Leukemia (2014) 28, 1407–1413

& 2014 Macmillan Publishers Limited All rights reserved 0887-6924/14

www.nature.com/leu

REVIEW
An overview on CALR and CSF3R mutations and a proposal
for revision of WHO diagnostic criteria for myeloproliferative
neoplasms
A Tefferi1, J Thiele2, AM Vannucchi3 and T Barbui4

Disease-specific mutations facilitate diagnostic precision and drug target discovery. In myeloproliferative neoplasms (MPN), this is
best exemplified by the chronic myeloid leukemia-associated BCR-ABL1. No other mutation in MPN has thus far shown a similar
degree of diagnostic accuracy or therapeutic relevance. However, JAK2 and KIT mutations are detected in more than 90% of
patients with polycythemia vera and systemic mastocytosis, respectively, and are therefore used as highly sensitive clonal markers
in these diseases. JAK2 and MPL mutations also occur in essential thrombocythemia (ET) and primary myelofibrosis (PMF), but their
diagnostic value is limited by suboptimal sensitivity and specificity. The molecular diagnostic gap in JAK2/MPL-unmutated ET/PMF is
now partially addressed by the recent discovery of calreticulin (CALR) mutations in the majority of such cases. However, bone
marrow morphology remains the central diagnostic platform and is essential for distinguishing ET from prefibrotic PMF and
diagnosing patients those do not express JAK2, MPL or CALR (triple-negative). The year 2013 was also marked by the description of
CSF3R mutations in the majority of patients with chronic neutrophilic leukemia (CNL). Herein, we argue for the inclusion of CALR
and CSF3R mutations in the World Health Organization classification system for ET/PMF and CNL, respectively.

Leukemia (2014) 28, 1407–1413; doi:10.1038/leu.2014.35

Keywords: CALR; CSF3R; myeloproliferative; WHO; neutrophilic

INTRODUCTION examination showing CML-characteristic megakaryocyte

The World Health Organization (WHO) system uses peripheral proliferation.3,5 The morphologic denominator in BCR-ABL1-
blood (PB) counts and smear analysis, bone marrow (BM) negative MPN includes megakaryocytic proliferation associated
morphology, karyotype and molecular genetic tests to classify with variable amounts of granulocytic and erythroid proliferation.
myeloid malignancies into five major categories (Table 1).1,2 In addition, MPN diagnosis requires absence of dysgranulopoiesis
Myeloproliferative neoplasms (MPN) constitute one of these five or dyserythropoiesis. Table 2 summarizes the 2008 WHO
categories and are further classified into eight subcategories, diagnostic criteria for PV, ET and PMF and highlights the role of
including the BCR-ABL1-defined chronic myeloid leukemia (CML) JAK2 mutations and BM morphology.1
and the so-called ‘BCR-ABL1-negative MPN’: polycythemia vera CNL is considered in the presence of marked PB leukocytosis
(PV), essential thrombocythemia (ET) and primary myelofibrosis (X25  10(9)/l), 480% segmented/band neutrophils, o10%
(PMF) (Table 1). Also included under the umbrella of MPN are immature myeloid cells, o1% circulating blasts and absence of
chronic neutrophilic leukemia (CNL), systemic mastocytosis (SM), dysgranulopoiesis or monocytosis. CNL should be distinguished
chronic eosinophilic leukemia, not otherwise specified (CEL-NOS) from other MPN, including CML, PV, ET and PMF, as well as entities
and MPN, unclassifiable (MPN-U).2 In the current perspective, we currently classified under MDS/MPN overlap (Table 3). The latter
focus on two classes of mutations (CALR and CSF3R) and their include atypical CML, BCR-ABL1-negative (aCML), chronic myelo-
potential role in refining the classification and diagnostic criteria monocytic leukemia (CMML), juvenile myelomonocytic leukemia
for ET, PMF and CNL. (JMML) and MDS/MPN, unclassifiable (MDS/MPN-U). CNL diagnosis
also requires exclusion of other causes of nonclonal neutrophilia,
including plasma cell neoplasms, solid tumor, infections and
inflammatory processes.6 BM in CNL is hypercellular and displays
CURRENT DIAGNOSTIC CRITERIA FOR ET, PMF AND CNL increased number and percentage of neutrophils with very high
In the context of MPN, the WHO classification system considers myeloid to erythroid ratio and minimal left shift, myeloid dysplasia
the presence of BCR-ABL1 as being diagnostic for CML and its or reticulin fibrosis.1
absence for all other MPN. However, it has become increasingly
evident that a small fraction of patients with ET or PMF
might harbor low-allelic burden BCR-ABL1 subclones that are of THE CALR STORY
undetermined significance.3,4 Such a scenario does not necessarily Although the molecular diagnostic gap in JAK2V617F-negative PV
imply a diagnosis of CML, which should be confirmed by BM was adequately addressed by the discovery of other JAK2
1
Division of Hematology, Department of Medicine, Mayo Clinic, Rochester, MN, USA; 2Department of Pathology, University of Cologne, Cologne, Germany; 3Department of
Hematology, University of Florence, Florence, Italy and 4Division of Hematology and Research Foundation, Papa Giovanni XXIII Hospital, Bergamo, Italy. Correspondence:
Professor A Tefferi, Division of Hematology, Department of Medicine, Mayo Clinic, 200 First Street, SW, Rochester, MN 55905, USA.
E-mail: tefferi.ayalew@mayo.edu
Received 10 December 2013; accepted 7 January 2014; accepted article preview 20 January 2014; advance online publication, 11 February 2014
 An overview on CALR and CSF3R mutations
A Tefferi et al
1408
Table 1. World Health Organization classification of myeloid malignancies

