Laboratory Features

Unlike cancers of solid organs, AML is widely disseminated at diagnosis; hence, a TNM-type staging system is of little practical value. For many years, classification of AML was almost exclusively morphologic and was dominated by the French-American-British (FAB) system.¹³ AML was diagnosed only if the marrow or blood contained >30% myeloid blasts, identified primarily by histochemical stains specific to, for example, the granulocytic (peroxidase) or monocytic lineages (esterase). Various FAB subtypes, M0 through M7, were recognized. A problem with the FAB system was that its clinical relevance was limited, with the very important exception of subtype M3, acute promyelocytic leukemia (APL), which unlike other types of AML, is routinely curable with all-trans retinoic acid and arsenic trioxide with or without an anthracycline such as idarubicin. Furthermore, as more was learned about the prognostic significance of cytogenetics and various molecular abnormalities, it became obvious that any classification system needed to incorporate this information. This recognition led to development of the World Health Organization (WHO) system, currently the most widely accepted means of classifying AML.¹⁴

Unlike the FAB system, the WHO criterion for a diagnosis of AML is >20% blasts, with the exceptions noted here. This change from >30% reflected observations that after accounting for various prognostic factors, such as cytogenetics or whether AML was de novo or secondary (unfavorable), patients with 21-30% blasts fared identically to those with >30% blasts when treated identically.¹⁵ Some experts have contended, however, that any minimum blast criterion (many cooperative groups use 10%) is too simplistic and ignores the complex medical decision making inherent in AML treatment. Rather than relying on histochemical stains to identify blasts of myeloid lineage, cell surface antigens now serve this purpose (Table 81.1).

The great majority of patients with AML fall into one of the first four WHO categories: AML with recurrent genetic changes, AML with myelodysplasia-related changes, therapy-related AML, or AML not otherwise specified (NOS) (Table 81.2).

AML with Recurrent Genetic Changes

Patients with the inv(16), t(16;16), or t(8;21) cytogenetic abnormality are lumped together into a “core-binding factor” (CBF) group. CBF refers to regulators of transcription that have a β subunit and three α subunits. RUNX1 and RUNX1T1 t(8;21) affects one of the α subunits, whereas CBFB-MYH11 affects the β subunit. Depending on the age of the population in question, CBF AML comprises 5-15% of cases and is relatively sensitive to cytosine arabinoside (ara-C), thus benefiting from “high” doses of this drug and having a relatively good prognosis.¹⁶ Despite being considered a single group, there are important clinical differences between t(8;21) and inv(16). For example, the latter is much more likely to have multiple remissions and thus longer survival.¹⁷

The t(15;17) abnormality is characteristic of FAB subtype M3, as discussed earlier. About 5% of cases of APL have normal cytogenetics, thus it is critical to assess for the PML-RARα aberration associated with the t(15;17).¹⁸ The abnormal juxtaposition of the PML and RARα genes results in the impaired promyelocytic differentiation that is the hallmark of APL. The abnormal positioning of PML leads to a characteristic pattern of nuclear immunofluorescence, demonstration of which is the quickest way to diagnose APL.¹⁹

Abnormalities in the mixed lineage leukemia (MLL) gene are associated with deletions or translocations in the long arm of chromosome 11 (11q). The clinical patterns associated with different MLL partners vary considerably. For example, pairing of MLL with MLLT3, resulting in t(9;11), leads to a prognosis better than that seen when MLL is paired with other partners.²⁰ Thus, as with CBF, classifying an AML as “MLL AML” is itself a gross oversimplification. MLL abnormalities are particularly important to recognize because a class of drugs (DOT1L inhibitors) appear to target MLL and have produced (mainly minor) responses in clinical trials.²¹

The t(6;9) (which transposes the DEK and NUP214 genes) often is associated with basophilia but is perhaps mainly of interest because 80% of cases have an internal tand em duplication (ITD) of the FLT3 gene; t (6;9) is the AML subtype most frequently associated with FLT3 ITD.22 The 2008 version of WHO system does not recognize FLT3 ITD as falling in the category of recurrent genetic abnormalities (Table 81.2). This is because in perhaps 30% of cases, patients who present with an FLT3 ITD are either not FLT3 ITD positive at relapse or vice versa.23 Hence, it would appear that FLT3 ITDs are not foundational abnormalities. In contrast, a change in NPM1 or CEBPA mutation status between diagnosis and relapse is believed to be less common, leading to inclusion of these in the WHO system.²⁴ Nonetheless, FLT3 ITD, NPM1 mutations, and CEBPA double mutations all have major prognostic importance.²⁵NPM1 mutations and double (biallelic) CEBPA mutations convey a relatively favorable prognosis, whereas the opposite is true of FLT3 ITDs; mutations in the tyrosine kinase domain (TKD) of FLT3 are of less prognostic relevance. Of course, the significance of these aberrations depends on the context in which they occur. For example, patients in whom >50% of alleles have an FLT3 ITD have a poorer prognosis than patients with lower allelic burdens, whereas patients with NPM1 mutations and FLT3 ITDs do worse than those with only NPM1 mutations.²⁶

