review

The power and potential of integrated diagnostics in acute

myeloid leukaemia

Torsten Haferlach and Ines Schmidts

MLL Munich Leukemia Laboratory, Munich, Germany

based solely on cytomorphological and cytochemical features.

Summary
In the past, present and future, cytomorphology was, is and
The field of acute myeloid leukaemia (AML) diagnostics, will still be at the forefront of haematological diagnostics. It
initially based solely on morphological assessment, has inte- provides fast assessments of specimens and thus enables
grated more and more disciplines. Today, state-of-the-art time- and cost-effective step-wise diagnostics. Abnormalities
AML diagnostics relies on cytomorphology, cytochemistry, in cell morphology are readily identified by the trained
immunophenotyping, cytogenetics and molecular genetics. haematologist and allow for distinction between normal and
Only the integration of all of these methods allows for a aberrant (and potentially leukaemic) cells. Evaluation of the
comprehensive and complementary characterisation of each percentage and relative distribution of erythropoiesis, granu-
case, which is prerequisite for optimal AML diagnosis and lopoiesis and monocytopoiesis identifies a range of haemato-
management. Here, we will review why multidisciplinary logical disorders. Cytochemical staining of non-specific
diagnostics is mandatory today and will gain even more esterase and myeloperoxidase and iron staining in many
importance in the future, especially in the context of cases enable or facilitate cell lineage determination and evalu-
precision medicine. We will discuss ideas and strategies that ation of dysplasia. It is required for disease classification, val-
are likely to shape and improve multidisciplinary diagnos- idation of diagnoses, differential diagnostics and assessment
tics in AML and may even overcome some of today’s gold of disease kinetics and response.
standards. This includes recent technical advances that pro- A bone marrow biopsy provides complementary informa-
vide genome-wide molecular insights. The enormous tion on cells in the tissue context, for example on cellularity
amount of data obtained by these latter techniques repre- and histotopography as well as the proportion and matura-
sents a great challenge, but also a unique chance. We will tion of haematopoietic cells (Swerdlow et al., 2017). In cases
reflect on how this increase in knowledge can be incorpo- with fatty marrow or acute myeloid leukaemia (AML) with
rated into the routine to pave the way for personalised fibrosis, the blast count can only be reliably assessed in the
medicine in AML. biopsy specimen. Moreover, histologic evaluation aided by
immunohistochemistry can validate classification, especially
Keywords: multidisciplinary diagnostics, acute myeloid leu- for cases of the subgroup of AML not otherwise specified
kaemia, precision medicine, AML diagnosis and manage- (NOS), and facilitate the differentiation between myelodys-
ment, artificial intelligence. plastic syndromes (MDS) and AML (Orazi, 2007).

Immunophenotyping
Current diagnostic workup in acute myeloid
The presence and expression strength of antigens constitute
leukaemia
the immunophenotype of a cell, which is indicative of cell
lineage identity as well as the degree of maturation. Modern
Cytomorphology
flow cytometers allow the parallel inspection of 8–10 mark-
Cytomorphology is the indispensable starting point in the ers. Leukaemic cells show aberrations in their immunophe-
diagnostics of haematological diseases. This is also true from notype, which can be broadly categorised as follows
a historical perspective — the first classification efforts were (Swerdlow et al., 2017):
1 Expression of cross-lineage antigens (e.g. expression of
lymphoid markers in AML cells, such as CD19+ AML).
*Correspondence: Torsten Haferlach, MLL Munich Leukemia
2 Asynchronous expression of maturational markers (e.g.
Laboratory, Max-Lebsche-Platz 31, 81377 Munich, Germany.
concomitant expression of CD34 and CD11b in AML).
E-mail: torsten.haferlach@mll.com

First published online 6 December 2019 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
doi: 10.1111/bjh.16360 and John Wiley & Sons Ltd.British Journal of Haematology, 2020, 188, 36–48
This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and
reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
 Review

3 Absent or decreased antigen expression of typical markers development of targeted therapeutics and is increasingly used
(e.g. HLA-DR-negative AML). as a therapeutic decision-making tool.
4 Overexpression of antigens (e.g. CD33++ CD34++ AML). With respect to AML diagnostics, polymerase chain reac-
tion (PCR)-based approaches as well as next-generation
Flow cytometry is a crucial tool for detection, characterisa-
sequencing (NGS) represent the gold standard.
tion as well as quantification of healthy and malignant cell
The PCR allows the specific amplification of known target
populations. Depending on the respective disorder, it plays
sequences. By the method of quantitative PCR (qPCR), aber-
an essential or supporting role for classification and differen-
rations cannot only be detected but also sensitively moni-
tial diagnostics. In a myriad of haematological neoplasms, it
tored. Template amplification is measured in real time using
has a pivotal function in the assessment of response, disease
fluorescent probes and quantification is performed relative to
kinetics and especially in AML also in the detection of mini-
a standard. The input of cDNA, that is reverse transcribed
mal/measurable residual disease (MRD). Immunophenotyp-
(RT) RNA, permits transcript detection and quantification
ing also encompasses the method of immunohistochemistry,
(RT-qPCR). The major advantage of PCR-based assays is
which can be performed on bone marrow biopsy specimens.
their high sensitivity of up to 10 6.
The stereotypical distribution patterns of myeloid leukaemias
Next-generation sequencing, in contrast to older sequenc-
become apparent when the biopsy specimen is stained with
ing techniques (e.g. Sanger sequencing), offers the capability
an antibody directed against the blast marker CD34. Using
for massive parallelisation. This enables sequencing of hun-
suitable antibodies, a multitude of diagnostic questions can
dreds of samples and/or genomic loci in one run. Panel-
be addressed, for example on cell lineage identity and degree
based sequencing represents the current state-of-the-art NGS
of maturation (Swerdlow et al., 2017).
methodology — such a panel could for example comprise all
genes known to be associated with myeloid neoplasms that
Cytogenetics show diagnostic and/or clinical relevance today. Panel-based
sequencing has led to better molecular characterisation in
Cytogenetics encompasses the techniques of chromosome AML and it can be used also now for MRD.
analysis and fluorescence in situ hybridisation (FISH). Chro- Next-generation sequencing is a highly versatile platform
mosome analysis is performed by chromosome banding of and in the future new innovative NGS applications are likely
metaphases. Non-malignant cells generally have a normal to transition from research to routine diagnostics. With the
karyotype (46,XX or 46,XY), while the leukaemic karyotype technique of whole genome or whole exome (i.e. all protein-
might show acquired numerical or structural chromosomal coding genes) sequencing (WGS/WES), sequence variations
aberrations. FISH relies on the use of fluorescent probes that as well as numerical and structural aberrations can be
are directed against specific chromosomal loci. This tech- detected. Sequencing the whole transcriptome (WTS/RNA-
nique can be performed on interphase as well as on meta- Seq) allows for genome-wide gene expression analysis, the
phase chromosomes. Probes can be either used to screen for detection of fusion transcripts and also for mutational analy-
known and/or suspected cytogenetic aberrations or, if direc- sis of expressed loci.
ted against centromeres, to detect numerical aberrations. The
use of so-called 24-colour FISH allows characterisation/vali-
dation of complex aberrations found in chromosome analysis Optional diagnostic methods
after banding. While chromosome analysis enables a gen-
ome-wide, comprehensive evaluation, FISH provides a tar- Gene expression profiling
geted, but fast approach. Subtypes of MDS and acute Gene expression profiling (GEP) has shown its potential to
leukaemia are defined by specific cytogenetic aberrations. finally be integrated into routine diagnostic settings. Several
Beside its relevance for WHO classification, the most crucial studies had demonstrated that classification can benefit
role for cytogenetics in acute leukaemia is prognostic stratifi- from GEP. Differentiation between AML and acute lym-
cation. Cytogenetics is also important for the monitoring of phoblastic leukaemia (ALL) can be realized based solely on
disease kinetics, response assessment and the characterisation expression profiles (Golub et al., 1999; Haferlach et al.,
of clonal evolution. 2010) and GEP-based approaches were able to reproduce
classification of genetically defined subtypes, while at the
same time providing insight into the underlying pathobiol-
Molecular genetics
ogy (Schoch et al., 2002; Debernardi et al., 2003; Valk
Molecular genetics has rapidly evolved into an indispensable et al., 2004; Haferlach et al., 2010; Visani et al., 2018). One
diagnostic discipline and has brought about major advances study proposed the improvement of classification by intro-
in our understanding of the molecular landscape of cancers, duction of GEP data for subclassification of AML with a
including AML. It has significantly contributed to optimisa- normal karyotype (AML-NK) (Bullinger et al., 2004). Here,
tion of not only classification, but also of prognostication GEP identified two distinct groups among AML-NK cases
and residual disease monitoring. Moreover, it has aided the and stratification was of prognostic relevance, a finding that

