplasms into the following prioritized list: (1) AML with recurrent genetic abnormalities (requiring ≥10% myeloblasts, with the exception of AML with t(9;22)(q34.1lq11.2)/BCR::ABL, which requires ≥20% myeloblasts), (2) AML or MDS/AML (>20% or 10–20% myeloblasts, respectively), and (3) myeloid sarcoma (Table 2). In addition to these three broad categories, there are four diagnostic modifiers. Unique to the 2022 ICC is the hierarchical organization of AML with TP53 mutation, then myelodysplasia-related gene mutations (ASXL2, BCOR, EZH2, SF3B1, SRSF2, STAG2, U2AF1, ZRSR2, and RUNX1), followed by myelodysplasia-related cytogenetic abnormalities [complex karyotype, del(5q)/t(5q)/add(5q), −7/del(7q), +8, del(12p)/t(12p)/add(12p), i(17q), −17/add(17p) or del (17p), del(20q), or idic(x)(q14)]. If AML does not contain any of the specified genetic lesions, it is assigned the classification NOS.

Table 2.

ICC Classification of AML and related neoplasms

AML with recurrent genetic abnormalities (≥10% blasts) 
 Acute promyelocytic leukemia with t(15;17)/PML::RARA fusion 
 AML with t(8;21)/RUNX1::RUNX1T1 
 AML with inv(16) or t(16;16)/CBFB::MYH11 
 AML with t(9;11)/MLLT3::KMT2A 
 AML with t(6;9)/DEK::NUP214 
 AML with inv(3) or t(3;3)/GATA2, MECOM(EVI1) 
 AML with other rare recurring translocations 
 AML with mutated NPM1 
 AML with in-frame bZIP mutated CEBPA 
 AML with t(9;22)/BCR::ABL1** requires 20% myeloblasts 
Categories designated AML (≥20% blasts) or MDS/AML (if 10–19% blasts) 
 AML with mutated TP53 
 Acute myelodysplasia-related gene mutations 
  ASXL1 
  BCOR 
  EZH2 
  SF3B1 
  SRSF2 
  STAG2 
  U2AF1 
  ZRSR2 
  RUNX1 
 Acute myelodysplasia-related cytogenetic abnormalities 
  Complex karyotype (≥3 abnormalities) 
  Del(5q)/t(5q)/add(5q) 
  −7/del(7q) 
  +8 
  Del(12p)/t(12p)/add(12p) 
  Del(17p)/Iso(17q), −17/add(17p) 
  Del(20q) 
  Idic(X)(q13) 
AML, not otherwise specified (AML-NOS) 
AML with recurrent genetic abnormalities (≥10% blasts) 
 Acute promyelocytic leukemia with t(15;17)/PML::RARA fusion 
 AML with t(8;21)/RUNX1::RUNX1T1 
 AML with inv(16) or t(16;16)/CBFB::MYH11 
 AML with t(9;11)/MLLT3::KMT2A 
 AML with t(6;9)/DEK::NUP214 
 AML with inv(3) or t(3;3)/GATA2, MECOM(EVI1) 
 AML with other rare recurring translocations 
 AML with mutated NPM1 
 AML with in-frame bZIP mutated CEBPA 
 AML with t(9;22)/BCR::ABL1** requires 20% myeloblasts 
Categories designated AML (≥20% blasts) or MDS/AML (if 10–19% blasts) 
 AML with mutated TP53 
 Acute myelodysplasia-related gene mutations 
  ASXL1 
  BCOR 
  EZH2 
  SF3B1 
  SRSF2 
  STAG2 
  U2AF1 
  ZRSR2 
  RUNX1 
 Acute myelodysplasia-related cytogenetic abnormalities 
  Complex karyotype (≥3 abnormalities) 
  Del(5q)/t(5q)/add(5q) 
  −7/del(7q) 
  +8 
  Del(12p)/t(12p)/add(12p) 
  Del(17p)/Iso(17q), −17/add(17p) 
  Del(20q) 
  Idic(X)(q13) 
AML, not otherwise specified (AML-NOS) 

Categorization of AML according to the ICC. Categories are listed in numerically prioritized, descending order.

May append the following qualifiers to any above diagnosis:

Therapy-related;

prior MDS;

prior MDS/MPN;

germline predisposition.

Diagnostic modifiers are not discrete categories of AML but are instead appended to the primary AML classification. These include (1) therapy-related (chemotherapy, radiotherapy, immune modifiers), (2) progressing from MDS, (3) progressing from MDS/MPN, and (4) with germline predisposition. Unlike the WHO 5th edition, the ICC does not specify the cytotoxic agents that merit addition of the “therapy-related” qualifier.