1. Acute myeloid leukemia (AML) and related precursor neoplasmsa

2. Myeloproliferative neoplasms (MPN)
2.1. Chronic myelogenous leukemia, BCR-ABL1 positive (CML)
2.2. Polycythemia vera (PV)
2.3. Primary myelofibrosis (PMF)
2.4. Essential thrombocythemia (ET)
2.5. Chronic neutrophilic leukemia (CNL)
2.6. Chronic eosinophilic leukemia, not otherwise specified (CEL-NOS)
2.7. Mastocytosis
2.8. Myeloproliferative neoplasm, unclassifiable (MPN-U)
3. Myelodysplastic syndromes (MDS)
3.1. Refractory cytopeniab with unilineage dysplasia (RCUD)
3.1.1. Refractory anemia (ring sideroblasts o15% of erythroid precursors)
3.1.2. Refractory neutropenia
3.1.3. Refractory thrombocytopenia
3.2. Refractory anemia with ring sideroblasts (RARS; dysplasia limited to erythroid lineage and ring sideroblasts X15% of bone marrow erythroid
precursors)
3.3. Refractory cytopenia with multi-lineage dysplasia (RCMD; ring sideroblast count does not matter)
3.4. Refractory anemia with excess blasts (RAEB)
3.4.1. RAEB-1 (2–4% circulating or 5–9% marrow blasts)
3.4.2. RAEB-2 (5–19% circulating or 10–19% marrow blasts or Auer rods present)
3.5. MDS associated with isolated del(5q)
3.6. MDS, unclassifiable (MDS-U)
4. MDS/MPN overlap
4.1. Chronic myelomonocytic leukemia (CMML)
4.2. Atypical chronic myeloid leukemia, BCR-ABL1 negative (aCML)
4.3. Juvenile myelomonocytic leukemia (JMML)
4.4. MDS/MPN, unclassifiable (MDS/MPN-U)
4.4.1. Provisional entity: Refractory anemia with ring sideroblasts associated with marked thrombocytosis (RARS-T)
5. Myeloid and lymphoid neoplasms with eosinophilia and abnormalities of PDGFRA, PDGFRB or FGFR1c
5.1. Myeloid and lymphoid neoplasms with PDGFRA rearrangement
5.2. Myeloid neoplasms with PDGFRB rearrangement
5.3. Myeloid and lymphoid neoplasms with FGFR1 abnormalities
a
AML-related precursor neoplasms include ‘therapy-related MDS’ and ‘myeloid sarcoma’. bEither mono- or bi-cytopenia: hemoglobin level o10 g/dl, absolute
neutrophil count o1.8  109/l or platelet count o100  109/l. However, higher blood counts do not exclude the diagnosis in the presence of unequivocal
histological/cytogenetic evidence for MDS. cGenetic rearrangements involving platelet-derived growth factor receptor a/b (PDGFRA/PDGFRB) or fibroblast
growth factor receptor 1 (FGFR1).

Table 2. 2008 World Health Organization diagnostic criteria for myeloproliferative neoplasms

Polycythemia vera (PV)a Essential thrombocythemia (ET)a Primary myelofibrosis (PMF)a

Major criteria
1 Hemoglobin (Hgb) Platelet count X450  109/l Megakaryocyte proliferation and atypia,c accompanied
418.5 g/dl (men) by either reticulin and/or collagen fibrosis, ord
416.5 g/dl (women) orb
2 Presence of JAK2V617F or Megakaryocyte proliferation with large and mature Not meeting WHO criteria for CML, PV, MDS, or other
JAK2 exon 12 mutation morphology. myeloid neoplasm
3 Not meeting WHO criteria for CML, PV, PMF, MDS or Demonstration of JAK2V617F or other clonal marker or
other myeloid neoplasm no evidence of reactive BM fibrosis
4 Demonstration of JAK2V617F or other clonal
marker or no evidence of reactive thrombocytosis

Minor criteria
1 BM trilineage — Leukoerythroblastosis
myeloproliferation
2 Subnormal serum — Increased serum LDH level
erythropoietin level
3 Endogenous erythroid — Anemia
colony growth
4 — Palpable splenomegaly
Abbreviations: BM, bone marrow; CML, chronic myelogenous leukemia; LDH, lactate dehydrogenase; MDS, myelodysplastic syndrome; WHO, World Health
Organization. aPV diagnosis requires meeting either both major criteria and one minor criterion or the first major criterion and two minor criteria. ET diagnosis
requires meeting all four major criteria. PMF diagnosis requires meeting all three major criteria and two minor criteria. bHgb or hematocrit 499th percentile of
reference range for age, sex or altitude of residence or red cell mass 425% above mean normal predicted or Hgb 417 g/dl (men) and 415 g/dl (women) if
associated with a sustained increase of X2 g/dl from baseline that cannot be attributed to correction of iron deficiency cSmall to large megakaryocytes with
aberrant nuclear/cytoplasmic ratio and hyperchromatic and irregularly folded nuclei and dense clustering. dIn the absence of reticulin fibrosis, the
megakaryocyte changes must be accompanied by increased marrow cellularity, granulocytic proliferation and often decreased erythropoiesis (that is,
prefibrotic PMF).

Leukemia (2014) 1407 – 1413 & 2014 Macmillan Publishers Limited

 An overview on CALR and CSF3R mutations
A Tefferi et al
1409
Table 3. World Health Organization (WHO) diagnostic criteria for CNL, aCML and CMML

Chronic neutrophilic leukemia (CNL) Atypical chronic myeloid leukemia (aCML) Chronic myelomonocytic leukemia (CMML)
9 9
PB leukocytosis X25  10 /l PB leukocytosis X13  10 /l Persistent PB monocytosis 41  109/l
480% Segmented neutrophils/bands mNeutrophils/precursors with prominent
dysgranulopoiesis No BCR-ABL1
o10% PB immature granulocytesa X10% Immature granulocytesa
o1% PB myeloblasts o20% PB myeloblasts No PDGFRA or PDGFRBmutations

BM hypercellular
Neutrophils increased in % and number No BCR-ABL1 o20% Blasts or promonocytes in BM or PB
o5% myeloblasts No PDGFRA or PDGFRBmutations Dysplasia in one or more myeloid lineages or
Normal neutrophilic maturation Clonal cytogenetic or molecular abnormality
Megakaryocytes normal/left shifted o2% PB basophilia Monocytosis has lasted for X3 months AND
all other causes of monocytosis excluded