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 37
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

was validated in an independent study (Bullinger et al., molecular genetic methods have challenged the comprehen-
2004; Radmacher et al., 2006). Further research efforts have siveness of AML classification, yet again.
highlighted GEP’s value for risk stratification (Valk et al., Since its introduction in 2001 (Jaffe et al., 2001), the
2004; Huang et al., 2017; Visani et al., 2018). With respect WHO classification has unified well established (cytomor-
to AML therapy, gene expression profiles have been found phology, immunophenotyping, chromosome analysis) and
to differ between responders and non-responders in induc- molecular-orientated diagnostic disciplines (FISH, molecular
tion therapy (Heuser et al., 2005; Herold et al., 2018) and genetics) for a comprehensive classification of haematological
can be used to predict drug sensitivity (Raponi et al., 2008; neoplasms. Today, 11 AML subtypes are defined by genetics
Visani et al., 2017; Tyner et al., 2018). Despite all these (eight cytogenetically, three by gene mutations). The com-
promising studies, also reviewed in Visani et al. (2018), plete transition from the FAB to the WHO classification
microarray-based GEP has never been implemented into therefore also signifies the paradigm change from phenotype
routine diagnostics in AML — today’s options with RNA- to genotype (Swerdlow et al., 2017).
Seq will hopefully change this.
The WHO classification
Epigenetic analysis
Table I provides an overview of the different categories and
Similarly, epigenetics plays an important role in leukaemoge- subtypes of AML and related neoplasms. Further details on
nesis and AML disease biology (Wouters & Delwel, 2016). AML subtypes are given in the following paragraph. Here,
Partly, aberrant DNA methylomes with resulting gene expres- we only focus on details of classification that need specific
sion deregulation in AML can be explained by recurrent diagnostic approaches and sometimes directly lead to prog-
aberrations in epigenetic regulators, such as DNMT3A, nostic and therapeutic consequences.
ASXL1, TET2, KMT2A, IDH1 and IDH2 (Wouters & Delwel, Although genetics plays a crucial and partly entity-defining
2016). However, even in the absence of said somatic aberra- role for the classification of AML according to WHO, the
tions, distinct classes, defined by their DNA methylome, are presence of ≥20% blasts in peripheral blood or bone marrow
discernible and of prognostic relevance (Figueroa et al., is a requirement for AML diagnosis, making cytomorphology
2010). Further research into the epigenome could lead to essential for AML classification.
improved classification — especially in cases for which no However, there are three exceptions, for which the AML
leukaemia-driving (cyto-)genetic event can be identified. Epi- diagnosis is made independent of blast count:
genetic compounds (e.g. hypomethylating agents) are already
1 AML with t(8;21)(q22;q22.1); RUNX1-RUNX1T1.
an integral part of the therapeutic arsenal and AML diagnos-
2 AML with inv(16)(p13.1q22) or t(16;16)(p13.1;q22);
tics would be likely to benefit from inclusion of epigenetic
CBFB-MYH11.
analytics. However, there is, as of yet, no prospect of epige-
3 APL (acute promyelocytic leukaemia) with t(15;17)(q22;
netic analysis in the clinical routine.
q11-12); PML-RARA.

In the absence of subtype-defining aberrations, cases with

The need for integrated diagnostics
AML can be categorised into AML with myelodysplasia-re-
lated changes, therapy-related neoplasms or AML, NOS.
From phenotype to genotype
Classification of AML with myelodyplasia-related changes
The introduction of the French–American–British (FAB) depends on multidisciplinary diagnostics. In addition to the
classification, which was initially based solely on morphologi- presence of ≥20% blasts, cases must exhibit MDS-related fea-
cal characteristics (Bennett et al., 1976; Bennett et al., 1985), tures which are defined as (i) a history of MDS or MDS/
set the stage for modern AML diagnostics by providing MPN (myelodysplastic/myeloproliferative neoplasm); (ii)
objective criteria for patient stratification. The soon reached MDS-related cytogenetic abnormalities; or (iii) multilineage
conclusion was that morphology alone cannot uncover the dysplasia (as defined by the presence of ≥50% dysplastic cells
full range of heterogeneity in AML, which led to the intro- in 2–3 haematopoietic lineages, assessed by bone marrow
duction of immunophenotypic criteria to AML classification cytomorphology). Up to ~25% of de novo AML cases present
(Bennett et al., 1991; Catovsky et al., 1991). In parallel, sev- with multilineage dysplasia (Haferlach et al., 2003). The pres-
eral AML-specific cytogenetic aberrations have been ence of multilineage dysplasia, however, does not influence
described and some of them could be linked to a specific the prognosis in AML with NPM1 mutation (Falini et al.,
phenotype (Second MIC Cooperative Study Group, 1988). 2010) or AML with biallelic CEBPA mutation (Bacher et al.,
For the following decade, cytomorphology and 2012), NPM1 and biallelic CEBPA mutations thus take diag-
immunophenotyping in some cases together with cytogenet- nostic precedence over multilineage dysplasia (Swerdlow
ics were the basis for AML classification (Second MIC Coop- et al., 2017).
erative Study Group, 1988). Meanwhile, new methods Prior exposure to cytotoxic therapy or radiotherapy iden-
increasingly allowed molecular insight and FISH as well as tifies therapy-related neoplasms. A differentiation into

38 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
 Review

Table I. Acute myeloid leukaemia (AML) and related precursor neo- subtypes. With the potential exception of pure erythroid leu-
plasms according to the WHO classification (2017). kaemia, AML, NOS subclassification does not provide prog-
Subclassfication Subtypes nostic information per se based on its morphoplogy
(Swerdlow et al., 2017).
AML with
recurrent • AML with t(8;21)(q22;q22.1); RUNX1-
genetic RUNX1T1 Multidisciplinarity is a prerequisite for an optimal
abnormalities • AML with inv(16)(p13.1q22) or t(16;16) diagnosis
(p13.1;q22); CBFB-MYH11
• Acute promyelocytic leukaemia with PML- Multidisciplinarity aids fast diagnosis as well as differential
RARA diagnosis of acute leukaemias (AML vs. ALL). While today
• AML with t(9;11)(p21.3;q23.3); KMT2A- two methods can be applied: cytomorphology including cyto-
MLLT3 chemistry (turnaround time 2–4 h) and immunophenotyping
• AML with t(6;9)(p23;q34.1); DEK-NUP214
(turnaround time 2–6 h), in many cases both methods are
• AML with inv(3)(q21.3q26.2) or t(3;3)
used in parallel, depending on the respective facility —
(q21.3;q26.2); GATA2, MECOM
• AML (megakaryoblastic) with t(1;22)(p13.3; which consolidates and validates findings from either
q13.1); RBM15-MKL1 method. However, if for morphological questions trephine
• Provisional entity: AML with BCR-ABL1 biopsies including histochemistry and immunohistochemistry
are the methods of choice, immunophenotyping is the
AML with gene mutations
quicker technique to be applied.
• AML with mutated NPM1 In addition to classification, risk stratification already at
• AML with biallelic mutation of CEBPA diagnosis is of great clinical importance. Cytogenetics and
• Provisional entity: AML with mutated
molecular genetics provide the most powerful information
RUNX1
for prognosis. In order to determine the risk group of a
AML with given case according to the recommendations of the Euro-
myelodysplasia- pean LeukemiaNet (ELN), a comprehensive genetic charac-
related terisation is required that reaches significantly beyond the 11
changes
genetic aberrations that define an entity (see also Table II).
Therapy-related
The importance of genetic evaluation both for classification
myeloid neoplasms
• AML with minimal differentiation and risk stratification firmly establishes chromosome analysis,
AML, not
• AML without maturation FISH and molecular genetics in the diagnostic evaluation of
otherwise
• AML with maturation
specified every (suspected) case of AML.
• Acute myelomonocytic leukaemia
In conclusion, state-of-the-art classification in AML in
• Acute monoblastic and monocytic leukaemia
accordance with the WHO guidelines and also for ELN prog-
• Pure erythroid leukaemia
• Acute megakaryoblastic leukaemia nostication relies on the diagnostic disciplines of cytomor-
• Acute basophilic leukaemia phology, immunophenotyping, cytogenetics and molecular
• Acute panmyelosis with myelofibrosis genetics. A combined and interdisciplinary approach is
needed to harmonise reports and increase the quality of any
Myeloid sarcoma
• Transient abnormal myelopoiesis associated single method by implementation of knowledge already avail-
Myeloid
with Down syndrome able from others. Of note, turnaround times are differing
proliferations • Myeloid leukaemia associated with Down
associated with
from some hours to a week.
syndrome
Down syndrome
The emerging molecular landscape of AML
In AML classification, it is foreseeable that more subtypes
defined by specific molecular aberrations or their co-occu-
individual subtypes of myeloid neoplasms is not intended in rance will be introduced. Prerequisite for the recognition of
the WHO classification, although it might be clinically mean- an entity by the WHO is its clinical relevance and biological
ingful from our perspective. This category includes not only homogeneity, which makes it discernible not only by a given
cases with AML, but also with t-MDS or t-MDS/MPN-over- genetic aberration, but also by its clinical, morphological
laps (Swerdlow et al., 2017). and/or immunophenotypical characteristics (Swerdlow et al.,
Cases that do not fulfil the diagnostic criteria of all AML 2017).
subtypes mentioned above are by exclusion categorised as A comprehensive effort to gain deeper insight into the
AML, NOS. Within this category, subclassification is based molecular landscape of AML led to the suggestion of 11 dis-
on morphological and immunophenotypical criteria, and tinct classes — based solely on genetic features (Papaem-
most entities of this group are synonymous to the old FAB manuil et al., 2016). Basis for this proposition was a study

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 39
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