AML classification by the ICC and 5th edition of WHO share two common themes – the first being recognition that quantification of myeloblast percentage at a single timepoint may be imprecise and insufficient to describe disease behavior. The second is that genetic characterization of a myeloid malignancy best predicts clinical behavior and response to therapy. Both the ICC and 5th edition of WHO combine genetic descriptors with more flexible myeloblast cutoffs for the diagnosis of AML. Even when blasts are <20%, AML can be diagnosed in either classification system so long as specific genetic lesions are present. This can be thought of as a myeloblast “sliding scale” that hinges on disease genetics (Table 3). Although the ICC and WHO both acknowledge that myeloblast percentage does not define AML, the relative weight given to blast percentage in the two systems is evident in the following two cases (Table 4).

Table 3.

Comparison of WHO 5th edition and ICC AML classifications

Overall Framework – defined groups (WHO) versus progressive hierarchy (ICC) 
 WHO 5th edition 
  AML divided into two groups, in descending order of priority 
   AML defined by genetic abnormalities 
   AML defined by differentiation 
   Secondary myeloid neoplasms retain MN nomenclature above, but included within separate group 
 ICC AML classification 
  Hierarchical prioritization of genetic abnormalities, in descending order of priority 
   AML with recurrent genetic abnormalities (10% blasts) 
   Mutated TP53 (VAF >10%) 
   AML with myelodysplasia-related gene mutations 
   AML with myelodysplasia-related cytogenetic abnormalities 
   AML, not otherwise specified 
AML-defining blast threshold in blood or bone marrow – defined by specific subcategories (both), with ICC including new MDS/AML intermediary as “leukemia” 
 WHO 5th edition 
  AML defined by genetic abnormalities: no lower blast threshold, with the exceptions of AML defined by BCR::ABL1, CEBPA mutation, and myelodysplasia-related (each requires ≥20% blasts) 
  AML defined by differentiation: ≥20% blasts 
 ICC AML classification 
  AML with recurrent genetic abnormalities: ≥ 10% blasts, except for BCR::ABL1 (≥20% blasts) 
   Myeloid malignancy with TP53 mutation, myelodysplasia-related genetic or cytogenetic abnormalities, and AML-NOS divided into MDS/AML (10–19% blasts) versus AML (≥20% blasts) 
Definition of myelodysplasia-related AML (AML-MR vs. AML with myelodysplasia-related gene mutations or cytogenetic abnormalities) – specificity of genetic lesions, role of prior MDS or MDS/MPN in application of classification 
 Specific genetic lesions 
  Included in WHO 5th edition, excluded from ICC 
   t(7q), del(11a), −13, del(13q), t(17p)−7/del(7q) 
  Included in ICC, excluded from WHO 5th edition 
   Add(5q), +8, add(12p), −17, add(17p), del(20q) 
   RUNX1 
 Role of prior MDS or MDS/MPN in classification 
  WHO 5th edition 
   Prior known MDS or MDS/MPN is sufficient for inclusion within AML-MR 
   No additional modifier or separate classification 
  ICC 
   AML with specified genetic or cytogenetic abnormalities is a genetically defined classification, falling within a strict hierarchical classification schema 
   Prior known MDS or MDS/MPN may be added as a qualifier to any AML classification 
Therapy-related or germline-associated AML – utilization of antecedent MDS or MDS/MPN to classify AML and addition of a modifier or qualifier to connote contributing factors 
 WHO 5th edition – considered a separate group 
  “Myeloid neoplasm”- post-cytotoxic therapy (MN-pCT), with defined cytotoxic therapies that merit this addition to primary MN diagnosis 
  “Myeloid neoplasm” associated with germline predisposition 
 ICC – diagnostic qualifier added to primary diagnosis 
  Therapy related – does not specify cytotoxic agents 
  With germline predisposition 
  Progressing from MDS or MDS/MPN 
Overall Framework – defined groups (WHO) versus progressive hierarchy (ICC) 
 WHO 5th edition 
  AML divided into two groups, in descending order of priority 
   AML defined by genetic abnormalities 
   AML defined by differentiation 
   Secondary myeloid neoplasms retain MN nomenclature above, but included within separate group 
 ICC AML classification 
  Hierarchical prioritization of genetic abnormalities, in descending order of priority 
   AML with recurrent genetic abnormalities (10% blasts) 
   Mutated TP53 (VAF >10%) 
   AML with myelodysplasia-related gene mutations 
   AML with myelodysplasia-related cytogenetic abnormalities 
   AML, not otherwise specified 
AML-defining blast threshold in blood or bone marrow – defined by specific subcategories (both), with ICC including new MDS/AML intermediary as “leukemia” 
 WHO 5th edition 
  AML defined by genetic abnormalities: no lower blast threshold, with the exceptions of AML defined by BCR::ABL1, CEBPA mutation, and myelodysplasia-related (each requires ≥20% blasts) 
  AML defined by differentiation: ≥20% blasts 
 ICC AML classification 
  AML with recurrent genetic abnormalities: ≥ 10% blasts, except for BCR::ABL1 (≥20% blasts) 
   Myeloid malignancy with TP53 mutation, myelodysplasia-related genetic or cytogenetic abnormalities, and AML-NOS divided into MDS/AML (10–19% blasts) versus AML (≥20% blasts) 
Definition of myelodysplasia-related AML (AML-MR vs. AML with myelodysplasia-related gene mutations or cytogenetic abnormalities) – specificity of genetic lesions, role of prior MDS or MDS/MPN in application of classification 
 Specific genetic lesions 
  Included in WHO 5th edition, excluded from ICC 
   t(7q), del(11a), −13, del(13q), t(17p)−7/del(7q) 
  Included in ICC, excluded from WHO 5th edition 
   Add(5q), +8, add(12p), −17, add(17p), del(20q) 
   RUNX1 
 Role of prior MDS or MDS/MPN in classification 
  WHO 5th edition 
   Prior known MDS or MDS/MPN is sufficient for inclusion within AML-MR 
   No additional modifier or separate classification 
  ICC 
   AML with specified genetic or cytogenetic abnormalities is a genetically defined classification, falling within a strict hierarchical classification schema 
   Prior known MDS or MDS/MPN may be added as a qualifier to any AML classification 
Therapy-related or germline-associated AML – utilization of antecedent MDS or MDS/MPN to classify AML and addition of a modifier or qualifier to connote contributing factors 
 WHO 5th edition – considered a separate group 
  “Myeloid neoplasm”- post-cytotoxic therapy (MN-pCT), with defined cytotoxic therapies that merit this addition to primary MN diagnosis 
  “Myeloid neoplasm” associated with germline predisposition 
 ICC – diagnostic qualifier added to primary diagnosis 
  Therapy related – does not specify cytotoxic agents 
  With germline predisposition 
  Progressing from MDS or MDS/MPN 