Hepatosplenomegaly No monocytosis and o10% PB monocytes

Absence of reactive neutrophiliab BM hypercellular

Or presence of cytogenetic/molecular mGranulocyte proliferation and granulocytic
abnormality dysplasia with or without dysplasia in
erythroid or megakaryocytic lineages

No BCR-ABL1

No PDGFRA, PDGFRB or FGRF1mutations o20% Myeloblasts

No evidence PV, PMF or ET

No evidence of MDS or MDS/MPN overlap

No granulocytic dysplasia
No MDS-like changes
PB monocytes o1  109/l
Abbreviations: BM, bone marrow; ET, essential thrombocythemia; MDS, myelodysplastic syndromes; MPN, myeloproliferative neoplasms; PB, peripheral blood;
PMF, primary myelofibrosis; PV, polycythemia vera. aImmature granulocytes include myeloblasts, promyelocytes, myelocytes and metamyelocytes bCauses of
reactive neutrophilia include plasma cell neoplasms, solid tumor, infections and inflammatory processes.

mutations,7 the same had not been realized for the 35–40% of In the second study by Nangalia et al,14 CALR mutations were
patients with ET or PMF who do not express JAK2 mutations.8–11 not seen in 258 patients with PV/post-PV MF or 253 patients with
The discovery of MPL mutations in 200612 only partially addressed JAK2/MPL-mutated ET/PMF/ post-ET MF. CALR mutational
the problem because of its low mutational frequency, estimated at frequencies were 71%, 56% and 86% in JAK2/MPL-unmutated ET
10–20% in JAK2-unmutated ET or PMF.10,11,13 In December 2013, (n ¼ 112), PMF (n ¼ 32) or post-ET MF (n ¼ 14), respectively. In this
two groups reported the occurrence of novel calreticulin (CALR) study by Nangalia et al.,14 CALR mutations were infrequently seen
mutations in JAK2/MPL-unmutated PMF or ET.14,15 Both groups in MDS (8.3%; n ¼ 120) including RARS (11%; n ¼ 27), CMML (3%;
found mutual exclusivity between CALR, JAK2 and MPL mutations. n ¼ 33) and atypical CML (3.4%; n ¼ 29) but not in RARS-T (n ¼ 6),
CALR is a multi-functional Ca2 þ -binding protein chaperone CML (n ¼ 28), AML (n ¼ 48), systemic mastocytosis (n ¼ 114),
mostly localized in the endoplasmic reticulum (ER). Located on eosinophilic disorders (n ¼ 2), ‘idiopathic’ erythrocytosis (n ¼ 5),
chromosome 19p13.2, CALR contains nine exons and its transient abnormal myelopoiesis (n ¼ 10), lymphoid malignancies
protein three domains: N-domain (residues 1–180), P-domain (n ¼ 287), solid tumors (n ¼ 502) or control samples (n ¼ 550).
(residues 181–290) and C-domain (residues 291–400). Knocking Similar to the observations by Klampf et al.,15 Nangalia et al.14 also
out CALR in mice is lethal and causes impaired cardiac found an association between CALR mutations and higher platelet
development.16 count and lower hemoglobin level in ET; in addition, their study
In the study by Klampfl et al.,15 CALR mutations were not suggested an increased incidence of fibrotic transformation
seen in 382 cases of PV but were detected in 25% of patients in CALR-mutated ET without apparent survival difference.14
with ET (n ¼ 311) and 35% of those with PMF (n ¼ 203). The All CALR mutations seen in the above two studies14,15 were
authors subsequently enriched their JAK2/MPL-unmutated patient exon 9 frame-shift mutations with somatic insertions or deletions.
population by adding 211 such cases with ET or PMF and reported Two variants constituted more than 80% of the CALR mutations
CALR mutational frequencies of 67% and 88% in JAK2/MPL- seen: type 1 variant (p.L367fs*46) resulted from 52 bp deletion and
unmutated ET (n ¼ 289) and PMF (n ¼ 120), respectively. They also was more frequent in PMF, and type 2 variant (p.K385fs*47)
found CALR mutations in 3 (13%) of 24 patients with refractory resulted from 5-bp TTGTC insertion. In the study by Klampfl et al.,
15
anemia with ring sideroblasts and marked thrombocytosis these two common variants, respectively, accounted for 53%
(RARS-T). In this particular study,15 CALR mutations were not and 32% of all mutant CALR; the corresponding frequencies
seen in acute myeloid leukemia (AML; n ¼ 254), MDS (n ¼ 73), in the study by Nangalia et al.14 were 45% and 41%. Only three
CMML (n ¼ 64) or CML (n ¼ 45). In regards to clinical and cases of homozygous CALR mutations were reported and all three
laboratory correlative studies, CALR, compared with JAK2, displayed p.K385fs*47. Several other distinct variants were seen
mutations were associated with lower hemoglobin level, lower infrequently. All CALR mutations resulted in one base pair reading
leukocyte count, higher platelet count, lower risk of thrombosis frame shift and an altered C-terminal that is missing the KDEL
and better survival in ET and lower leukocyte count, higher (lysine, aspartic acid, glutamic acid and leucine) endoplasmic
platelet count and better survival in PMF.15 reticulum retention motif and is positively rather than negatively

