Article Text

Download PDFPDF

Immunohistochemistry and molecular analyses in myeloid sarcoma of the breast in a patient with relapse of NPM1-mutated and FLT3-mutated AML after allogeneic stem cell transplantation
  1. Matthias Choschzick1,
  2. Ulrike Bacher2,
  3. Francis Ayuk2,
  4. Annette Lebeau1
  1. 1Department of Pathology, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany
  2. 2Interdisciplinary Clinic for Stem Cell Transplantation, University Medical Centre Hamburg-Eppendorf, Hamburg, Germany
  1. Correspondence to Dr Matthias Choschzick, Department of Pathology, University Medical Centre Hamburg-Eppendorf, Martinistraße 52, 20246 Hamburg, Germany; mchoschz{at}


Myeloid sarcoma of the breast is a rare manifestation of acute myeloid leukaemia (AML). This report describes a patient who was diagnosed with AML FAB M2. Molecular analysis showed evidence of an NPM1 mutation (subtype A) and internal tandem duplications of the FLT3 gene (FLT3-ITD). Eight months after allogeneic stem cell transplantation, the patient developed a palpable mass in the left breast initially suspected as breast carcinoma. Core needle biopsy of the lesion resulted in diagnosis of myeloid sarcoma. Molecular analysis of formalin-fixed specimens of the breast tumour confirmed the known FLT3 and NPM1 gene mutations. Immunohistochemically, an aberrant cytoplasmic staining pattern for NPM1 and overexpression of FLT3 were demonstrated. The myeloid sarcoma showed complete transient resolution following treatment with the kinase inhibitor sorafenib. However, the patient developed bone marrow relapse and died in fatal cerebral haemorrhage 1 year after initial diagnosis of AML. In summary, combined molecular and immunohistochemical examination of NPM1 and FLT3 is helpful in the diagnosis of extramedullary manifestations of AML in core needle biopsies.

  • Acute myeloid leukaemia (AML)
  • FLT3-ITD
  • immunohistochemistry
  • myeloid sarcoma
  • myeloproliferative disease
  • NPM1 mutation
  • PCR
View Full Text

Statistics from


According to the recently modified World Health Organization classification, myeloid sarcoma is a tumour mass of myeloid precursor cells with variable maturation.1 The lesions occur mostly in the setting of acute myeloid leukaemia (AML) or other myeloid malignancies such as chronic myeloid leukaemia. The overall occurrence of myeloid sarcoma has been reported in up to 9% of AML cases, with variable frequencies depending on the subtypes. Common sites of involvement are bone, especially the skull, and lymph nodes and skin.

Cytogenetic alterations represent a strong prognostic parameter in AML and can be identified in the bone marrow in about 55% of patients with AML. In tumour specimens from patients with myeloid sarcoma, clonal chromosomal aberrations have also been described in more than 50% of cases by Alexiev et al.2 Unfortunately, detection of these cytogenetic changes is usually not feasible in histological sections. Furthermore, small crushed biopsies, or an unusual solitary extramedullary manifestation without concomitant infiltration of the bone marrow by the AML, may hamper the diagnosis or impair later follow-up of patients.

Several leukaemia-specific somatic mutations with relevance for a biological subclassification and for prognostic predictions have been identified in AML with normal karyotype. Mutations of the nucleophosmin gene (NPM1) and the fms-like tyrosine kinase 3 gene (FLT3) are the most frequent genetic aberrations in cytogenetically normal AML. NPM1 mutations have been detected in around 50% of these cases.3 Nucleophosmin is an ubiquitously expressed nucleolar protein, and it plays a role in cell proliferation, centrosome duplication and p53 tumour suppressor pathway regulation. The protein has the ability to shuttle between nucleus and cytoplasm. Mutations lead to aberrant cytoplasmic displacement of the nucleophosmin protein. About 30% of adult AML normal karyotype cases harbour a FLT3 mutation, mostly as internal tandem duplications (FLT3-ITD).4 FLT3 is predominantly expressed in haemopoietic progenitor cells and its mutation confers ligand-independent receptor activation and constitutive activation of downstream signalling pathways. Whereas isolated NPM1 mutations are predictive for more favourable outcomes in normal karyotype patients, presence of the FLT3-ITD is an adverse prognostic parameter. This adverse prognostic impact of the FLT3-ITD is also seen in AML patients with coincident NPM1 mutations.

We report a case of AML that showed evidence of NPM1 mutation in combination with FLT3-ITD. Allogeneic stem cell transplantation (SCT) was followed by relapse of the AML presenting as a myeloid sarcoma of the breast. We discuss the impact of molecular mutation screening and immunohistochemical staining of NPM1 and FLT3 for correct diagnosis of extramedullary manifestations of AML in biopsies.