This table summarizes the primary similarities and differences between WHO 5th edition and ICC classification of AML.

Table 4.

Summary of Examples

CaseDiagnosis: WHO 5th EdDiagnosis: ICC
Example 1 AML defined by genetic abnormality AML with recurrent genetic abnormalities 
Blood and marrow with >20% myeloblasts. Normal karyotype (46XY), isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%) 
Example 2 MDS-IB2 AML with recurrent genetic abnormalities 
Blood and marrow with 12% myeloblasts. Normal karyotype (46XY), isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%) 
Example 3 AML with other defined genetic alterations AML with myelodysplasia-related gene mutation 
Marrow with 40% blasts. Cytogenetics include del(20q) and trisomy 8. RUNX1 mutation 
Example 4 AML, myelodysplasia-related AML with mutated TP53 
Marrow with 20% blasts. Complex karyotype, TP53 mutation (VAF 20%) 
CaseDiagnosis: WHO 5th EdDiagnosis: ICC
Example 1 AML defined by genetic abnormality AML with recurrent genetic abnormalities 
Blood and marrow with >20% myeloblasts. Normal karyotype (46XY), isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%) 
Example 2 MDS-IB2 AML with recurrent genetic abnormalities 
Blood and marrow with 12% myeloblasts. Normal karyotype (46XY), isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%) 
Example 3 AML with other defined genetic alterations AML with myelodysplasia-related gene mutation 
Marrow with 40% blasts. Cytogenetics include del(20q) and trisomy 8. RUNX1 mutation 
Example 4 AML, myelodysplasia-related AML with mutated TP53 
Marrow with 20% blasts. Complex karyotype, TP53 mutation (VAF 20%) 

This table summarizes the case and WHO 5th edition and ICC classification of each example discussed in this review.

Example 1

Fifty-six-year-old male with peripheral blood notable for thrombocytopenia, anemia, and leukocyte count of 2,000 cells/µL, of which 80% are myeloblasts, was presented. Bone marrow examination reveals a hypercellular marrow with sheets of myeloblasts. Cytogenetics reveals normal karyotype (46XY), and next-generation sequencing identifies an isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%). The patient has no personal history of previous cytopenias, malignancy, chemotherapy, or radiation, and no family history of malignancy.

In this example, the diagnosis is AML with recurrent genetic abnormalities (ICC), or the analogous AML with defining genetic abnormalities (WHO 5th edition). The myeloblast percentage being >20% provides straightforward clarity to the diagnosis of AML; the in-frame bZIP CEBPA mutation is a “defining genetic abnormality” in both classification systems [17‒19].