& 2014 Macmillan Publishers Limited Leukemia (2014) 1407 – 1413

 An overview on CALR and CSF3R mutations
A Tefferi et al
1410
charged. CALR mutations were found in hematopoietic with CSF3 and close to 80% of such patients harbor acquired
progenitors and did not appear to affect the intracellular nonsense mutations of the gene encoding the receptor for CSF3
localization of the mutant protein.14 Overexpression of the most (CSF3R); mutational frequency in SCN patients without AML is
frequent CALR deletion in Ba/F3 cell lines caused cytokine- estimated at 30%.19 SCN-associated CSF3R mutations occur in the
independent growth, and activation of signal transducer and region of the gene (nucleotides 2300–2600) coding for the
activator of transcription 5 that was sensitive to pharmacologic cytoplasmic domain of CSF3R and result in truncation of the
JAK2 inhibition.15 C-terminal-negative regulatory domain.20
Several other papers on CALR mutations in MPN have since In 2009, Plo et al. 21 described a germline CSF3R transmembrane
been published.10,11,17,18 Rotunno et al.10 studied 576 patients mutation (C-to-A substitution at nucleotide 2088; T617N) in
with WHO-defined ET and reported mutational frequencies of autosomal dominant hereditary neutrophilia. The same exon 16
JAK2, CALR, MPL and ‘triple-negative’ at 64%, 15%, 4% and 16%, mutation is now identified as T640N and was previously described
respectively. In JAK2/MPL-unmutated cases, CALR mutational as an acquired event in AML.22CSF3RT617I was shown to induce
frequency was 49%, which was significantly lower than those CSF3R-independent granulocyte proliferation and differentiation
reported above by Nangalia et al.14 (71%) and Klampfl and JAK2 inhibitor-sensitive constitutive phosphorylation of JAK2,
et al.15(67%). This discrepancy probably reflects differences in STAT3, AKT and ERK.21 In mice, the mutation produced
patient selection in terms of the methodology used in diagnosing neutrophilia and splenomegaly.21 In 2012, Beekman et al.23
ET; the study by Rotunno et al.10 used strict WHO criteria, whereas described a novel CSF3R autoactivating mutation (CSF3RT595I,
the other studies did not and might have therefore included which is the same mutation as CSF3RT618I) in a patient with
patients with early/prefibrotic PMF. CALR vs JAK2 mutations were SCN/AML that co-existed with a truncating CSF3R mutation. The
associated with male sex, younger age, lower leukocyte count, same mutation was noted in a single patient among 199 AML
lower hemoglobin level and higher platelet count. The association cases from a separate database.
with male sex but not young age, leukocyte count or hemoglobin In 2013, Maxson et al.24 established the connection between
level was sustained when CALR mutations were compared with CNL and activating CSF3R mutations. The authors detected CSF3R
MPL mutations or triple-negative cases. Platelet counts were mutations in 16 (59%) of 27 patients with CNL (n ¼ 9) or aCML
similar between CALR- and MPL-mutated cases. CALR-mutated and (n ¼ 18) and 1 of 92 patients with AML; mutational frequency was
triple-negative cases displayed superior thrombosis-free survival.10 89% in CNL and 44% in aCML. The main mutations were
In another ET study, Rumi et al. 17 reported mutational membrane proximal and included T618I (n ¼ 12; exon 14 C-to-T
frequencies of 62, 24, 4 and 10% for JAK2, CALR, MPL and triple- substitution at nucleotide 1853) and T615A (n ¼ 2) but other
negative cases. Compared with JAK2-mutated cases, CALR- truncating mutations, often in association with T618I/T615A, were
mutated patients were younger and displayed lower leukocyte also seen; 9 of 12 T618I mutations occurred without associated
count, lower hemoglobin level and higher platelet count. This truncating mutations. In vitro drug sensitivity assays suggested
study also found no difference in overall survival, risk of leukemic JAK inhibitor sensitivity for membrane proximal but not truncating
or fibrotic transformation but better thrombosis-free survival in mutations, which were instead sensitive to dasatinib. These
CALR- vs JAK2-mutated patients. In a third study of 289 patients authors subsequently reported CSF3RT618I-induced lethal
referred for evaluation of persistent thrombocytosis but not myeloproliferative disorder in a murine bone marrow transplant
confirmed ET,18 mutational frequencies were 65 JAK2, 9 CALR, 3 model.25 Treatment with ruxolitinib was reported to have a
MPL and 23% triple-negative. The latter group had lower hemo- salutary effect both in this mouse model and a patient with
globin and platelet levels compared with JAK2-mutated cases and CSF3RT618I mutation.24
lower platelet count compared with CALR-mutated cases.18 Subsequent to the report by Maxson et al.,24 Pardanani et al.26
In PMF, Tefferi et al. studied 254 patients who were performed CSF3R mutation analysis on 54 patients with clinically
cytogenetically characterized and were also screened for several suspected CNL (n ¼ 35) or aCML (n ¼ 19). After central pathology
MPN-characteristic mutations including ASXL1, EZH2, IDH and review, 12 cases were confirmed to have WHO-defined CNL and 9
spliceosome mutations (SF3B1, SRSF2 and U2AF1). Mutational WHO-defined aCML. CSF3R mutations were detected in 13
frequencies were 58 JAK2, 25 CALR, 8 MPL and 9% triple-negative. patients and included CSF3RT618I in 10 cases. The latter
Mutational frequency in JAK2/MPL-unmutated cases was 74%. occurred exclusively in WHO-defined CNL for a mutational
Interestingly, one patient expressed both JAK2 and CALR muta- frequency of 83% and not seen in WHO-defined aCML, PMF
tions. CALR mutations were associated with younger age, higher (n ¼ 76) or CMML (n ¼ 94). Moreover, the two T618I-negative
platelet count and lower DIPSS-plus score. CALR-mutated patients patients with WHO-defined CNL harbored other CSF3R mutations
were also less likely to be anemic, require transfusions or display (M696T and I598I). Interestingly, 4 of the 10 T618I-mutated cases
leukocytosis. Spliceosome mutations were infrequent in CALR- also expressed SETBP1 mutations. In a subsequent study of a
mutated patients, but no other molecular or cytogenetic patient harboring both CSF3RT618I and SETBP1D868N, the authors
associations were evident. In multivariable analysis, CALR muta- demonstrated the presence of both mutations in granulocytes,
tions had a favorable impact on survival that was independent of mononuclear cells and CD34 þ myeloid progenitors but not in T
both DIPSS-plus risk and ASXL1 mutation status. Compared with cells. Additional observations in the particular study included the
CALR-mutated cases, triple-negative patients displayed inferior sighting of homozygous CSF3RT618I single colonies, relative
leukemia-free survival. The study by Tefferi et al identified resistance to in vitro treatment with JAK1 and JAK2 inhibitors
‘CALR–ASXL1 þ ’ and ‘triple-negative’ mutation profiles to be and lack of clinical response to ruxolitinib. Other studies have
prognostically detrimental. subsequently reported the rare occurrence of T618I (n ¼ 5) or
T617I (n ¼ 2) CSF3R mutations among 1446 consecutive patients
with de novo AML27 and no exon 14 mutations among 354
THE CSF3R STORY patients with CMML; approximately 4% of the latter displayed
Granulocyte colony-stimulating factor (G-CSF), also known as other CSF3R mutations including E808K and M696T.28
colony-stimulating factor 3 (CSF3), contributes to the proliferation
and granulocyte differentiation of myeloid progenitor cells.
Recombinant CSF3 has been used for the treatment of severe PROPOSALS FOR REVISION OF THE DIAGNOSTIC CRITERIA FOR
neutropenia associated with a variety of conditions, including PV, ET AND PMF
severe congenital neutropenia (SCN).19 More than a third of The discovery of JAK2V617F in 2004 led to a sharp turn in our
patients with SCN transform into AML after 15 years of treatment diagnostic approach and therapeutic direction in BCR-ABL1-