Table II. Risk stratification according to European LeukemiaNet Both methods provide genome-wide and unbiased insight
(ELN) recommendations (D€ ohner et al., 2017). into chromosomal and sequence aberrations (WGS/WES) as
Risk group Genetic aberration well as into (aberrant) gene expression and sequence alter-
ations for expressed genes (RNA-Seq). This not only repro-
Favourable t(8;21)(q22;q22.1); RUNX1-RUNX1T1 duces already known but also might lead to identification of
inv(16)(p13.1q22) or t(16;16)(p13.1;q22);
new leukaemia-driving or -promoting molecular aberrations
CBFB-MYH11
with not only diagnostic, but also prognostic, predictive or
Mutated NPM1 without FLT3-ITD or with low
therapeutic impact.
FLT3-ITD allelic ratio (<0
5)
Biallelic mutated CEBPA
Why germline matters — Is the outbreak of
Intermediate Mutated NPM1 and high FLT3-ITD allelic ratio
AML predictable?
(≥0
5)
Wild-type NPM1 without FLT3-ITD or with Germline mutations in CEBPA, DDX41, RUNX1, ANKRD26,
FLT3-ITD low allelic ratio (<0
5) (without ETV6 and GATA2 as well as inherited bone marrow failure
adverse-risk genetic lesions) and telomere syndromes predispose an individual to myeloid
t(9;11)(p21.3;q23.3); MLLT3-KMT2A neoplasia. The recognition of distinct disease entities within
Cytogenetic abnormalities not classified as
the recently introduced category of ‘myeloid neoplasms with
favourable or adverse
germline predisposition’ (Swerdlow et al., 2017) will go a
long way to increase our knowledge on hereditary factors
Adverse t(6;9)(p23;q34.1); DEK-NUP214
t(v;11q23.3); KMT2A rearranged that drive or promote AML pathobiology. In addition to
t(9;22)(q34.1;q11.2); BCR-ABL1 that, only few risk factors have been identified, such as
inv(3)(q21.3q26.2) or t(3;3)(q21.3;q26.2); smoking (Fircanis et al., 2014) or prior exposure to cytotoxic
GATA2,MECOM(EVI1) compounds or radiotherapy. Two recently published retro-
5 or del(5q); 7; 17/abn(17p) spective studies have dealt with the question whether one
Complex karyotype, monosomal karyotype can predict the onset of AML within the general population.
Wild-type NPM1 and high FLT3-ITD allelic A predictive AML ‘prodrome’ could be identified by molecu-
ratio (≥0
5) lar genetic screening and the laboratory parameter of red cell
Mutated RUNX1
distribution width (Abelson et al., 2018; Desai et al., 2018).
Mutated ASXL1
Prospective studies will be necessary to determine whether
Mutated TP53
screening for AML will one day be feasible and clinically
With minor adaptions republished, with permission of the American meaningful.
Society of Hematology, from: D€
ohner et al. (2017).

Multidisciplinary diagnostics for prognosis and

risk stratification
with 1540 AML patients for which mutational and cytoge-
netic analysis was performed and correlated with clinical In addition to a patient’s age and performance, genetics rep-
data. For mutational analysis, 111 cancer genes were targeted. resent the single most relevant marker for risk stratification
At least one driver mutation was identified in 96% of cases. in AML (De Kouchkovsky & Abdul-Hay, 2016). Accordingly,
The majority of classes identified by this genetic approach the currently used risk stratification systems consider cytoge-
equalled entities recognised by the WHO classification, while netic and molecular aberrations of high prognostic relevance.
three represented novel genetic classes (compare Table III). Two stratification models are well established: risk stratifica-
However, a subset of patients (4%) fulfilled the criteria for tion recommended by the ELN (see also Table II) (D€ ohner
two or more classes. This poses the question which aberra- et al., 2017) and the risk model of the Medical Research
tion would take diagnostic precedence. In the end, 11% of Council (MRC) (Grimwade et al., 2016), which is shown in
patients remained unclassified (Papaemmanuil et al., 2016). Table IV.
However, this is far better than the current AML, NOS cate- Acute myeloid leukaemia with NPM1 provides a case study
gory which includes 25–30% of AML patients (Swerdlow for the prognostic importance of the genetic context. A recent
et al., 2017). It is still a goal to find meaningful and relevant study highlighted for example that the presence of NPM1 muta-
characteristics that allow for classification of all these unclas- tions did not compensate for the negative impact of concomi-
sifiable cases in the future. This is likely to improve patient tant adverse cytogenetic aberrations. The study was restricted to
risk stratification, treatment and outcome. patients without concomitant FLT3-ITD mutation or with a
Advances in diagnostic techniques have led to improve- low FLT3-ITD allelic ratio (Angenendt et al., 2019). A DNMT3A
ments of AML classification and will continue to do so. mutation is the most frequently occurring co-mutation in AML
Among the innovative techniques that will likely drive opti- with mutated NPM1 (Ivey et al., 2016; Cappelli et al., 2019).
misation of AML classification are WGS/WES and RNA-Seq. The prognostic impact of this NPM1–DNMT3A co-mutational

40 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
 Review

Table III. Genetic subclassification of acute myeloid leukaemia Table IV. Medical Research Council risk stratification according to
(AML) according to Papaemmanuil et al. (2016). Grimwade et al. (2016).

Class recognised by the WHO Risk group Genetic aberration

Suggested genetic class classification (2017)
Favourable t(15;17)(q22;q21)/PML-RARA
inv(16) AML with inv(16) or t(16;16); t(8;21)(q22;q22)/RUNX1-RUNX1T1
CBFB-MYH11 inv(16)(p13q22)/t(16;16)(p13;q22)/CBFB-MYH11
t(15;17) APL with PML-RARA NPM1 mutation (in absence of FLT3-ITD or
t(8;21) AML wit t(8;21); RUNX1-RUNX1T1 DNMT3A mutation)
KMT2A fusions AML with t(9;11); KMT2A-MLLT3 Biallelic CEBPA mutation
inv(3) AML with inv(3) or t(3;3); GATA2, Intermediate Cytogenetic/molecular genetic abnormalities
not classified as favourable or adverse
MECOM Adverse In the absence of favourable risk cytogenetic/
t(6;9) AML with t(6;9); DEK-NUP214 molecular genetic abnormalities:
NPM1 AML with mutated NPM1 abn(3q) [excluding t(3;5)(q21~25;q31~35)/
Biallelic CEBPA AML with biallelic mutation of CEBPA NPM1-MLF1]
TP53-aneuploidy inv(3)(q21q26)/t(3;3)(q21;q26)/GATA2/EVI1
Chromatin-spliceosome add(5q)/del(5q), 5
IDH2R172 t(5;11)(q35;p15.5)/NUP98-NSD1
t(6;9)(p23;q34)/DEK-NUP214
add(7q)/del(7q), 7
pattern is itself determined by genetic interdependencies. Muta- t(11q23) [excluding t(9;11)(p21~22;q23) and
tions that affect the glycines at position 12 or 13 of NRAS posi- t(11;19)(q23;p13)]
tively influence the prognosis in NPM1–DNMT3A AML t(9;22)(q34;q11)/BCR-ABL
(Papaemmanuil et al., 2016), while an additional FLT3-ITD 17/abn(17p)/TP53 mutation
mutation has a negative prognostic impact (Papaemmanuil Complex karyotype (≥4 unrelated abnormalities)
et al., 2016; Cappelli et al., 2019). Still, risk stratification accord- ASXL1 mutation
ing to current guidelines takes into account only FLT3-ITD DNMT3A mutation
(both ELN and MRC) or DNMT3A (MRC only) co-mutations FLT3-ITD
MLL-PTD
(Grimwade et al., 2016; D€ ohner et al., 2017). With this excep-
RUNX1 mutation
tion, the complete genomic landscape can only partly be consid-
ered by both the ELN and MRC stratification system as it is too With minor adaptions republished, with permission of the American
complex for routine application. Society of Hematology, from: Grimwade et al. (2016).
Several large-scale studies have highlighted the complexity of
the genetic landscape in AML (Cancer Genome Atlas Research
Network, 2013; Papaemmanuil et al., 2016; Tyner et al., 2018). algorithm correlates with sample size (Gerstung et al., 2017), it
However, the mutation status of only a few genes is taken into should be an incentive to combine databases to reach a sample
account by the currently used risk stratification models. This size that allows accurate prediction even for cases with rare (ge-
includes mutations in NPM1, FLT3-ITD, CEBPA (biallelic), netic) features. Data harmonisation and the use of a common
RUNX1, ASXL1 and TP53 (Grimwade et al., 2016; D€ ohner et al., data model (e.g. the OHDSI OMOP: https://www.ohdsi.org/)
2017). In addition, the MRC classification considers KMT2A- are integral to any effort to create a comprehensive unified data-
PTD and DNMT3A mutations (Grimwade et al., 2016). base. Proof-of-concept for the utility of such a knowledge bank-
Although prognostic stratification today is far from trivial, based approach has been demonstrated recently (Gerstung
current risk stratification systems are still oversimplified. Nei- et al., 2017). The authors highlighted how a multistage model,
ther clinical nor patient-specific parameters are incorporated trained on data of the AMLSG cohort (n = 1540) (Papaem-
nor are interdependencies of genetic aberrations. Further manuil et al., 2016) and validated on the TCGA cohort
studies suggest that gene expression analysis as well as DNA (n = 186) (Cancer Genome Atlas Research Network, 2013),
methylation analysis might also provide complementary may aid the decision for or against allogeneic transplantation. A
prognostic information (Valk et al., 2004; Bullinger et al., machine learning algorithm that was trained on the data of
2010; Figueroa et al., 2010; Tyner et al., 2018). 3421 patients is currently under development and already out-
It will be a challenge to determine and validate the influence performs ELN risk classification (Shreve et al., 2019). We envi-
of all possible parameters and to model a universally applicable sion that web applications will make personalised risk
risk stratification system will prove increasingly difficult — if predictions applicable within the clinical setting.
not impossible. Personalised risk stratification might be the We envision several benefits by implementation of a
solution to this problem. For the training of risk prediction patient-specific risk prediction into the clinical routine.
algorithms, large databases that match genomic with clinical Firstly, such a personalised risk stratification algorithm
data are required. Since the predictive accuracy of the trained can be designed to include not only genetic parameters, but