Example 2

Fifty-six-year-old male with peripheral blood notable for thrombocytopenia, anemia, and leukocyte count of 2,000 cells/µL, of which none are myeloblasts, was presented. Bone marrow examination reveals a hypercellular marrow with 12% myeloblasts. Cytogenetics reveals normal karyotype (46XY), and next-generation sequencing identifies an isolated in-frame bZIP monoallelic CEBPA mutation (VAF of 35%), without concurrent NPM1 mutation. The patient has no personal history of previous cytopenias, malignancy, chemotherapy, or radiation, and no family history of malignancy.

With the myeloblast percentage in blood/marrow reduced to 12%, the differences between these two classification systems are more apparent. Per the ICC, the diagnosis of “AML with recurrent genetic abnormality” remains so long as the myeloblast percentage is 10% or greater. In contrast, if CEBPA mutation is present in isolation, the WHO requires that the myeloblast percentage be at least 20% to diagnose AML, and therefore would classify this as MDS with increased blasts-2 (MDS-IB2). With the exclusion of isolated CEBPA mutations and BCR::ABL1 translocations, the 5th edition of WHO removes the lower cutoff for myeloblast percentage altogether, so long as specific AML-defining genetic abnormalities are present and correlate with morphologic features of the myeloid malignancy (Tables 1-3).

In many instances, the discrepancies between ICC and WHO feel like semantics – the minute differences in nomenclature do not often change management unless it includes or excludes patients from clinical trials. However, in some cases, such as Example 2, the name given to a disease entity may be highly relevant. Is it appropriate to treat a patient with 12% myeloblasts in the marrow and a CEBPA mutation with daunorubicin and cytarabine-based AML induction chemotherapy? By ICC criteria, this patient has “AML with recurrent genetic abnormality,” whereas by WHO criteria they do not. The absence of clinical data supporting the application of AML induction therapy in this type of case underlies the rationale behind the WHO’s strict 20% or greater cutoff for myeloblasts in a myeloid neoplasm with isolated CEBPA mutation. The same conundrum impacts the decision of whether to advance this patient to allogeneic hematopoietic cell transplantation. AML with a normal karyotype and an in-frame bZIP CEBPA mutation is unlikely to be transplanted in first remission given the overall good prognosis of the diagnosis. If instead, this patient is diagnosed with MDS-IB2, albeit a surprising CEBPA mutation, they would deserve consideration for transplant.

In reality, AML with isolated bZIP CEBPA mutation is uncommon. It is more likely that bZIP CEBPA mutations will be identified in combination with other mutations and that clinicians will need to make treatment recommendations based on the relative risk of the other associated genetic abnormalities. For instance, had the same case (12% myeloblasts) also included an NPM1 mutation along with the in-frame bZIP CEBPA mutation, many clinicians would treat this patient with curative-intent chemotherapy. That WHO would classify this entity as AML with recurrent genetic abnormalities gives additional support for this therapeutic strategy. In contrast, were there a STAG2 mutation along with the in-frame bZIP CEBPA mutation, treatment decisions would be driven by the higher risk STAG2 lesion. Again, classification of this disease as MDS-IB2 supports inclusion of allogeneic hematopoietic cell transplantation as part of curative-intent therapy for this disease. Perhaps the take-home message from both classification systems is that emphasis should be on the full genetic landscape of a myeloid malignancy and not exclusively based on blast percentage.

Another important difference between the ICC and WHO is their classification of “high-grade” MDS (Table 3). In the 4th edition of WHO, high-grade MDS with myeloblasts between 10 and 19% were classified as MDS with excess blasts-2 [4, 7]. The ICC now categorizes this cohort of myeloid malignancies as “MDS/AML,” and those with myeloblasts of 20% or greater as AML, myelodysplasia-related. Assignation to either group is dependent on myelodysplasia-related genetic abnormalities and not on morphologic dysplasia.

Both the WHO 5th edition and ICC identify myelodysplasia-related genetic lesions as significant contributors to AML ontogeny. However, the genetic abnormalities identified in each classification system do not fully overlap. The potential diagnostic impact of these discrepancies is presented in Example 3 and described in Table 3 and 4.

Example 3

Sixty-five-year-old female with peripheral blood notable for leukopenia, anemia, and thrombocytopenia, without peripheral myeloblasts, was presented. Bone marrow examination reveals a hypercellular marrow with 40% myeloblasts, cytogenetics reveals the following karyotype: 47,XX,+8,del(20q)[18]/46,XX[2], and next-generation sequencing identifies a mutation in RUNX1 but no translocations.