Leukemia (2014) 1407 – 1413 & 2014 Macmillan Publishers Limited

 An overview on CALR and CSF3R mutations
A Tefferi et al
1411
negative MPN.29 Soon thereafter, other JAK27 and MPL12 JAK2/MPL/CALR (16% of cases),10 the first three major criteria and
mutations were described and further enriched the pool of the one minor criterion (Table 4).
clonal markers for the distinction of MPN from reactive In PMF, about 90% of patients harbor JAK2, CALR or MPL
myeloproliferation. JAK2V617F is present in more than 95% of mutations.11 Furthermore, in the presence of a typical BM mor-
patients with PV, and other JAK2 mutations are detected in the phology and absence of evidence for other myeloid malignancies,
majority of the remaining JAK2V617F-negative cases.30,31 these clonal markers are specific enough to confirm the diagnosis.
Therefore, in its 2008 revision of the WHO document, it was Therefore, it is reasonable to make the diagnosis of PMF in the
appropriate for the WHO MPN subcommittee to include the presence of three major criteria: (i) typical BM morphology, (ii)
presence of JAK2 mutations as a major criterion for the diagnosis absence of evidence for another myeloid malignancy including
of PV (Table 2).2 Furthermore, because JAK2 mutations also occur CML, PV and ET, and (iii) presence of JAK2, CALR or MPL mutations
in 60–65% of patients with ET or PMF,10,11 they were also listed as (Table 4). Such a strategy excludes the possibility of nonclonal
useful clonal markers in the diagnosis of ET and PMF (Table 2). The causes of bone marrow fibrosis, and inter-MPN diagnostic confusion
2008 WHO diagnostic criteria for BCR-ABL1-negative MPN also is minimized by strict adherence to the second major criterion
allowed the use of other clonal markers, such as cytogenetic (Table 4). In the absence of JAK2, CALR or MPL mutations (that is,
abnormalities or presence of MPL mutations, in distinguishing triple-negative cases), diagnosis of PMF requires not only meeting
between clonal and reactive thrombocytosis or bone marrow the first two major criteria but also (i) exclusion of reactive bone
fibrosis (Table 2). The committee also recognized the fact that marrow fibrosis and (ii) presence of clinical and laboratory features
neither JAK2 nor MPL mutations were disease-specific and able to that are typical of PMF. These additional criteria are now listed as
distinguish between the three variants of BCR-ABL1-negative MPN. minor criteria in the proposed revised scheme (Table 4). In other
Accordingly, BM morphology was included as a minor criterion in words, in the absence of JAK2/CALR/MPL mutations, the first minor
PV and major criterion in both ET and PMF (Table 2). criterion helps exclude the possibility of non-clonal bone marrow
Where do CALR mutations fit in the above scheme? CALR fibrosis, whereas the second and third minor criteria reinforce the
mutations are frequent in JAK2/MPL-unmutated ET/PMF (esti- morphologic impression of PMF.
mated at 49% in strictly WHO-defined ET and 74% in WHO- In PV, according to the 2008 WHO system, the first major
confirmed PMF)10,11 and thus provide a much needed clonal diagnostic criterion requires one of the following four compo-
marker in such cases. Therefore, they should be included as nents: (i) hemoglobin level 418.5 g/dl in men and 416.5 g/dl in
named clonal markers, along with JAK2 and MPL mutations, in the women, or (ii) hemoglobin or hematocrit level that is 499th
list of major diagnostic criteria for ET and PMF (Table 4). percentile of reference range for age, sex, or altitude of residence,
Furthermore, in the context of consistent morphology, CALR, as or (iii) red cell mass that is 425% above mean normal predicted,
well as JAK2 and MPL mutations, is relatively specific to MPN and or (iv) hemoglobin level 417 g/dl (415 g/dl in women) associated
should therefore be separated from ‘other’ clonal markers (for with a sustained increase of X2 g/dl from baseline that cannot be
example, abnormal karyotype) and bundled together as a distinct attributed to correction of iron deficiency. The latter three
category in the criteria list (Table 4). However, CALR mutations do components were designed to capture PV patients with borderline
not fully address the molecular gap in JAK2/MPL-unmutated increased hemoglobin that is 418.5 g/dl in men and 416.6 g/dl
disease or distinguish between ET and early/prefibrotic PMF. As in women. In a somewhat similar gesture, we have recently
such, their availability does by no means undercut the necessity of re-introduced32 the term ‘masked PV’ for JAK2-mutated patients
BM morphology as a major diagnostic criterion in both ET and who display PV-characteristic BM morphology but display
PMF (Table 4). Accordingly, ET diagnosis still requires meeting the hemoglobin levels between 16 and 18.5 g/dl for men and 15
four major diagnostic criteria listed in Table 4, or, in the absence of and 16.5 g/dl for women.33,34 We subsequently determined a

Table 4. 2014 proposed revision for World Health Organization (WHO) diagnostic criteria for BCR-ABL1-negative myeloproliferative neoplasms

Polycythemia vera (PV)a Essential thrombocythemia (ET)b Primary myelofibrosis (PMF)c