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 41
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

also patient-specific parameters, such as age, fitness and After induction treatment a cytomorphological evaluation
comorbidities as well as laboratory parameters of potential of the bone marrow is performed to determine whether the
prognostic significance (e.g. white blood cell count). patient has achieved a CR, which is defined, among other
Secondly, data on interdependencies of genetic aberrations criteria, by <5% blasts in the bone marrow.
can be considered by such an algorithm. In contrast to this, The criteria of CR and other response definitions as well
it would never be feasible to incorporate all possible genetic as the definition of treatment failure and relapse are indi-
scenarios and co-mutation patterns into a generally applica- cated in Table V.
ble risk stratification model. Integrated diagnostics should assist treatment decision
Thirdly, an accordingly designed and trained algorithm especially in patients above the age of 65. This is underlined
could identify the most relevant and targetable contributors to by a recent study that showed improved outcome for
a patient’s prognosis and thereby aid therapeutic decisions. patients ≥65 years with adverse cytogenetics under azacy-
tidine treatment compared to conventional therapy (which
included 7 + 3 chemotherapy, low-dose cytarabine and best
The power of integrated diagnostics as a supportive care). For a subset of patients the study also eval-
therapeutic decision-making tool uated the influence of molecular aberrations on therapy out-
come. While patients carrying TP53 and NRAS mutations
In general, the therapeutic algorithm for AML can be divided
benefitted from azacytidine treatment, patients with FLT3
into two separate phases:
and TET2 mutations had better outcome under conventional
1 Induction therapy with the goal to achieve a complete therapy regimens (D€ ohner et al., 2018).
remission (CR).
2 Post-remission therapy with the goal to erase residual dis-
Response monitoring and MRD
ease and prevent relapse.
The detection of residual disease in haematological neo-
For both scenarios and time points and also at relapse a
plasms has been improved in parallel to therapy optimisa-
comprehensive and individual diagnostic approach is needed.
tion. Sensitivities of 1:20 (cytomorphology) (Schuurhuis
Applicable techniques include: morphology, cytogenetics,
et al., 2018), or 1:100 (FISH) (Ravandi et al., 2018) were
immunophenotyping and an increasing number of molecular
never thought to be sufficient to reliably monitor diseases
genetic tests. All of the latter assays are needed at diagnosis,
kinetics in AML. This is also reflected by the abbreviation
and morphology at a minimum for the definition of first CR.
‘MRD’, which was initially defined as minimal residual dis-
However, several aspects are important:
ease, and is of today more correctly defined as measurable
1 Diagnostic parameters lead to risk classification that needs residual disease. Sensitivities of 10 4 and 10 6 are needed to
to guide further strategies including allogeneic stem cell assess residual disease, and this can be achieved by using
transplantation (allo-SCT) in first CR in accordance with state-of-the-art molecular approaches or multiparameter flow
current guidelines (MRC, ELN, NCCN) (Grimwade et al., cytometry (MFC) with 8–10 colours (Schuurhuis et al.,
2016; D€ ohner et al., 2017; Tallman et al., 2019). 2018). For MFC two differing approaches are implemented
2 Diagnostic information increasingly leads to individualised in MRD diagnostics: the different-from-normal approach
treatment not only in PML-RARA-positive AML but also (DfN) and the leukaemia-associated immunophenotype
in FLT3-mutated AML. Other findings such as the detec- (LAIP). The latter has to be determined at diagnosis, since it
tion of mutations of IDH, or SF3B1, KIT and others can is patient-specific. With modern flow cytometry a LAIP can
influence choice of drugs. This is also true for the applica- be identified in up to 90% of cases (Swerdlow et al., 2017).
tion of drugs such as the anti-CD33 monoclonal antibody The DfN approach focuses on leukaemia- rather than on
gemtuzumab ozogamicin. patient-specific aberrant markers and allows flow cytometric
3 Age in combination with cytogenetic or molecular data MRD monitoring even if the leukaemic immunophenotype
influence treatment and drug choice, for example at diagnosis is unknown.
in including venetoclax, azacytidine or CPX-351 according Molecular MRD monitoring is dependent on suitable mark-
to the guideline of the National Comprehensive ers, which can be categorised as follows (Schuurhuis et al., 2018):
Cancer Network (NCCN) (Version 3.2019) (Tallman
1 Fusion gene transcripts (PML-RARA for APL).
et al., 2019).
2 Somatic mutations (e.g. NPM1).
4 MRD parameters that can be followed by immunopheno-
3 Aberrant gene expression (e.g. WT1 and EVI1).
typing and/or molecular assays (PCR, digital PCR, NGS)
lead to individual follow-up strategies and treatment However, when choosing a marker for molecular MRD
(Schuurhuis et al., 2018). assessment, several limitations have to be taken into consid-
5 At relapse, the genetic landscape may differ from that at eration. Some potential MRD markers cannot be measured
first diagnosis and complete workup is recommended for with the required sensitivity of 10 4 to 10 6 (e.g. WT1 gene
best rescue therapy. expression levels). There are also various biological situations

42 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
 Review

Table V. Definitions of response, treatment failure and relapse in acute myeloid leukaemia (AML) according to D€
ohner et al. (2017), excerpt of
the table on ‘Response criteria in AML’ in ‘Diagnosis and management of AML in adults: 2017 ELN recommendations from an international
expert panel’.

Category Definition

Response
CR without minimal residual disease (CRMRD ) If studied pretreatment, CR with negativity for a genetic marker by RT-qPCR, or
CR with negativity by multiparameter flow cytometry
Complete remission (CR) • Bone marrow blasts <5%
• Absence of circulating blasts and blasts with Auer rods
• Absence of extramedullary disease
• ANC ≥ 1
0 9 109/l (1000/µl)
• Platelet count ≥100 9 109/l (100 000/µl)

CR with incomplete haematologic recovery (CRi) All CR criteria except for residual neutropenia (<1
0 9 109/l [1000/µl]) or
thrombocytopenia (<100 9 109/l [100 000/µl])
Morphologic leukaemia-free state (MLFS) • Bone marrow blasts <5%
• Absence of blasts with Auer rods
• Absence of extramedullary disease
• No haematologic recovery required

Partial remission (PR) All haematologic criteria of CR; decrease of bone marrow blast percentage
to 5–25%; and decrease of pretreatment bone marrow blast percentage by at least 50%
Treatment failure
Primary refractory disease No CR or CRi after two courses of intensive induction treatment; excluding patients
with death in aplasia or death due to indeterminate cause
Death in aplasia Deaths occurring ≥7 days following completion of initial treatment while cytopenic;
with an aplastic or hypoplastic bone marrow obtained within 7 days of death, without
evidence of persistent leukaemia
Death from indeterminate cause Deaths occurring before completion of therapy, or <7 days following its completion; or
deaths occurring ≥7 days following completion of initial therapy with no blasts in the
blood, but no bone marrow examination available
Relapse
Haematologic relapse (after CRMRD , CR, CRi) Bone marrow blasts ≥5%; or reappearance of blasts in the blood; or development of
extramedullary disease
Molecular relapse (after CRMRD ) If studied pretreatment, reoccurrence of MRD as assessed by RT-qPCR or by
multiparameter flow cytometry

With minor adjustments republished with permission of American Society of Hematology, from: D€
ohner et al. (2017).

that complicate the use of molecular aberrations for MRD Quantitative PCR represents the gold standard for molec-
monitoring (Schuurhuis et al., 2018): ular assessment of the MRD status. All PCR-based
approaches require the use of aberration- and often also
1 At relapse, some genetic loci are prone to chromosomal
patient-specific assays. For qPCR, which depends on stan-
losses or gains (e.g. FLT3-ITD/TKD mutations, EVI1).
dards for relative quantification, this increases the labour
2 The somatic origin of a given mutation might be unclear,
intensity of the technique. In digital PCR (dPCR), the com-
because of their recurrence in the germline (e.g. RUNX1,
partmentalisation of the reaction volume permits a binary
CEBPA).
fluorescence signal read out (signal or no signal) after PCR
3 Clonal haematopoiesis of indeterminate potential (CHIP)-
and thus absolute quantification (Sykes et al., 1992; Vogel-
associated gene mutations often persist (Busque et al., 2012;
stein & Kinzler, 1999). Compared to qPCR it offers several
Jaiswal et al., 2014; Xie et al., 2014; Genovese et al., 2015;
advantages: in addition to an improved signal-to-noise ratio
Steensma et al., 2015), even during CR (DNMT3A, ASXL1,
and the independence from standards, potentially present
TET2) (Jongen-Lavrencic et al., 2018; H€ ollein et al., 2018b)
PCR inhibitors and PCR efficiency have a much smaller
Only the RUNX1-RUNX1T1, PML-RARA and CBFB- influence on the measurement (Huggett et al., 2015; Quan
MYH11 rearrangements as well as the NPM1 mutation are et al., 2018). Based on its properties, dPCR is a suitable and
currently fully recommended as sole MRD markers by the feasible method for sensitive MRD monitoring (Cilloni et al.,
ELN (Schuurhuis et al., 2018). 2019) and is likely to prove its value in the clinical setting.