This neoplasia would be classified as “AML with other defined genetic alterations” using the WHO framework. By comparison, the ICC annotates this entity as “AML with a myelodysplasia-related gene mutation.” The differences in classification arise from the subtle variations in the cytogenetic abnormalities and somatic mutations that define myelodysplasia-informed AMLs in the WHO 5th edition and ICC. This example highlights the inclusion of +8, del(20q) and RUNX1 mutations as genetic abnormalities that define AML, myelodysplasia-related genetic/cytogenetic abnormalities in the ICC but do not define AML-MR according to WHO criteria. This distinction is important for patient access to clinical trials designed to specifically address myelodysplasia-related AML and also impacts decision-making regarding induction chemotherapy. For example, liposomal daunorubicin and cytarabine (CPX-351; Vyxeos) are approved for the treatment of AML that arises from myelodysplasia. According to ICC, but not WHO criteria, liposomal daunorubicin and cytarabine would be an appropriate treatment choice for this clinical example.

While not included in this example, there are additional variations between the WHO and ICC with respect to their nomenclature for leukemias that arise from a known MDS, MDS/MPN, or MPN (Table 3). By the WHO 5th edition system, leukemic progression of MDS or MDS/MPN is sufficient to classify a leukemia as an AML-MR irrespective of genetic abnormalities. Leukemia occurring in a patient with a prior MPN is not classified within the AML rubric but rather within the MPN classification structure. In contrast, the ICC adds a modifier to the primary leukemia classification to connote prior MDS, MDS/MPN, or MPN, so long as the MDS or MPN was diagnosed 3 months or more prior to diagnosis of AML. They do not change the primary classification of the leukemia.

Presumably most AMLs with antecedent MDS would contain a genetic lesion that facilitates classification as AML-MR by the WHO, but as exemplified by the genetic abnormalities presented in Example 3, this may not be the case. Had the patient detailed in Example 3 been diagnosed with MDS prior to her leukemia diagnosis, it would be sufficient to modify the WHO 5th edition diagnosis to AML-MR. The revised WHO diagnosis would then be more consistent with the diagnosis provided by the ICC framework, which would be only subtly modified to “AML with myelodysplasia-related gene mutation and with prior MDS.”

While not included in this example, both the ICC and current WHO systems highlight the significance of -pCT and germline predisposition in the classification of AML. In both the WHO 5th edition and the ICC, annotation of a myeloid malignancy as “secondary” is appended to the primary diagnosis as a modifier for both antecedent exposure to cytotoxic therapy and for concurrent germline cancer predisposition. The WHO specifies the medication class that could trigger application of the modifier -pCT and includes PARP1 inhibitors and excludes methotrexate. The ICC does not specify agents whose exposure prior to development of a myeloid malignancy merits notation as “secondary to cytotoxic therapy.”

While the ICC and WHO 5th edition have some “defining genetic abnormalities” in common, they do not overlap fully. This is most notable for TP53, which is not included as a defining mutation by the WHO but is for the ICC. The following case exemplifies the challenges that this discrepancy can cause (Tables 3, 4).

Example 4

Seventy-year-old female with peripheral blood notable for leukopenia, anemia, and thrombocytopenia, without peripheral myeloblasts, was presented. Bone marrow examination reveals a hypercellular marrow with 20% myeloblasts. Cytogenetics reveals a complex karyotype, and next-generation sequencing identifies a mutation in the TP53 gene, with a VAF of 20%. The patient has no personal history of previous cytopenias, malignancy, chemotherapy, or radiation, and no family history of malignancy.

The WHO classifies this entity as AML-MR, given the complex karyotype. The ICC classifies this entity as AML with mutated TP53. This is based on detection of mutated TP53 and myeloblasts of at least 20%. In the ICC diagnostic hierarchy, mutated TP53 with a VAF ≥10% is heavily weighted and supersedes AML with myelodysplasia-related cytogenetic abnormalities. Given the overwhelming and negative clinical impact of TP53 lesions in AML, these cases are perhaps best annotated within the ICC system. AML with complex karyotype but without TP53 mutation is more likely to respond to hypomethylating agent and venetoclax doublet than is AML with complex karyotype and a TP53 mutation.

A category that exists within the ICC but not WHO is MDS/AML and AML NOS. This nomenclature identifies MDS/AML and AML entities that lack defining recurrent genetic abnormalities, TP53 mutation, or myelodysplasia-related gene or cytogenetic abnormalities. This category highlights the near-exclusive reliance of ICC on genetic characterization of myeloid neoplasms. If a genetic lesion is identified within the malignancy but has not risen to the significance of discrete subcategorization, the WHO annotates this as “AML with defining genetic abnormalities,” but subclassified as “AML with other defined genetic alterations.” The rationale being that this classification allows for newly described recurrent genetic abnormalities to inform AML nomenclature while facilitating further investigation into their significance as driver versus secondary genetic lesions. Should a defining genetic abnormality not be identified whatsoever, the WHO retains “AML, defined by differentiation” as a class, which is equivalent to the ICC-designated “AML, NOS.”