It is also worth noting that patients with RUNX1-RUNX1TI [i.e., t(8;21)] or CBFB-MYH11 [i.e., inv16 or t(16;16)] are considered by WHO to have AML regardless of blast count.¹⁴ This can readily be justified on clinical grounds, given that these patients have responses to AML therapy similar to those of patients with >20% blasts.¹⁵ It is likely that blast count eventually will be considered immaterial to a diagnosis of AML in patients with NPM1 mutations and /or CEBPA double mutations.

AML with Myelodysplasia-Related Changes

Patients fall into this category if they have >=20% blasts and either morphologic features of myelodysplasia (MDS), a history of MDS, or a myeloproliferative neoplasm (MPN), or MDS-related cytogenetic abnormalities without the specific genetic abnormalities characteristic of AML, as noted in Table 81.2. MDS-related cytogenetic abnormalities are complex karyotypes (at least three distinct clonal abnormalities), monosomy of or deletion of the long arm of chromosomes 5 and /or 7 (-5,del 5q, -7, del 7q), t(3;5)(q25;q34) and changes in chromosome 11q not involving MLL.

This formulation has several problems. For example, concordance is far from perfect among pathologists in describing dysplasia. The prognostic relevance of dysplasia is in doubt.²⁷ Finally, as noted earlier, the distinction between de novo and secondary AML (e.g., arising on a background of MDS or MPN) is probably better made based on genomics than clinical history.³

Therapy-Related AML (t-AML)

t-AMLs generally are divided into those associated with use of alkylating agents (e.g., cyclophosphamide or melphalan) and those associated with use of topoisomerase II-reactive drugs (e.g., anthracyclines or etoposide).² The former are characterized by a 5- to 10-year latency period and complex cytogenetics. The latter more often present with abnormalities of chromosome 11q or balanced translocations, such as t(15;17) or t(8;21). Such cases generally are regarded as “AML with recurrent genetic abnormalities.”¹⁴ This is reasonable from a prognostic stand point, as cases of t-AML with t(15;17), t(8;21), or inv(16) have prognoses similar to those of cases in which these abnormalities arise seemingly spontaneously.²⁸

Assignment of a case to the t-AML category is not always straightforward, such as when the latency period is >10 years after administration of alkylating agents. Here again, defining secondary AML based on genomic patterns likely is better than doing so based on clinical history.

AML NOS

These cases fall outside the other types of AML specified in Table 81.2. They comprise about 25% of all cases. Essentially, they are classified using FAB nomenclature. There likely will be fewer cases as more genetic subtypes of AML are recognized by the WHO.

Using data from 5,848 patients with AML NOS treated in studies by the Southwest Oncology Group (SWOG), MD and erson Cancer Center, U.K. Medical Research Council/National Cancer Research Institute (MRC/NCRI), and Dutch-Belgian Cooperative Trial Group for Hematology/Oncology and the Swiss Group for Clinical Cancer Research (HOVON/SAKK), Walter et al.²⁹ studied the prognostic effect of different AML NOS subtypes, such as AML with minimal differentiation (M0), AML without maturation (M1), and AML with maturation (M2). After multivariate adjustment, FAB M0 was independently associated with a significantly lower likelihood of achieving complete remission and inferior relapse-free and overall survival compared with M1, M2, M4, M5, and M6, with inconclusive data regarding M7. However, when attention was restricted to known NPM1-negative patients, FAB M0 was no longer associated with worse outcomes; restricting attention to patients known to be NPM1 negative/CEPBA negative (i.e., excluding the provisional entities of “AML with mutated NPM1” and “AML with mutated CEBPA”; Table 81.2) did not affect this result. Hence, FAB subclassification of AML NOS does not provide prognostic information if data on NPM1 and CEBPA mutations are available.