Major criteria
1 Hemoglobin 416.5 g/dl (men) 416 g/ Platelet count X450  109/l Megakaryocyte proliferation and atypiad,
dl (women) or hematocrit 449% (men) accompanied by either reticulin and/or
448% (women) collagen fibrosis ore
2 BM trilineage myeloproliferation with Megakaryocyte proliferation with large and Not meeting WHO criteria for CML, PV, ET, MDS
pleomorphic megakaryocytes mature morphology or other myeloid neoplasm
3 Presence of JAK2 mutation Not meeting WHO criteria for CML, PV, PMF, Presence of JAK2, CALR or MPL mutation
MDS or other myeloid neoplasm
4 Presence of JAK2, CALR or MPL mutation

Minor criteria
1 Subnormal serum erythropoietin level Presence of a clonal marker (e.g. abnormal Presence of a clonal marker (e.g. abnormal
karyotype) or absence of evidence for karyotype) or absence of evidence for reactive
reactive thrombocytosis bone marrow fibrosis
2 Presence of anemia or palpable splenomegaly
3 Presence of leukoerythroblastosisf, or increased
lactate dehydrogenasef
Abbreviations: BM, bone marrow; CML, chronic myelogenous leukemia; MDS, myelodysplastic syndrome. aPV diagnosis requires meeting either all three major
criteria or the first two major criteria and one minor criterion. bET diagnosis requires meeting all four major criteria or first three major criteria and one minor
criterion. cPMF diagnosis requires meeting all three major criteria or the first two major criteria and all three minor criteria. dSmall-to-large megakaryocytes
with aberrant nuclear/cytoplasmic ratio and hyperchromatic and irregularly folded nuclei and dense clustering. eIn the absence of reticulin fibrosis, the
megakaryocyte changes must be accompanied by increased marrow cellularity, granulocytic proliferation and often decreased erythropoiesis (that is,
prefibrotic PMF). fDegree of abnormality can be borderline or marked and institutional reference range should be used for lactate dehydrogenase level.

& 2014 Macmillan Publishers Limited Leukemia (2014) 1407 – 1413

 An overview on CALR and CSF3R mutations
A Tefferi et al
1412
with hemoglobin (hematocrit) levels of 418.5 g/dl (452%) in
Table 5. Proposal for revision of World Health Organization (WHO)
men and 416.5 g/dl (448%) in women.
criteria for diagnosis of chronic neutrophilic leukemia

Major criteria 1 Peripheral blood (PB) leukocytosis X13  109/l

2 480% Neutrophils or bands PROPOSALS FOR REVISION OF THE DIAGNOSTIC CRITERIA
3 Presence of CSF3RT618I or other membrane- FOR CNL
proximal CSF3R mutations The current WHO diagnostic criteria for CNL are primarily driven
by the absence of a clonal marker and disease-specific BM
Minor criteria 1 Bone marrow:
morphologic traits.37 Accordingly, they require a relatively
Hypercellular
Increased granulocytic proliferation without high leukocyte count threshold (X25  10(9)/l) and a list of
significant left shift exclusion criteria that focus on reactive granulocytosis, plasma
No dysgranulopoiesis cell neoplasm-associated neutrophilia, CML, aCML and CMML
2 Peripheral blood: (Table 3). In clinical practice, only a fraction of clinically suspected
o10% immature granulocytes cases of CNL meets strict WHO criteria and this might also be true
o2% myeloblasts for some published cases.26,38 Conversely, it is equally possible to
p1  109 absolute monocyte count or misdiagnose some cases of CNL as aCML or CMML (Table 3).
o10% monocyte percentage The recent discovery of CSF3R mutations and their almost
No dysgranulopoiesis
invariable association with WHO-defined CNL presents the
3 Presence of a clonal marker or absence of
evidence for reactive/secondary opportunity to make significant changes in our diagnostic
granulocytosis, including plasma cell approach (Table 5).24,26 Such availability of a clonal marker
neoplasm for the majority of patients with CNL should allow lowering of
4 Absence of BCR-ABL1 the diagnostic leukocyte count threshold level from 25  109/l to
5 Not meeting WHO diagnostic criteria for any 13  109/l because the latter represents a level above 3 s.d. from
other myeloid neoplasm the normal mean and would be consistent with what is currently
Diagnosis requires presence of all three major criteria or the first two major
being used for the diagnosis of WHO-defined aCML.1
criteria and all minor criteria. In addition to the above, we propose separate sets of major and
minor criteria to accommodate the diagnostic possibility in both
CSF3R-mutated and unmutated CNL (Table 5). The presence of a
hemoglobin level of 16.5 g/dl in men and 16 g/dl for women or a membrane proximal CSF3R mutation in a patient with predomi-
hematocrit level of 49% in men and 48% in women to be the nantly neutrophilic granulocytosis should be sufficient for the
optimal cutoff levels for distinguishing JAK2-mutated ET from diagnosis of CNL.24,26 This is now signified in the form of three
masked PV (manuscript in preparation). major diagnostic criteria including (i) a PB leukocyte level of
Strict adherence to all four components of the first major X13  109/l, (ii) a PB neutrophil/band percent distribution of
criterion for diagnosis of PV might capture a substantial number of 480% and (iii) presence of a CSF3RT618I or other membrane
patients with masked PV. In contrast, lowering the diagnostic proximal CSF3R mutations. We realize that one can argue for
hemoglobin/hematocrit level to 16.5 g/dl (49%) in men and 16 g/ including the absence of dysgranulopoiesis as a major criterion
dl (48%) in women, as suggested by the aforementioned masked based on the discrepancy in current literature regarding the
PV study, might simplify the first major diagnostic criterion, but occurrence of CSF3R mutations in aCML.24,26 However, careful
would require morphologic confirmation. Furthermore, regardless analysis of Mayo Clinic cases26 and additional evidence from other
of whether or not one buys into the concept of masked PV, such a centers have not confirmed the occurrence of CSF3R mutations
measure conjures therapeutic relevance because of the recent in strictly WHO-defined aCML. Regardless, we believe that the
associations between increased thrombotic complications inclusion of ‘CSF3R-mutated aCML’ in the molecularly similarly
and borderline increased hematocrit (45–50%) in PV and defined CNL category makes better scientific sense.
JAK2 mutation in ET.35,36 We therefore propose consolidation Unfortunately, several exclusionary criteria still need to be met
of the currently listed four components of the first major for diagnosing CNL in the absence of CSF3R mutations, and these
diagnostic criterion into one component based on a lower are now listed as minor criteria (Table 5). Finally, before embarking
threshold hemoglobin level and hematocrit reinforced by the on a testing spree for CSF3R mutations, one should be aware of
inclusion of BM morphology as a major criterion (Table 4). The the fact that CNL is an extremely rare disorder and that it is very
inclusion of BM morphology as a major criterion also allows the unlikely to be the cause of neutrophilia that is encountered in
diagnosis of PV based on major criteria alone without the need for routine clinical practice.37,38
additional minor criteria. However, in order to address the rare
possibility of a JAK2-unmutated PV, we have included ‘subnormal
erythropoietin level’ as a minor criterion. Endogenous CONCLUSIONS
erythroid colony growth is now removed from the list of The last 10 years have been marked by remarkable genetic
minor criteria because of redundancy in value and limited discoveries in MPN, which should in time lead to equally impressive
practical use. therapeutic advances. The identification of disease-specific muta-
In terms of routine clinical practice, JAK2 mutation screening tions justifies, in certain cases, the use of peripheral blood mutation
should start with JAK2V617F because it represents more than 95% analysis for diagnosis, prognosis and monitoring of treatment
of all JAK2 mutations. Other JAK2 mutations, including JAK2 exon response. It is inevitable that more mutations will be discovered in
12 mutations, should be studied only in the absence of JAK2V617F the near future, which is welcome news not only for patients but
and presence of subnormal serum erythropoietin level. Mutation also for the highly profitable in vitro diagnostics market. However,
screening in ET or PMF should also start with JAK2V617F, followed new rules on reimbursement for molecular tests dictate a joint effort
by CALR mutations in JAK2-unmutated case, whereas MPL between clinicians and pathologists in order to formulate affordable
mutation screening, because of its low mutational frequency, is testing schedules with measurable clinical benefit.39
best reserved for JAK2 and CALR unmutated cases. In regards to
bone marrow examination, whereas it is mandatory for clinical trial
and research study purposes, it might not be necessary for CONFLICT OF INTEREST
clinically overt cases of PV. These include JAK2-mutated patients The authors declare no conflict of interest.