Whether working from the ICC or 5th edition of WHO, a takeaway message from both revisions to the revised 4th edition WHO guidelines is that precise diagnosis of AML requires both cytogenetic and molecular disease characterization. For myeloid neoplasms with 10–19% myeloblasts, even general classification as MDS, MDS/AML, or AML cannot be made without these data. At a practical level, this means that any clinician engaged in the care of patients with myeloid malignancy needs to partner not only with excellent hematopathologists for the purpose of morphologic and flow cytometric studies but also with robust genetics labs capable of returning rapid, accurate, and complete cytogenetic and molecular testing.

This begs the question: how fast is acceptable? The 2022 European Leukemia Network recommendations suggest 3–5 days as the preferred turn-around time for many genetic test results [20], and recent review papers have adopted this as a management goal [21]. However, while this may be currently achievable for some larger tertiary care centers, it is often not possible in smaller centers that send out their samples. If samples are being sent out to reference testing labs, sample transit time and incorporation of these results into the final pathology report within the home institution’s electronic record need to also be considered. It is incumbent on each group providing care for patients with myeloid malignancy to review their genetic testing resources and develop strategies to avoid misdiagnosis of patients. While traditionally, AML has been considered a disease needing urgent therapy, waiting for complete molecular and genetic data is feasible and often preferable in order to select the optimal management [22].

When appropriate diagnostic resources are in place, the next question to be addressed is the mechanism by which formal diagnoses are relayed to clinicians and patients. At a minimum, the pertinent morphologic description, blast count, flow cytometry-based lineage description, cytogenetics, and molecular results should be included on the final diagnostic summary. This is a standard reporting strategy at many if not most institutions. What is less clear-cut is how to coalesce these disease descriptors into a formal diagnosis. With the revised 4th edition of WHO still in use at many centers, as well as the ICC and 5th edition of WHO, some centers, including The Dana-Farber Cancer Institute, have moved to reporting all three in their diagnostic summary. The benefit to this approach is that it facilitates treatment selection based on expected disease behavior, including participation in clinical trials with specific diagnostic inclusion and exclusion criteria. It may also facilitate interpretation of treatment response in highly specific disease entities, which is increasingly important in the era of precision medicine and rapid development of targeted therapies.

Clinical researchers and trial authors need also be aware of these differences as they write inclusion and exclusion criteria. It may be more straightforward, in some cases, to describe the characteristics of the eligible patients by myeloblast percentage or by genetic lesion rather than the generalizations of AML or myelodysplastic syndrome. Over-specifying enrollment to a classification standard may, in some cases, eliminate patients who would be eligible in the competing system. The alternative risk is that inclusion criteria become long and unwieldy if one cannot simply refer to a single, standard diagnostic classification.

We should also be mindful of the impact of two separate classification structures on how patients understand their own disease. Under the 21st Century Cures Act, patients now have immediate access to their medical records; different classification systems add to patient confusion and anxiety when accessing these data. Patients expect that their hematologist will provide them with a diagnosis, around which patient and physician will frame their discussions of treatment, supportive care, lifestyle modifications, and prognosis. How does one explain the current selection of diagnostic platforms to a patient? Sharing a new diagnosis of AML with a patient is complex, and mention of “genetic tests” that further characterize and inform treatment decisions throughout their disease course is standard. However, only rarely has there been equivocation between calling a disease MDS versus AML. It is important to patients that they understand whether their treatment is to “prevent transformation into AML” or to “treat AML,” even when prognostically there is little difference between MDS-IB2 or MDS/AML. Even more troubling would be for a patient to have the impression that their disease cannot be diagnosed – it is a subtle distinction between a disease that is challenging to diagnose and a disease that can carry several names based upon the lens of classification that is applied.

How do we move forward with dual classification schemes? Many of us will likely rely heavily on our hematopathology colleagues for clarification when two options appear possible. There are also several published schematics that may be helpful in moving forward with clinical decisions (Fig. 1) [10]. For some myeloid neoplasms, the nomenclature differences between the ICC and WHO are negligible. These are highlighted by the first case iteration in “Example 1” and the important but rare AML with BCR::ABL1 translocation. But for many other aggressive myeloid neoplasms, the waters are murky. Perhaps the morphologic, flow cytometric, and genetic characterization of a myeloid malignancy are sufficient. These data along with the nomenclature that could be applied by using the revised 4th WHO, 5th WHO, and ICC provide ample information about the predicted disease behavior.

Fig. 1.