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 43
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

Next-generation sequencing, with sensitivities of approx. experienced a paradigm change from phenotype to genotype
1% mutational load (Schuurhuis et al., 2018), is a valuable tool and an ever-increasing importance of multidisciplinary diag-
to identify potential MRD markers at diagnosis, but not yet for nostics. Today, FISH and molecular techniques are indis-
their monitoring. However, efforts have been made to increase pensable. All diagnostic disciplines are needed to inform
sensitivity by optimising experimental parameters and bioin- and/or assist classification, prognostication, therapeutic deci-
formatic algorithms (Thol et al., 2018). This will allow for reli- sion and monitoring of residual disease. Table VI gives an
able NGS-based MRD quantification in the future. overview of the respective essential diagnostic tool set.
Already, there are a few examples of how MRD status Data obtained by integration of all of the different diag-
informs therapeutic decisions. Pre-emptive therapy for APL nostic disciplines are immense, especially since the introduc-
patients with MRD positivity strongly reduced relapse risk tion of NGS. All guidelines published need a minimum of
(Grimwade et al., 2009). Patients who underwent allo-SCT diagnostic information not only at diagnosis but also for
have been shown to benefit from MRD monitoring and pre- MRD measurement and at relapse. Today, this includes
emptive therapy in the case of a positive MRD status (Schroe- genetic data at all time points. However, up to now, the gen-
der et al., 2013). Moreover, molecular monitoring of patients eral approach was a targeted one, for example screening a
with t(8;21)/RUNX1-RUNX1T1 or with mutated NPM1 after patient for entity-defining genetic aberrations. In the future,
induction and consolidation therapy identified those at high diagnostics will entail data on the global genomic and tran-
risk of relapse and thus beneficiaries of allo-SCT (Zhu et al., scriptomic level, instead of focusing on single aberrations.
2013) and/or high-dose cytarabine (Kr€ onke et al., 2011; Why are next steps needed and how can they be implemented?
Shayegi et al., 2013; Ivey et al., 2016; H€
ollein et al., 2018a).
1 Capabilities of NGS increase and prices will go down.
Currently, MRD status is monitored either by MFC or
2 Turnaround time of NGS-based methods is below
molecular approaches. A recent study, however, found that
seven days and can already influence first-line treatment.
combining both methods strongly improved prediction of
3 MRD diagnostics will be possible and can follow individ-
relapse risk. When residual disease was detected using either
ual findings in nearly all patients.
method, relapse risk was ~50%. When MRD positivity was
4 WGS, WES and WTS will be feasible for routine use in
ascertained by both methods, the relapse risk was 73
3%
the next five years and will outperform methods such as
(Jongen-Lavrencic et al., 2018). This highlights the comple-
chromosome banding analysis, FISH, array comparative
mentarity of both methods and strongly argues for synergis-
genomic hybridisation (CGH) or panel testing using NGS
tic multidisciplinary diagnostics in MRD detection.
at diagnosis and at relapse to define the complete land-
In addition, the study showed that all identified mutations
scape of AML and foster personalised treatment.
were suitable MRD markers, with the exception of CHIP-asso-
ciated genes: DNMT3A, TET2 and ASXL1. Molecular MRD Depending on the respective reimbursement structures
markers were identified by the authors at diagnosis using a and countries, the costs for methods such as cytogenetics,
panel of 54 genes associated with myeloid neoplasms (Jongen- FISH, and especially molecular testing differ over a broad
Lavrencic et al., 2018). If this was to be validated in broader range. It is beyond the intention of this review to discuss this
prospective studies, molecular MRD detection for almost every in detail. However, in the future methods like WGS and
AML patient would be feasible, since 96% of patients carry at WTS will challenge the gold standards from today not only
least one driver mutation (Papaemmanuil et al., 2016). with respect to reproducibility and sensitivity but also with
In order to firmly establish MRD diagnostics in AML, stan- respect to turnaround times and costs. Parallel studies are
dardisation of molecular and immunophenotypic MRD assess- needed to define the respective advantages and drawbacks.
ments is an absolute must. This would allow for the definition All data available now, but even more data from WGS/WES
of valid and reliable response criteria, for the identification of and transcriptomics represent a great challenge, as the
clinically meaningful MRD thresholds and for determination obtained information for one patient, let alone the genetic and
of the clinical utility of MRD for different AML subtypes. First gene expression landscape of AML in general, will be beyond
standardisation attempts are under way, such as the UK human comprehension. However, it also provides us with the
NEQAS pilot project for minimal residual disease evaluation unique opportunity to translate the advance in knowledge into
in AML by flow cytometry (http://www.ukneqasli.co.uk/eqa-pt-
programmes/flow-cytometry-programmes/minimal-residual-dise Table VI. Mandatory diagnostic techniques in 2020.
ase-for-aml-by-flow-cytometry-pilot-not-accredited/)
Choice Measurable
of residual
Today’s needs and future directions for Diagnostic technique Diagnosis Prognosis therapy disease
integrated diagnostics in AML Cytomorphology X X
Immunophenotyping X X X
Just 15 years ago the diagnostic state-of-the-art in AML
Cytogenetics X X X
included only cytomorphology, immunophenotyping and
Molecular genetics X X X X
metaphase chromosome banding analysis. We have since

44 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
 Review

Fig 1. Precision medicine will be driven by multidisciplinary diagnostics and targetable genetic aberrations. Icons depict the diagnostic disciplines
of (A) cytomorphology, (B) chromosome banding analysis, (C) FISH, (D) molecular genetics and (E) immunophenotyping. Only by combining
findings of all diagnostic techniques a comprehensive characterisation of the underlying pathobiology can be attained. Mutational profiling plays
a key role in identifying acute myeloid leukaemia drivers and targetable genetic aberrations, while carefully distinguishing between somatic and
germline aberrations. Personalised therapies will significantly contribute to improved outcome. In the future, it will not suffice to describe leukae-
mia at initial diagnosis; instead multidisciplinary diagnostics will be required to monitor disease and response kinetics, clonal dynamics as well as
residual disease iteratively. This ensures that every patient is treated adaptively and in the best possible way. Graphic by Dr. Wencke Walter, MLL
Munich Leukemia Laboratory.

improved classification and prognostication and pave the way pathobiology attained by integrated diagnostics and continu-
for precision medicine in AML in the truest sense of the term. ous monitoring (see also Fig 1).
To reach this goal, physicians and scientists will need
assistance to make sense of the data, identify clinically mean-
Conclusion
ingful disease patterns as well as leukaemia-driving or -defin-
ing events and aid therapeutic decisions. This all cannot be We envision a future where artificial intelligence with oversight
done without streamlined workflows, automation of sample by trained haematologists and scientists will find the best ther-
handling, databases, and a complex armamentarium of soft- apeutic algorithm and drugs for every patient — be it the par-
ware tools for interpretation. Said assistance will for sure ticipation in one or several (basket) studies, the choice for or
include artificial intelligence and cloud computing. It will against allogeneic SCT, the treatment with targeted therapeu-
bring up new challenges and solutions for data security and tics or the ideal sequence of therapeutic regimens. At the same
interpretation will also lead to ethical discussions how to time, special consideration must not only be given to a
handle germline findings. These data then need to be trans- patient’s genetic setup but also to treatment guidelines as well
lated into reports understandable for doctors and patients. as to known and validated interdependencies between genetic
As artificial intelligence using deep learning algorithms is aberrations and the influence of individual genetic aberrations
on its way for routine applications also in diagnostics, several or aberration patterns on drug sensitivity. Furthermore, AI can
interesting approaches are ongoing, including AI-based image assist the detection of MRD and relapse prediction.
analysis of blood and bone marrow smears or the drawing of In perspective, all these goals will only be achievable if we
karyograms based on captured metaphases. So far none of use and integrate all diagnostic and technological tools from
these approaches have been used for routine diagnostics and today, combine their information and in parallel test new
they should be studied in prospective trials in comparison to options such as WGS, WTS and the implementation of AI
gold standard approaches. and automation into our future thinking and doing.
In the next five years the initial workup will not change: Never before were the options in AML diagnostics so close
the diagnostic basis is and will be determined by cytomor- to meet the needs, and cure for every single patient might be
phology, immunophenotyping and genetic analysis. The ther- possible if we test all new options and challenge all state-of-
apeutic aim is and ever will be to provide the best possible the-art workflows without prejudice.
treatment for every patient, possibly a cure, while avoiding
unnecessary risks and toxicities. However, owing to the
Conflict of interest
heterogeneity of the disease, in the end the ideal approach to
reach these ambiguous therapeutic aims will differ for every TH is part owner and founder of the MLL Munich Leukemia
patient. This requires a deep understanding of the individual Laboratory. IS is employed by the MLL Munich Leukemia

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 45
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

Laboratory. The MLL offers diagnostic services for leukae- research studies based on a combination of methods for rou-
mias and lymphomas, including cytomorphology, cytochem- tine use and also including whole genome sequencing and
istry, immunophenotyping, cytogenetics, FISH, and a broad whole transcriptome sequencing.
spectrum of molecular assays. In addition, MLL runs several