Case report

The patient in the present report was diagnosed with AML of the FAB M2 subtype. Cytogenetic analysis of the bone marrow revealed a normal karyotype. In molecular analysis, a prognostically unfavourable FLT3-ITD with a 58 bp insert, together with a type A mutation of the NPM1 gene, were detected. The patient received induction chemotherapy with idarubicin, cytarabine and etoposide (according to the AML Study Group protocol), but was refractory to therapy. Subsequently, salvage therapy with gemtuzumab ozogamicin in combination with all-trans retinoic acid, high dosage cytarabine, and mitoxantrone (GO-A-HAM) was performed. Initially, complete remission was achieved but was followed by early relapse 2 months later. Thus, allogeneic SCT was performed from an HLA-matched unrelated donor using a dose-reduced conditioning regimen (fludarabine, amsacrine, cytarabine and busulfan) 4 months after the initial diagnosis. Following the allogeneic SCT, the cytological examination showed a hypocellular bone marrow without blasts. In the bone marrow, NPM1 mutation was not detected with quantitative real-time PCR (RQ-PCR) at this time. Six months after SCT, there was a slight increase of the mutation load in the peripheral blood to an expression ratio of 0.003% (performing normalisation with the HCK gene).5 One year after first diagnosis and 8 months from SCT, the patient consulted for a newly appeared palpable mass in the left breast. The NPM1 mutation analysis from peripheral blood showed a 0.086% ratio corresponding to molecular relapse, while cytomorphology showed no blasts in the peripheral blood. An ultrasound scan demonstrated a hypoechoic tumour measuring 4.4 cm in the upper outer quadrant of the breast. Additionally several enlarged lymph nodes in the left axilla were detected. The appearance was indistinguishable from breast carcinoma, and needle core biopsy of the tumour was performed.

Histological examination of core biopsies displayed diffuse infiltration of breast parenchyma by immature cells with polymorph nuclei and ill-defined cell borders (figure 1A). In most areas neoplastic cells were arranged in a discohesive single-file pattern. Sometimes a targetoid growth pattern around pre-existing lobules was observed. Focally tumour cells infiltrated lobular epithelial structures. The surrounding stroma was fibrotic with few lymphocytic cell infiltrates. Neoplastic cells were non-reactive for cytokeratin marker (antibody AE1/3), but reacted positively with the naphthol AS-D chloroacetate esterase stain (figure 1B). Upon further immunohistochemical examination, the atypical cells showed a strong expression of CD45 (leukocyte common antigen) and myeloperoxidase. The blasts were negative for CD20, CD3, CD79a, CD 117 and CD34. Antibodies against nucleophosmin clearly stained the cytoplasm of the tumour cells (figure 1C). Nevertheless, in most neoplastic cells a nuclear signal was also visible. However, when compared with normal cells, there was a clear shift of the nucleophosmin protein to the cytoplasm. Adjacent normal mammary lobules exhibited a moderate to strong nuclear signal in the epithelial cells. Antibodies for FLT3 showed a moderate cytoplasmic signal in neoplastic cells (figure 1D) corresponding to FLT3 overexpression. Non-neoplastic epithelial and mesenchymal structures reacted negatively at all times. Furthermore, the previously known FLT3-ITD mutation was detected by PCR, and gel electrophoresis in the core biopsies and RQ-PCR showed clear evidence of the NPM1A mutation (figure 2).

Figure 1

Myeloid sarcoma of the breast. (A) Diffuse infiltration of breast parenchyma by immature leukaemic blasts (H&E ×100). (B) Histochemically, tumour cells show a positive reaction for naphthol AS-D chloroacetate esterase (×100). (C) Immunohistochemical staining for nucleophosmin reveals a positive cytoplasmic reaction. (D) Positive cytoplasmic reaction of leukaemic blasts with antibodies against FLT3.

Figure 2

NPM1A and FLT3 mutations in myeloid sarcoma of the breast. (A) Quantitative real-time PCR analysis. *Curves showing the results of breast biopsy analysis 8 months after stem cell transplantation. The other curves represent the results as assessed by dilution of the OCI/AML3 cell line carrying the NPM1A mutation for standardisation of the mutation load in terms of percentage (before normalisation of the DNA content). (B) Gel electrophoresis following FLT3-amplification by PCR: a, FLT3-wildtype; b, the patient's FLT3-ITD mutated sample with two separated amplification lines (the upper line is marked with a white arrow for better visualisation; c, the MV4-11 cell line carrying a FLT3-ITD (positive control).

After diagnosis of myeloid sarcoma, the patient was treated with sorafenib (Nexavar), a multikinase inhibitor (200 mg daily). A significant reduction of the tumour mass was achieved about 2 weeks later. There was persistent sorafenib-induced thrombocytopenia, with platelets lower than 5×109/l in the follow-up, requiring thrombocyte substitution. After a short period of partial molecular remission with a decrease of the NPM1 ratio to 0.11% in the peripheral blood, there was another increase of the NPM1 mutation load up to 1.43% at an interval of 9 months from SCT. The patient died 15 months after primary diagnosis, and 10 months from SCT, from cerebral haemorrhage probably related to the thrombocytopenia.