Recommended approach for unification of ICC and 5th edition of WHO strategies for classification of AML. This schematic proposes one strategy for coalescing ICC and WHO classification strategies for AML and was developed by leading experts in AML biology and clinical care [10] (re-published with permission from Elsevier).

Fig. 1.

Recommended approach for unification of ICC and 5th edition of WHO strategies for classification of AML. This schematic proposes one strategy for coalescing ICC and WHO classification strategies for AML and was developed by leading experts in AML biology and clinical care [10] (re-published with permission from Elsevier).

Close modal

A conundrum faced by both the 5th WHO and ICC is classification of myeloid malignancies arising in individuals with prior exposure to cytotoxic therapy. There are generally predictable timelines, cytogenetic changes, and molecular mutation profiles in myeloid neoplasms that are the biological consequence of topoisomerase II (1–5 year latency, KMT2A, CBFB mutations) or alkylator/radiation (5–10 year latency, 5q/7q deletion, monosomal karyotype, TP53 mutations) therapy. Both antecedent genotoxic therapy and genetic features are consistently “high risk” in an AML with KMT2A mutation arising shortly after doxorubicin exposure. In this instance, the modifiers “-pCT” or “therapy-related” are accurate descriptors of the patient’s medical history and of their AML’s high-risk biology. However, if a normal karyotype AML with isolated NPM1 mutation arises greater than 30 years after genotoxin exposure, do the “-pCT” and “therapy-related” modifiers accurately relay the expected behavior of disease? This challenge, in addition to that of interpreting and discriminating the significance of germline mutations identified by next-generation myeloid malignancy sequencing panels, has yet to be met by either system. Hopefully, as more is learned about the genetics and molecular aspects of myeloid malignancies and the correspondence with prognosis, the WHO and ICC classification schemas can be unified into a single, cohesive platform.

Authors acknowledge and appreciate the insights provided by Dr. Sanam Loghavi, Department of Hematopathology, Division of Pathology-Lab Medicine Division, MD Anderson (Houston, TX, USA).

Karen-Sue Carlson served on the Institutional Review Board for the National Marrow Donor Program. Ashley Cunningham has no conflicts of interest to report. Maximilian Stahl served on the advisory board for Novartis, Kymera, Sierra Oncology, GSK and Rigel; consulted for Boston Consulting and Dedham group; and participated in GME activity for Novartis, Curis Oncology, Haymarket Media, and Clinical Care Options. Eric Winer has served on advisory boards for Novartis, Takeda, and Curis. Laura C. Michaelis served as a consultant/advisor for Jazz Pharmaceuticals, Incyte, Celgene Corp, Novartis Corp, Sierra Oncology, and Nkarta. She received research funding from Jazz Pharmaceuticals.

This review article was written and revised without external funding sources.

K.-S.C. and L.C.M. prepared the initial draft of this manuscript, which was reviewed and substantially edited by K.-S.C., A.C., M.S., E.W., and L.C.M.