Leukemia (2014) 1407 – 1413 & 2014 Macmillan Publishers Limited

 An overview on CALR and CSF3R mutations
A Tefferi et al
1413
REFERENCES 20 Germeshausen M, Ballmaier M, Welte K. Incidence of CSF3R mutations in severe
1 Swerdlow SHCE, Harris NL, Jaffe ES, Pileri SA, Stein H, Thiele J et al. WHO congenital neutropenia and relevance for leukemogenesis: Results of a long-term
Classification Of Tumours Of Haematopoietic And Lymhoid Tissues. IARC press: Lyon, survey. Blood 2007; 109: 93–99.
France, 2008. 21 Plo I, Zhang Y, Le Couedic JP, Nakatake M, Boulet JM, Itaya M et al. An activating
2 Vardiman JW, Thiele J, Arber DA, Brunning RD, Borowitz MJ, Porwit A et al. The mutation in the CSF3R gene induces a hereditary chronic neutrophilia. J Exp Med
2008 revision of the World Health Organization (WHO) classification of myeloid 2009; 206: 1701–1707.
neoplasms and acute leukemia: rationale and important changes. Blood, 2009; 22 Forbes LV, Gale RE, Pizzey A, Pouwels K, Nathwani A, Linch DC. An activating
114(5): 937–951. mutation in the transmembrane domain of the granulocyte colony-stimulating
3 Heller P, Kornblihtt LI, Cuello MT, Larripa I, Najfeld V, Molinas FC. BCR-ABL factor receptor in patients with acute myeloid leukemia. Oncogene 2002; 21:
transcripts may be detected in essential thrombocythemia but lack clinical 5981–5989.
significance. Blood 2001; 98: 1990. 23 Beekman R, Valkhof MG, Sanders MA, van Strien PM, Haanstra JR, Broeders L et al.
4 Pieri L, Spolverini A, Scappini B, Occhini U, Birtolo S, Bosi A et al. Concomitant Sequential gain of mutations in severe congenital neutropenia progressing to
occurrence of BCR-ABL and JAK2V617F mutation. Blood 2011; 118: 3445–3446. acute myeloid leukemia. Blood 2012; 119: 5071–5077.
5 Thiele J, Kvasnicka HM, Fischer R. Bone marrow histopathology in chronic 24 Maxson JE, Gotlib J, Pollyea DA, Fleischman AG, Agarwal A, Eide CA et al.
myelogenous leukemia (CML)-evaluation of distinctive features with clinical Oncogenic CSF3R mutations in chronic neutrophilic leukemia and atypical CML.
impact. Histol Histopathol 1999; 14: 1241–1256. New Engl J Med 2013; 368: 1781–1790.
6 Granger JM, Kontoyiannis DP. Etiology and outcome of extreme leukocytosis in 25 Fleischman AG, Maxson JE, Luty SB, Agarwal A, Royer LR, Abel ML et al. The CSF3R
758 nonhematologic cancer patients: a retrospective, single-institution study. T618I mutation causes a lethal neutrophilic neoplasia in mice that is responsive to
Cancer 2009; 115: 3919–3923. therapeutic JAK inhibition. Blood 2013; 122: 3628–3631.
7 Scott LM, Tong W, Levine RL, Scott MA, Beer PA, Stratton MR et al. JAK2 exon 12 26 Pardanani A, Lasho TL, Laborde RR, Elliott M, Hanson CA, Knudson RA et al. CSF3R
mutations in polycythemia vera and idiopathic erythrocytosis. New Engl J Med T618I is a highly prevalent and specific mutation in chronic neutrophilic leukemia.
2007; 356: 459–468. Leukemia 2013; 27: 1870–1873.
8 Tefferi A, Lasho TL, Abdel-Wahab O, Guglielmelli P, Patel J, Caramazza D et al. 27 Beekman R, Valkhof M, van Strien P, Valk PJ, Touw IP. Prevalence of a new
IDH1 and IDH2 mutation studies in 1473 patients with chronic-, fibrotic- or auto-activating colony stimulating factor 3 receptor mutation (CSF3R-T595I) in
blast-phase essential thrombocythemia, polycythemia vera or myelofibrosis. acute myeloid leukemia and severe congenital neutropenia. Haematologica 2013;
Leukemia 2010; 24: 1302–1309. 98: e62–e63.
9 Vannucchi AM, Lasho TL, Guglielmelli P, Biamonte F, Pardanani A, Pereira A et al. 28 Kosmider O, Itzykson R, Chesnais V, Lasho T, Laborde R, Knudson R et al.
Mutations and prognosis in primary myelofibrosis. Leukemia 2013; 27: Mutation of the colony-stimulating factor-3 receptor gene is a rare event
1861–1869. with poor prognosis in chronic myelomonocytic leukemia. Leukemia 2013; 27:
10 Rotunno G, Mannarelli C, Guglielmelli P, Pacilli A, Pancrazzi A, Pieri L et al. Impact 1946–1949.