Core needle biopsies of the breast tumour were fixed in 10% neutral buffered formalin and embedded in paraffin according to standard procedures. Immunohistochemical staining for NPM1 was performed using commercially available monoclonal antibody clone 376 (1:200; Dako, Glostrup, Denmark) according the recommendations of the manufacturer. For FLT3 detection, a polyclonal antibody was used (1:200; Acris Antibodies, Herford, Germany) as recommended by the supplier.

Following DNA extraction from the sample, screening for NPM1 mutations of the subtype A was performed with highly sensitive (TaqMan) RQ-PCR as described previously.5 Briefly, NPM1A expression was determined with a standard curve as assessed by serial dilution of the NPM1Amut-positive OCI/AML3 cell line. The undiluted cell line was used for the definition of a 100% NPM1A mutation ratio. Duplex PCRs were carried out using the HCK gene as a DNA content standard, and the relative NPM1Amut content was assessed for all available follow-up time points using the ΔCT method, giving the percentages of NPM1Amut expression in relation to the expression of the above cell line. Analysis of the FLT3-ITD was done with conventional PCR and gel electrophoresis.


Myeloid sarcoma of the breast is an uncommon condition with only few reported cases in the literature and is seen either isolated or as an extramedullary manifestation of AML or other myeloproliferative disorders.6 Without simultaneous morphological involvement of the bone marrow, the diagnosis may be especially difficult. The differential diagnosis includes undifferentiated carcinoma, invasive lobular breast carcinoma, and lymphoma.7 In rare cases, the differentiated type of myeloid sarcoma could be histomorphologically confused with benign extramedullary myeloid proliferations.8 Radiographic findings are not pathognomonic or specific in diagnosis of these lesions. In our case, diagnosis of the extramedullary manifestation of the AML was obtained by a combination of the clinical history, conventional histopathology and immunohistochemical analysis. Helpful markers are MPO, CD34, CD117 and negative staining for cytokeratin. Nevertheless, in cases with only small amounts of neoplastic cells, or in cases with an unusual immunohistochemical profile, molecular analysis may be helpful. The assessment of NPM1 mutations is a specific and sensitive test because of limited variations in the mutation pattern and its stability during disease progression.3 9 In the present case, the known NPM1 mutation of the A subtype was detected by RQ-PCR in formalin-fixed core biopsies. In contrast, the insertions in the FLT3-ITD range from 3 bp to over 400 bp, so DNA fragmentation might hamper the detection of this kind of mutation in fixed specimens. However, the specific FLT3 58 bp insertion of the leukaemic cell clone was detectable with PCR amplification in our material. Additionally, the aberrant cytoplasmic immunohistochemical staining pattern of NPM1 can facilitate the identification of the immature myeloid blasts. These molecular and immunohistochemical results confirmed the diagnosis of an extramedullary manifestation of the AML with the previously known NPM1A mutation and the prognostically adverse FLT3-ITD.

In our sections, non-neoplastic epithelial cells, inflammatory cells and stromal cells stained strictly nuclear for nucleophosmin, whereas the malignant blasts showed an additional cytoplasmic staining pattern. Nucleophosmin is expressed in a wide variety of normal tissues in contrast to FLT3, the expression pattern of which seems to be more restricted. FLT3 is detectable in a small fraction of lymphoid cells in non-neoplastic lymph nodes. Unfortunately, a comprehensive survey for expression in different malignant and benign tissues is not available at present. Nevertheless, a strong staining reaction for FLT3, as shown in our report, might be an indicator for overexpression and eventually, gene mutation. Therefore, immunostaining for FLT3 might be helpful to identify atypical immature myeloid cells and may be important for further diagnostic steps (e.g. PCR analysis) and therapeutic decisions.

FLT3-mutated AML can be targeted with a new generation of tyrosine kinase inhibitors. Sorafenib is a small-molecule inhibitor of various kinases that directly targets mutant FLT3.10 In the presented case, the myeloid sarcoma in the breast showed a complete response to sorafenib therapy, while in peripheral blood complete molecular remission was only transient. Finally, the patient died from cerebral haemorrhage.

In summary, this report illustrates the usefulness of immunohistochemical detection and mutation analysis of NPM1 and FLT3 in the diagnosis of extramedullary myeloid proliferations. These techniques can be highly valuable in addition to standard diagnostic procedures in the difficult differential diagnosis of myeloid sarcoma.

Take-home messages

  • We present a case of myeloid sarcoma in the breast corresponding to a relapse of acute myeloid leukaemia with known mutations in NPM1 and FLT3 genes.

  • Immunohistochemical staining of NPM1 and FLT3 may be a helpful completion of standard diagnostic procedures in myeloid sarcoma cases with a known positive mutation status.

  • Molecular detection of NPM1 mutations in core biopsies might be useful in difficult cases of extramedullary relapses with only small amounts of blasts or ambiguous morphology.


The authors would like to thank Anita Badbaran from the Laboratory of the Clinic for Stem Cell Transplantation for excellent technical work.


View Abstract


  • Competing interests None.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.