1.
Bennett
JM
,
Catovsky
D
,
Daniel
MT
,
Flandrin
G
,
Galton
DA
,
Gralnick
HR
.
Proposals for the classification of the acute leukaemias. French-American-British (FAB) co-operative group
.
Br J Haematol
.
1976
;
33
(
4
):
451
8
.
2.
Jaffe
ES
Pathology and genetics of tumours of haematopoietic and lymphoid tissues
IARC
2001
.
3.
Harris
NL
,
Jaffe
ES
,
Diebold
J
,
Flandrin
G
,
Muller-Hermelink
HK
,
Vardiman
J
.
The World Health Organization classification of neoplastic diseases of the hematopoietic and lymphoid tissues. Report of the clinical advisory committee meeting, airlie house, Virginia, November, 1997
.
Ann Oncol
.
1999
;
10
(
12
):
1419
32
.
4.
Swerdlow
SH
,
Campo
E
,
Harris
NL
,
Jaffe
ES
,
Pileri
SA
,
Stein
H
WHO classification of tumours of haematopoietic and lymphoid tissues
International agency for research on cancer Lyon, France
2008
.
5.
Arber
DA
,
Orazi
A
,
Hasserjian
R
,
Thiele
J
,
Borowitz
MJ
,
Le Beau
MM
.
The 2016 revision to the World Health Organization classification of myeloid neoplasms and acute leukemia
.
Blood
.
2016
;
127
(
20
):
2391
405
.
6.
Swerdlow
SH
,
Campo
E
,
Pileri
SA
,
Harris
NL
,
Stein
H
,
Siebert
R
.
The 2016 revision of the World Health Organization classification of lymphoid neoplasms
.
Blood
.
2016
;
127
(
20
):
2375
90
.
7.
Swerdlow
SH
,
Campo
E
,
Harris
NL
,
Jaffe
ES
,
Pileri
SA
,
Stein
H
WHO classification of tumours of haematopoietic and lymphoid tissues, Revised 4th Edition
Revised 4th ed
IARC
2017
.
8.
Khoury
JD
,
Solary
E
,
Abla
O
,
Akkari
Y
,
Alaggio
R
,
Apperley
JF
.
The 5th edition of the World Health Organization classification of haematolymphoid tumours: myeloid and histiocytic/dendritic neoplasms
.
Leukemia
.
2022
;
36
(
7
):
1703
19
.
9.
Arber
DA
,
Orazi
A
,
Hasserjian
RP
,
Borowitz
MJ
,
Calvo
KR
,
Kvasnicka
HM
.
International Consensus classification of myeloid neoplasms and acute leukemias: integrating morphologic, clinical, and genomic data
.
Blood
.
2022
;
140
(
11
):
1200
28
.
10.
Shallis
RM
,
Daver
N
,
Altman
JK
,
Komrokji
RS
,
Pollyea
DA
,
Badar
T
.
Standardising acute myeloid leukaemia classification systems: a perspective from a panel of international experts
.
Lancet Haematol
.
2023
;
10
(
9
):
e767
76
.
11.
Shimony
S
,
Stahl
M
,
Stone
RM
.
Acute myeloid leukemia: 2023 update on diagnosis, risk-stratification, and management
.
Am J Hematol
.
2023
;
98
(
3
):
502
26
.
12.
Estey
E
,
Hasserjian
RP
,
Döhner
H
.
Distinguishing AML from MDS: a fixed blast percentage may no longer be optimal
.
Blood
.
2022
;
139
(
3
):
323
32
.
13.
DiNardo
CD
,
Garcia-Manero
G
,
Kantarjian
HM
.
Time to blur the blast boundaries
.
Cancer
.
2022
;
128
(
8
):
1568
70
.
14.
Hasserjian
RP
,
Campigotto
F
,
Klepeis
V
,
Fu
B
,
Wang
SA
,
Bueso-Ramos
C
.
De novo acute myeloid leukemia with 20–29% blasts is less aggressive than acute myeloid leukemia with ≥30% blasts in older adults: a Bone Marrow Pathology Group study
.
Am J Hematol
.
2014
89
11
E193
9
.
15.
Engel
NW
,
Reinert
J
,
Borchert
NM
,
Panagiota
V
,
Gabdoulline
R
,
Thol
F
.
Newly diagnosed isolated myeloid sarcoma-paired NGS panel analysis of extramedullary tumor and bone marrow
.
Ann Hematol
.
2021
;
100
(
2
):
499
503
.
16.
Greenland
NY
,
Van Ziffle
JA
,
Liu
YC
,
Qi
Z
,
Prakash
S
,
Wang
L
.
Genomic analysis in myeloid sarcoma and comparison with paired acute myeloid leukemia
.
Hum Pathol
.
2021
;
108
:
76
83
.
17.
Tarlock
K
,
Lamble
AJ
,
Wang
YC
,
Gerbing
RB
,
Ries
RE
,
Loken
MR
.
CEBPA-bZip mutations are associated with favorable prognosis in de novo AML: a report from the Children’s Oncology Group
.
Blood
.
2021
;
138
(
13
):
1137
47
.
18.
Taube
F
,
Georgi
JA
,
Kramer
M
,
Stasik
S
,
Middeke
JM
,
Röllig
C
.
CEBPA mutations in 4,708 patients with acute myeloid leukemia: differential impact of bZIP and TAD mutations on outcome
.
Blood
.
2022
;
139
(
1
):
87
103
.
19.
Wakita
S
,
Sakaguchi
M
,
Oh
I
,
Kako
S
,
Toya
T
,
Najima
Y
.
Prognostic impact of CEBPA bZIP domain mutation in acute myeloid leukemia
.
Blood Adv
.
2022
;
6
(
1
):
238
47
.
20.
Döhner
H
,
Estey
E
,
Grimwade
D
,
Amadori
S
,
Appelbaum
FR
,
Büchner
T
.
Diagnosis and management of AML in adults: 2017 ELN recommendations from an international expert panel
.
Blood
.
2017
;
129
(
4
):
424
47
.
21.
DiNardo
CD
,
Erba
HP
,
Freeman
SD
,
Wei
AH
.
Acute myeloid leukaemia
.
Lancet
.
2023
;
401
(
10393
):
2073
86
.
22.
Röllig
C
,
Kramer
M
,
Schliemann
C
,
Mikesch
JH
,
Steffen
B
,
Krämer
A
.
Does time from diagnosis to treatment affect the prognosis of patients with newly diagnosed acute myeloid leukemia
.
Blood
.
2020
;
136
(
7
):
823
30
.