References (1985) Proposed revised criteria for the classifica- expression distinct from those with transloca-
tion of acute myeloid leukemia. A report of the tion-mediated fusion events. Genes, Chromo-
Abelson, S., Collord, G., Ng, S.W.K., Weissbrod, French-American-British Cooperative Group. somes and Cancer, 37, 149–158.
O., Mendelson, C.N., Niemeyer, E., Barda, N., Annals of Internal Medicine, 103, 620–625. Desai, P., Mencia-Trinchant, N., Savenkov, O.,
Zuzarte, P.C., Heisler, L., Sundaravadanam, Y., Bennett, J.M., Catovsky, D., Daniel, M.T., Flan- Simon, M.S., Cheang, G., Lee, S., Samuel, M.,
Luben, R., Hayat, S., Wang, T.T., Zhao, Z., Cir- drin, G., Galton, D.A., Gralnick, H.R. & Sultan, Ritchie, E.K., Guzman, M.L., Ballman, K.V.,
lan, I., Pugh, T.J., Soave, D., Ng, K., Latimer, C. (1991) Proposal for the recognition of mini- Roboz, G.J. & Hassane, D.C. (2018) Somatic
C., Hardy, C., Raine, K., Jones, D., Hoult, D., mally differentiated acute myeloid leukaemia mutations precede acute myeloid leukemia years
Britten, A., McPherson, J.D., Johansson, M., (AML-MO). British Journal of Haematology, 78, before diagnosis. Nature Medicine, 24, 1015–
Mbabaali, F., Eagles, J., Miller, J.K., Pasternack, 325–329. 1023.
D., Timms, L., Krzyzanowski, P., Awadalla, P., Bullinger, L., D€ ohner, K., Bair, E., Fr€ ohling, S., D€ohner, H., Estey, E., Grimwade, D., Amadori, S.,
Costa, R., Segal, E., Bratman, S.V., Beer, P., Beh- Schlenk, R.F., Tibshirani, R., D€ ohner, H. & Pol- Appelbaum, F.R., Buchner, T., Dombret, H.,
jati, S., Martincorena, I., Wang, J.C.Y., Bowles, lack, J.R. (2004) Use of gene-expression profil- Ebert, B.L., Fenaux, P., Larson, R.A., Levine,
K.M., Quiros, J.R., Karakatsani, A., La, V.C., ing to identify prognostic subclasses in adult R.L., Lo-Coco, F., Naoe, T., Niederwieser, D.,
Trichopoulou, A., Salamanca-Fernandez, E., acute myeloid leukemia. The New England Jour- Ossenkoppele, G.J., Sanz, M., Sierra, J., Tallman,
Huerta, J.M., Barricarte, A., Travis, R.C., nal of Medicine, 350, 1605–1616. M.S., Tien, H.F., Wei, A.H., L€ owenberg, B. &
Tumino, R., Masala, G., Boeing, H., Panico, S., Bullinger, L., Ehrich, M., D€ ohner, K., Schlenk, Bloomfield, C.D. (2017) Diagnosis and manage-
Kaaks, R., Kramer, A., Sieri, S., Riboli, E., R.F., D€ ohner, H., Nelson, M.R. & van den ment of AML in adults: 2017 ELN recommen-
Vineis, P., Foll, M., McKay, J., Polidoro, S., Sala, Boom, D. (2010) Quantitative DNA methylation dations from an international expert panel.
N., Khaw, K.T., Vermeulen, R., Campbell, P.J., predicts survival in adult acute myeloid leuke- Blood, 129, 424–447.
Papaemmanuil, E., Minden, M.D., Tanay, A., mia. Blood, 115, 636–642. D€ohner, H., Dolnik, A., Tang, L., Seymour, J.F.,
Balicer, R.D., Wareham, N.J., Gerstung, M., Busque, L., Patel, J.P., Figueroa, M.E., Vasanthaku- Minden, M.D., Stone, R.M., Del Castillo, T.B.,
Dick, J.E., Brennan, P., Vassiliou, G.S. & Shlush, mar, A., Provost, S., Hamilou, Z., Mollica, L., Al-Ali, H.K., Santini, V., Vyas, P., Beach, C.L.,
L.I. (2018) Prediction of acute myeloid leukae- Li, J., Viale, A., Heguy, A., Hassimi, M., Socci, MacBeth, K.J., Skikne, B.S., Songer, S., Tu, N.,
mia risk in healthy individuals. Nature, 559, N., Bhatt, P.K., Gonen, M., Mason, C.E., Mel- Bullinger, L. & Dombret, H. (2018) Cytogenetics
400–404. nick, A., Godley, L.A., Brennan, C.W., Abdel- and gene mutations influence survival in older
Angenendt, L., R€ ollig, C., Montesinos, P., Marti- Wahab, O. & Levine, R.L. (2012) Recurrent patients with acute myeloid leukemia treated
nez-Cuadron, D., Barragan, E., Garcia, R., somatic TET2 mutations in normal elderly indi- with azacitidine or conventional care. Leukemia,
Botella, C., Martinez, P., Ravandi, F., Kadia, T., viduals with clonal hematopoiesis. Nature Genet- 32, 2546–2557.
Kantarjian, H.M., Cortes, J., Juliusson, G., ics, 44, 1179–1181. Falini, B., Macijewski, K., Weiss, T., Bacher, U.,
Lazarevic, V., Hoglund, M., Lehmann, S., Cancer Genome Atlas Research Network (2013) Schnittger, S., Kern, W., Kohlmann, A., Klein,
Recher, C., Pigneux, A., Bertoli, S., Dumas, P.Y., Genomic and epigenomic landscapes of adult de H.U., Vignetti, M., Piciocchi, A., Fazi, P., Mar-
Dombret, H., Preudhomme, C., Micol, J.B., novo acute myeloid leukemia. The New England telli, M.P., Vitale, A., Pileri, S., Miesner, M.,
Terre, C., Racil, Z., Novak, J., Zak, P., Wei, A.H., Journal of Medicine, 368, 2059–2074. Santucci, A., Haferlach, C., Mandelli, F. &
Tiong, I.S., Wall, M., Estey, E., Shaw, C., Exeler, Cappelli, L.V., Meggendorfer, M., Dicker, F., Jero- Haferlach, T. (2010) Multilineage dysplasia has
R., Wagenfuhr, L., St€ olzel, F., Thiede, C., Stelljes, min, S., Hutter, S., Kern, W., Haferlach, T., no impact on biologic, clinicopathologic, and
M., Lenz, G., Mikesch, J.H., Serve, H., Ehninger, Haferlach, C. & H€ ollein, A. (2019) DNMT3A prognostic features of AML with mutated nucle-
G., Berdel, W.E., Kramer, M., Krug, U. & Schlie- mutations are over-represented in young adults ophosmin (NPM1). Blood, 115, 3776–3786.
mann, C. (2019) Chromosomal abnormalities with NPM1 mutated AML and prompt a dis- Figueroa, M.E., Lugthart, S., Li, Y., Erpelinck-Ver-
and prognosis in NPM1-mutated acute myeloid tinct co-mutational pattern. Leukemia, 33, 2741– schueren, C.A., Deng, X., Christos, P.J., Schi-
leukemia: a pooled analysis of individual patient 2746. fano, E., Booth, J., van Putten, W., Skrabanek,
data from nine international cohorts. Journal of Catovsky, D., Matutes, E., Buccheri, V., Shetty, V., L., Campagne, F., Mazumdar, M., Greally, J.M.,
Clinical Oncology, 21, 2632–2642. Hanslip, J., Yoshida, N. & Morilla, R. (1991) A Valk, P.J., L€
owenberg, B., Delwel, R. & Melnick,
Bacher, U., Schnittger, S., Macijewski, K., Gross- classification of acute leukaemia for the 1990s. A. (2010) DNA methylation signatures identify
mann, V., Kohlmann, A., Alpermann, T., Annals of Hematology, 62, 16–21. biologically distinct subtypes in acute myeloid
Kowarsch, A., Nadarajah, N., Kern, W., Hafer- Cilloni, D., Petiti, J., Rosso, V., Andreani, G., Dra- leukemia. Cancer Cell, 17, 13–27.
lach, C. & Haferlach, T. (2012) Multilineage gani, M., Fava, C. & Saglio, G. (2019) Digital Fircanis, S., Merriam, P., Khan, N. & Castillo, J.J.
dysplasia does not influence prognosis in PCR in myeloid malignancies: ready to replace (2014) The relation between cigarette smoking
CEBPA-mutated AML, supporting the WHO quantitative PCR? International Journal of Molec- and risk of acute myeloid leukemia: an updated
proposal to classify these patients as a unique ular Sciences, 20, 2249. meta-analysis of epidemiological studies. Ameri-
entity. Blood, 119, 4719–4722. De Kouchkovsky, I. & Abdul-Hay, M. (2016) can Journal of Hematology, 89, E125–E132.
Bennett, J.M., Catovsky, D., Daniel, M.T., Flan- Acute myeloid leukemia: a comprehensive Genovese, G., Jaiswal, S., Ebert, B.L. & McCarroll,
drin, G., Galton, D.A., Gralnick, H.R. & Sultan, review and 2016 update. Blood Cancer Journal, S.A. (2015) Clonal hematopoiesis and blood-
C. (1976) Proposals for the classification of the 6, e441. cancer risk. The New England Journal of Medi-
acute leukaemias. French-American-British Debernardi, S., Lillington, D.M., Chaplin, T., Tom- cine, 372, 1071–1072.
(FAB) co-operative group. British Journal of linson, S., Amess, J., Rohatiner, A., Lister, T.A. Gerstung, M., Papaemmanuil, E., Martincorena, I.,
Haematology, 33, 451–458. & Young, B.D. (2003) Genome-wide analysis of Bullinger, L., Gaidzik, V.I., Paschka, P., Heuser,
Bennett, J.M., Catovsky, D., Daniel, M.T., Flandrin, acute myeloid leukemia with normal karyotype M., Thol, F., Bolli, N., Ganly, P., Ganser, A.,
G., Galton, D.A., Gralnick, H.R. & Sultan, C. reveals a unique pattern of homeobox gene McDermott, U., D€ ohner, K., Schlenk, R.F.,