of calreticulin mutations on clinical and hematological phenotype and outcome 29 James C, Ugo V, Le Couedic JP, Staerk J, Delhommeau F, Lacout C et al. A unique
in essential thrombocythemia. Blood 2013; e-pub ahead of print 26 December clonal JAK2 mutation leading to constitutive signalling causes polycythaemia
2013; doi:10.1182/blood-2013-11-538983. vera. Nature 2005; 434: 1144–1148.
11 Tefferi A, Lasho TL, Finke CM, Knudson RA, Ketterling R, Hanson CH et al. CALR vs 30 Pardanani A, Lasho TL, Finke C, Hanson CA, Tefferi A. Prevalence and
JAK2 vs MPL- mutated or triple-negative myelofibrosis: clinical, cytogenetic and clinicopathologic correlates of JAK2 exon 12 mutations in JAK2V617F-negative
molecular comparisons. Leukemia 2014; 28: 1472–1477. polycythemia vera. Leukemia 2007; 21: 1960–1963.
12 Pikman Y, Lee BH, Mercher T, McDowell E, Ebert BL, Gozo M et al. MPLW515L is a 31 Tefferi A, Rumi E, Finazzi G, Gisslinger H, Vannucchi AM, Rodeghiero F et al.
novel somatic activating mutation in myelofibrosis with myeloid metaplasia. Survival and prognosis among 1545 patients with contemporary polycythemia
PLoS Med 2006; 3: e270. vera: an international study. Leukemia 2013; 27: 1874–1881.
13 Pardanani AD, Levine RL, Lasho T, Pikman Y, Mesa RA, Wadleigh M et al. MPL515 32 Shih LY, Lee CT. Identification of masked polycythemia vera from patients with
mutations in myeloproliferative and other myeloid disorders: a study of 1182 idiopathic marked thrombocytosis by endogenous erythroid colony assay. Blood
patients. Blood 2006; 108: 3472–3476. 1994; 83: 744–748.
14 Nangalia J, Massie CE, Baxter EJ, Nice FL, Gundem G, Wedge DC et al. Somatic 33 Barbui T, Thiele J, Gisslinger H, Finazzi G, Carobbio A, Rumi E et al. Masked
CALR mutations in myeloproliferative neoplasms with nonmutated JAK2. New polycythemia vera (mPV): results of an international study. Am J Hematol 2014; 89:
Engl J Med 2013; 369: 2391–2405. 52–54.
15 Klampfl T, Gisslinger H, Harutyunyan AS, Nivarthi H, Rumi E, Milosevic JD et al. 34 Barbui T, Thiele J, Vannucchi AM, Tefferi A. Rethinking the diagnostic criteria of
Somatic mutations of calreticulin in myeloproliferative neoplasms. New Engl J Med polycythemia vera. Leukemia 2014; 28: 1191–1195.
2013; 369: 2379–2390. 35 Marchioli R, Finazzi G, Specchia G, Cacciola R, Cavazzina R, Cilloni D et al. Cardi-
16 Mesaeli N, Nakamura K, Zvaritch E, Dickie P, Dziak E, Krause KH et al. Calreticulin is ovascular events and intensity of treatment in polycythemia vera. New Engl J Med
essential for cardiac development. J Cell Biol 1999; 144: 857–868. 2013; 368: 22–33.
17 Rumi E, Pietra D, Ferretti V, Klampfl T, Harutyunyan AS, Milosevic JD et al. JAK2 or 36 Tefferi A, Barbui T. Personalized management of essential thrombocythemia-
CALR mutation status defines subtypes of essential thrombocythemia with application of recent evidence to clinical practice. Leukemia 2013; 27: 1617–1620.
substantially different clinical course and outcomes. Blood 2013; e-pub ahead of 37 Elliott MA, Tefferi A. The molecular genetics of chronic neutrophilic leukaemia:
print 23 December 2013; doi:10.1182/blood-2013-11-539098. defining a new era in diagnosis and therapy. Curr Opin Hematol 2013; e-pub
18 Chi J, Nicolaou KA, Nicolaidou V, Kouma L, Mitsidou A, Pierides C et al. Calreticulin ahead of print 11 December 2013; doi:10.1097/MOH.0000000000000014.
exon 9 frameshift mutations in patients with thrombocytosis. Leukemia 2014; 28: 38 Bohm J, Schaefer HE. Chronic neutrophilic leukaemia: 14 new cases of an
1129. uncommon myeloproliferative disease. J Clin Pathol 2002; 55: 862–864.
19 Vandenberghe P, Beel K. Severe congenital neutropenia, a genetically hetero- 39 Engstrom PF, Bloom MG, Demetri GD, Febbo PG, Goeckeler W, Ladanyi M et al.
geneous disease group with an increased risk of AML/MDS. Pediatr Rep 2011; NCCN molecular testing white paper: effectiveness, efficiency, and reimburse-
3(Suppl 2): e9. ment. J Natl Compr Canc Netw 2011; 9(Suppl 6): S1–16.

& 2014 Macmillan Publishers Limited Leukemia (2014) 1407 – 1413