46 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
 Review

D€ ohner, H. & Campbell, P.J. (2017) Precision B. & Ganser, A. (2005) Gene-expression profiles Toal, M., Germing, U., Haase, D., Mergenthaler,
oncology for acute myeloid leukemia using a and their association with drug resistance in H.G., Krauter, J., Ganser, A., Gohring, G., Sch-
knowledge bank approach. Nature Genetics, 49, adult acute myeloid leukemia. Haematologica, legelberger, B., D€ ohner, H. & D€ ohner, K. (2011)
332–340. 90, 1484–1492. Monitoring of minimal residual disease in
Golub, T.R., Slonim, D.K., Tamayo, P., Huard, C., H€ollein, A., Jeromin, S., Meggendorfer, M., Fasan, NPM1-mutated acute myeloid leukemia: a study
Gaasenbeek, M., Mesirov, J.P., Coller, H., Loh, A., Nadarajah, N., Kern, W., Haferlach, C. & from the German-Austrian acute myeloid leuke-
M.L., Downing, J.R., Caligiuri, M.A., Bloom- Haferlach, T. (2018a) Minimal residual disease mia study group. Journal of Clinical Oncology,
field, C.D. & Lander, E.S. (1999) Molecular clas- (MRD) monitoring and mutational landscape in 29, 2709–2716.
sification of cancer: class discovers and class AML with RUNX1-RUNX1T1: a study on 134 Orazi, A. (2007) Histopathology in the diagnosis and
prediction by gene expression monitoring. patients. Leukemia, 32, 2270–2274. classification of acute myeloid leukemia, myelodys-
Science, 286, 531–537. H€ollein, A., Meggendorfer, M., Dicker, F., Jeromin, plastic syndromes, and myelodysplastic/myelopro-
Grimwade, D., Jovanovic, J.V., Hills, R.K., Nugent, S., Nadarajah, N., Kern, W., Haferlach, C. & liferative diseases. Pathobiology, 74, 97–114.
E.A., Patel, Y., Flora, R., Diverio, D., Jones, K., Haferlach, T. (2018b) NPM1 mutated AML can Papaemmanuil, E., Gerstung, M., Bullinger, L.,
Aslett, H., Batson, E., Rennie, K., Angell, R., relapse with wild-type NPM1: persistent clonal Gaidzik, V.I., Paschka, P., Roberts, N.D., Potter,
Clark, R.E., Solomon, E., Lo-Coco, F., Wheatley, hematopoiesis can drive relapse. Blood Advances, N.E., Heuser, M., Thol, F., Bolli, N., Gundem,
K. & Burnett, A.K. (2009) Prospective minimal 2, 3118–3125. G., Van Loo, L.P., Martincorena, I., Ganly, P.,
residual disease monitoring to predict relapse of Huang, R., Liao, X. & Li, Q. (2017) Identification Mudie, L., McLaren, S., O’Meara, S., Raine, K.,
acute promyelocytic leukemia and to direct pre- and validation of potential prognostic gene Jones, D.R., Teague, J.W., Butler, A.P., Greaves,
emptive arsenic trioxide therapy. Journal of Clin- biomarkers for predicting survival in patients M.F., Ganser, A., D€ ohner, K., Schlenk, R.F.,
ical Oncology, 27, 3650–3658. with acute myeloid leukemia. OncoTargets and D€ ohner, H. & Campbell, P.J. (2016) Genomic
Grimwade, D., Ivey, A. & Huntly, B.J. (2016) Therapy, 10, 5243–5254. classification and prognosis in acute myeloid
Molecular landscape of acute myeloid leukemia Huggett, J.F., Cowen, S. & Foy, C.A. (2015) Consid- leukemia. The New England Journal of Medicine,
in younger adults and its clinical relevance. erations for digital PCR as an accurate molecular 374, 2209–2221.
Blood, 127, 29–41. diagnostic tool. Clinical Chemistry, 61, 79–88. Quan, P.L., Sauzade, M. & Brouzes, E. (2018) dPCR:
Haferlach, T., Schoch, C., L€ offler, H., Gassmann, Ivey, A., Hills, R.K., Simpson, M.A., Jovanovic, a technology review. Sensors (Basel), 18, 1271.
W., Kern, W., Schnittger, S., Fonatsch, C., Lud- J.V., Gilkes, A., Grech, A., Patel, Y., Bhudia, N., Radmacher, M.D., Marcucci, G., Ruppert, A.S.,
wig, W.D., Wuchter, C., Schlegelberger, B., Farah, H., Mason, J., Wall, K., Akiki, S., Grif- Mrozek, K., Whitman, S.P., Vardiman, J.W.,
Staib, P., Reichle, A., Kubica, U., Eimermacher, fiths, M., Solomon, E., McCaughan, F., Linch, Paschka, P., Vukosavljevic, T., Baldus, C.D.,
H., Balleisen, L., Gr€uneisen, A., Haase, D., Aul, D.C., Gale, R.E., Vyas, P., Freeman, S.D., Rus- Kolitz, J.E., Caligiuri, M.A., Larson, R.A. &
C., Karow, J., Lengfelder, E., W€ ormann, B., Hei- sell, N., Burnett, A.K. & Grimwade, D. (2016) Bloomfield, C.D. (2006) Independent confirma-
necke, A., Sauerland, M.C., B€ uchner, T. & Hid- Assessment of minimal residual disease in stan- tion of a prognostic gene-expression signature
demann, W. (2003) Morphologic dysplasia in de dard-risk AML. The New England Journal of in adult acute myeloid leukemia with a normal
novo acute myeloid leukemia (AML) is related Medicine, 374, 422–433. karyotype: a Cancer and Leukemia Group B
to unfavorable cytogenetics but has no indepen- Jaffe, E.S., Harris, N.L. & Vardiman, J.W. (2001) study. Blood, 108, 1677–1683.
dent prognostic relevance under the conditions WHO Classification of Tumours. Pathology and Raponi, M., Lancet, J.E., Fan, H., Dossey, L., Lee,
of intensive induction therapy: results of a mul- Genetics of Tumours of Haematopoietic and G., Gojo, I., Feldman, E.J., Gotlib, J., Morris,
tiparameter analysis from the German AML Lymphoid Tissues, 3rd edn. IARC Press, Lyon. L.E., Greenberg, P.L., Wright, J.J., Harousseau,
Cooperative Group studies. Journal of Clinical Jaiswal, S., Fontanillas, P., Flannick, J., Manning, J.L., L€owenberg, B., Stone, R.M., De Porre, P.,
Oncology, 21, 256–265. A., Grauman, P.V., Mar, B.G., Lindsley, R.C., Wang, Y. & Karp, J.E. (2008) A 2-gene classifier
Haferlach, T., Kohlmann, A., Wieczorek, L., Basso, Mermel, C.H., Burtt, N., Chavez, A., Higgins, for predicting response to the farnesyltransferase
G., Kronnie, G.T., Bene, M.C., de Vos, J., Her- J.M., Moltchanov, V., Kuo, F.C., Kluk, M.J., inhibitor tipifarnib in acute myeloid leukemia.
nandez, J.M., Hofmann, W.K., Mills, K.I., Gilkes, Henderson, B., Kinnunen, L., Koistinen, H.A., Blood, 111, 2589–2596.
A., Chiaretti, S., Shurtleff, S.A., Kipps, T.J., Ras- Ladenvall, C., Getz, G., Correa, A., Banahan, Ravandi, F., Walter, R.B. & Freeman, S.D. (2018)
senti, L.Z., Yeoh, A.E., Papenhausen, P.R., Liu, B.F., Gabriel, S., Kathiresan, S., Stringham, Evaluating measurable residual disease in acute
W.M., Williams, P.M. & Foa, R. (2010) Clinical H.M., McCarthy, M.I., Boehnke, M., Tuomile- myeloid leukemia. Blood Advances, 2, 1356–
utility of microarray-based gene expression pro- hto, J., Haiman, C., Groop, L., Atzmon, G., Wil- 1366.
filing in the diagnosis and subclassification of son, J.G., Neuberg, D., Altshuler, D. & Ebert, Schoch, C., Kohlmann, A., Schnittger, S., Brors, B.,
leukemia: report from the International Microar- B.L. (2014) Age-related clonal hematopoiesis Dugas, M., Mergenthaler, S., Kern, W., Hidde-
ray Innovations in Leukemia Study Group. Jour- associated with adverse outcomes. The New Eng- mann, W., Eils, R. & Haferlach, T. (2002) Acute
nal of Clinical Oncology, 28, 2529–2537. land Journal of Medicine, 371, 2488–2498. myeloid leukemias with reciprocal rearrange-
Herold, T., Jurinovic, V., Batcha, A.M.N., Bamo- Jongen-Lavrencic, M., Grob, T., Hanekamp, D., ments can be distinguished by specific gene
poulos, S.A., Rothenberg-Thurley, M., Ksienzyk, Kavelaars, F.G., Al, H.A., Zeilemaker, A., Erpe- expression profiles. Proceedings of the National
B., Hartmann, L., Greif, P.A., Phillippou-Mas- linck-Verschueren, C.A.J., Gradowska, P.L., Mei- Academy of Sciences of the United States of Amer-
sier, J., Krebs, S., Blum, H., Amler, S., Schnei- jer, R., Cloos, J., Biemond, B.J., Graux, C., van ica, 99, 10008–10013.
der, S., Konstandin, N., Sauerland, M.C., Marwijk, K.M., Manz, M.G., Pabst, T., Passweg, Schroeder, T., Czibere, A., Platzbecker, U., Bug,
Gorlich, D., Berdel, W.E., Wormann, B.J., Tis- J.R., Havelange, V., Ossenkoppele, G.J., Sanders, G., Uharek, L., Luft, T., Giagounidis, A., Zoh-
cher, J., Subklewe, M., Bohlander, S.K., Braess, M.A., Schuurhuis, G.J., L€ owenberg, B. & Valk, ren, F., Bruns, I., Wolschke, C., Rieger, K., Fenk,
J., Hiddemann, W., Metzeler, K.H., Mansmann, P.J.M. (2018) Molecular minimal residual dis- R., Germing, U., Haas, R., Kroger, N. & Kobbe,
U. & Spiekermann, K. (2018) A 29-gene and ease in acute myeloid leukemia. The New Eng- G. (2013) Azacitidine and donor lymphocyte
cytogenetic score for the prediction of resistance land Journal of Medicine, 378, 1189–1199. infusions as first salvage therapy for relapse of
to induction treatment in acute myeloid leuke- Kr€onke, J., Schlenk, R.F., Jensen, K.O., Tschurtz, AML or MDS after allogeneic stem cell trans-
mia. Haematologica, 103, 456–465. F., Corbacioglu, A., Gaidzik, V.I., Paschka, P., plantation. Leukemia, 27, 1229–1235.
Heuser, M., Wingen, L.U., Steinemann, D., Cario, Onken, S., Eiwen, K., Habdank, M., Spath, D., Schuurhuis, G.J., Heuser, M., Freeman, S., Bene,
G., von Neuhoff, N., Tauscher, M., Bullinger, L., Lubbert, M., Wattad, M., Kindler, T., Salih, M.C., Buccisano, F., Cloos, J., Grimwade, D.,
Krauter, J., Heil, G., D€ohner, H., Schlegelberger, H.R., Held, G., Nachbaur, D., von Lilienfeld- Haferlach, T., Hills, R.K., Hourigan, C.S.,

ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology 47
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48
Review

Jorgensen, J.L., Kern, W., Lacombe, F., Maurillo, De Lima, M., Fathi, A.T., Fiorella, M., Foran, landscape of acute myeloid leukaemia. Nature,
L., Preudhomme, C., van der Reijden, B.A., J.M., Hall, A.C., Jacoby, M., Lancet, J., LeBlanc, 562, 526–531.
Thiede, C., Venditti, A., Vyas, P., Wood, B.L., T.W., Mannis, G., Marcucci, G., Martin, M.G., Valk, P.J., Verhaak, R.G., Beijen, M.A., Erpelinck,
Walter, R.B., D€ ohner, K., Roboz, G.J. & Mims, A., O’Donnell, M.R., Olin, R., Peker, D., C.A., van Waalwijk, Barjesteh, van Doorn-Khos-
Ossenkoppele, G.J. (2018) Minimal/measurable Perl, A., Pollyea, D.A., Pratz, K., Prebet, T., rovani, S., Boer, J.M., Beverloo, H.B., Moor-
residual disease in AML: consensus document Ravandi, F., Shami, P.J., Stone, R.M., Strickland, house, M.J., van der Spek, P.J., L€owenberg, B. &
from ELN MRD Working Party. Blood, 131, S.A., Wieduwilt, M., Gregory, K.M., Hammond, Delwel, R. (2004) Prognostically useful gene-ex-
1275–1291. L. & Ogba, N. (2019) Acute myeloid leukemia, pression profiles in acute myeloid leukemia. The
Second MIC Cooperative Study Group (1988) version 3.2019, NCCN clinical practice guideli- New England Journal of Medicine, 350, 1617–
Morphologic, immunologic and cytogenetic nes in oncology. Journal of the National Compre- 1628.
(MIC) working classification of the acute mye- hensive Cancer Network, 17, 721–749. Visani, G., Ferrara, F., Di Raimondo, F., Loscocco,
loid leukaemias. British Journal of Haematology, Thol, F., Gabdoulline, R., Liebich, A., Klement, P., F., Fuligni, F., Paolini, S., Zammit, V., Spina, E.,
68, 487–494. Schiller, J., Kandziora, C., Hambach, L., Stadler, Rocchi, M., Visani, A., Piccaluga, P.P. & Isidori,
Shayegi, N., Kramer, M., Bornhauser, M., Schaich, M., Koenecke, C., Flintrop, M., Pankratz, M., A. (2017) Low-dose lenalidomide plus cytara-
M., Schetelig, J., Platzbecker, U., R€ ollig, C., Hei- Wichmann, M., Neziri, B., Buttner, K., Heida, bine in very elderly, unfit acute myeloid leuke-
derich, C., Landt, O., Ehninger, G. & Thiede, C. B., Klesse, S., Chaturvedi, A., Kloos, A., Gohr- mia patients: Final result of a phase II study.
(2013) The level of residual disease based on ing, G., Schlegelberger, B., Gaidzik, V.I., Bullin- Leukemia Research, 62, 77–83.
mutant NPM1 is an independent prognostic fac- ger, L., Fiedler, W., Heim, A., Hamwi, I., Eder, Visani, G., Loscocco, F., Isidori, A. & Piccaluga,
tor for relapse and survival in AML. Blood, 122, M., Krauter, J., Schlenk, R.F., Paschka, P., P.P. (2018) Genetic profiling in acute myeloid
83–92. D€ohner, K., D€ ohner, H., Ganser, A. & Heuser, leukemia: a path to predicting treatment out-
Shreve, J., Meggendorfer, M., Awada, H., Mukher- M. (2018) Measurable residual disease monitor- come. Expert Review of Hematology, 11, 455–
jee, S., Walter, W., Hutter, S., Makhoul, A., ing by NGS before allogeneic hematopoietic cell 461.
Beau Hilton, C., Radakovich, N., Nagata, Y., transplantation in AML. Blood, 132, 1703–1713. Vogelstein, B. & Kinzler, K.W. (1999) Digital PCR.
Rouphail, Y., Adema, V., Kerr, C.M., Patel, B.J., Tyner, J.W., Tognon, C.E., Bottomly, D., Wilmot, Proceedings of the National Academy of Sciences
Kuzmanovic, T., Maciejewski, J.P., Haferlach, C., B., Kurtz, S.E., Savage, S.L., Long, N., Schultz, of the United States of America, 96, 9236–9241.
Sekeres, M.A., Haferlach, T. & Nazha, A. (2019) A.R., Traer, E., Abel, M., Agarwal, A., Blucher, Wouters, B.J. & Delwel, R. (2016) Epigenetics and
A personalized prediction model to risk stratify A., Borate, U., Bryant, J., Burke, R., Carlos, A., approaches to targeted epigenetic therapy in
patients with acute myeloid leukemia (AML) Carpenter, R., Carroll, J., Chang, B.H., Coblentz, acute myeloid leukemia. Blood, 127, 42–52.
using artificial intelligence. Blood, 134, 2091. C., d’Almeida, A., Cook, R., Danilov, A., Dao, Xie, M., Lu, C., Wang, J., McLellan, M.D., John-
Steensma, D.P., Bejar, R., Jaiswal, S., Lindsley, K.T., Degnin, M., Devine, D., Dibb, J., Edwards, son, K.J., Wendl, M.C., McMichael, J.F., Sch-
R.C., Sekeres, M.A., Hasserjian, R.P. & Ebert, D.K., Eide, C.A., English, I., Glover, J., Henson, midt, H.K., Yellapantula, V., Miller, C.A.,
B.L. (2015) Clonal hematopoiesis of indetermi- R., Ho, H., Jemal, A., Johnson, K., Johnson, R., Ozenberger, B.A., Welch, J.S., Link, D.C., Wal-
nate potential and its distinction from Junio, B., Kaempf, A., Leonard, J., Lin, C., Liu, ter, M.J., Mardis, E.R., DiPersio, J.F., Chen, F.,
myelodysplastic syndromes. Blood, 126, 9–16. S.Q., Lo, P., Loriaux, M.M., Luty, S., Macey, T., Wilson, R.K., Ley, T.J. & Ding, L. (2014) Age-
Swerdlow, S.H., Campo, E., Harris, N.L., Jaffe, MacManiman, J., Martinez, J., Mori, M., Nel- related mutations associated with clonal
E.S., Pileri, S.A., Stein, H., Thiele, J., Arber, son, D., Nichols, C., Peters, J., Ramsdill, J., hematopoietic expansion and malignancies. Nat-
D.A., Hasserjian, R.P., LeBeau, M.M., Orazi, A. Rofelty, A., Schuff, R., Searles, R., Segerdell, E., ure Medicine, 20, 1472–1478.
& Siebert, R. (2017) WHO Classification of Smith, R.L., Spurgeon, S.E., Sweeney, T., Thapa, Zhu, H.H., Zhang, X.H., Qin, Y.Z., Liu, D.H.,
Tumours of Haematopoietic and Lymphoid Tis- A., Visser, C., Wagner, J., Watanabe-Smith, K., Jiang, H., Chen, H., Jiang, Q., Xu, L.P., Lu, J.,
sues, 4th edn. International Agency for Research Werth, K., Wolf, J., White, L., Yates, A., Zhang, Han, W., Bao, L., Wang, Y., Chen, Y.H., Wang,
on Cancer (IARC), Lyon. H., Cogle, C.R., Collins, R.H., Connolly, D.C., J.Z., Wang, F.R., Lai, Y.Y., Chai, J.Y., Wang,
Sykes, P.J., Neoh, S.H., Brisco, M.J., Hughes, E., Deininger, M.W., Drusbosky, L., Hourigan, C.S., L.R., Liu, Y.R., Liu, K.Y., Jiang, B. & Huang,
Condon, J. & Morley, A.A. (1992) Quantitation Jordan, C.T., Kropf, P., Lin, T.L., Martinez, X.J. (2013) MRD-directed risk stratification
of targets for PCR by use of limiting dilution. M.E., Medeiros, B.C., Pallapati, R.R., Pollyea, treatment may improve outcomes of t(8;21)
BioTechniques, 13, 444–449. D.A., Swords, R.T., Watts, J.M., Weir, S.J., AML in the first complete remission: results
Tallman, M.S., Wang, E.S., Altman, J.K., Appel- Wiest, D.L., Winters, R.M., McWeeney, S.K. & from the AML05 multicenter trial. Blood, 121,
baum, F.R., Bhatt, V.R., Bixby, D., Coutre, S.E., Druker, B.J. (2018) Functional genomic 4056–4062.

48 ª 2019 The Authors. British Journal of Haematology published by British Society for Haematology
and John Wiley & Sons Ltd.. British Journal of Haematology, 2020, 188, 36–48