Annals of Diagnostic Pathology 18 (2014) 253–260
Contents lists available at ScienceDirect
Annals of Diagnostic Pathology
Review Article
A practical approach to diagnose soft tissue myeloid sarcoma preceding or coinciding with acute myeloid leukemia Robert P. Seifert, MD a, William Bulkeley III, MD b, Ling Zhang, MD a, c, Manuel Menes, MD d, Marilyn M. Bui, MD, PhD a, e,⁎ a
Department of Pathology and Cell Biology, University of South Florida Morsani College of Medicine, Tampa, FL James A. Haley Veterans' Hospital, Department of Pathology and Laboratory Medicine, Tampa, FL Department of Hematopathology and Laboratory Medicine the H. Lee Moffitt Cancer Center, Tampa, FL d Department of Pathology, Delray Beach Medical Center, Delray Beach, FL e Department of Anatomic Pathology, the H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL b c
a r t i c l e
i n f o
Keywords: Myeloid sarcoma Granulocytic sarcoma Pitfalls Differential Acute myeloid leukemia
a b s t r a c t Myeloid sarcoma involving soft tissue is rare and may present a pathologic diagnostic challenge, particularly when it precedes or coincides with hematological malignancies. Furthermore, it may mimic non-Hodgkin lymphoma, poorly differentiated carcinoma, melanoma, or round blue cell tumors, which is a potential diagnostic pitfall. In addition to a retrospective review of myeloid sarcoma (MS) cases seen at our institution, we describe differential diagnoses, diagnostic pitfalls, and practical approaches to diagnosing soft tissue MS preceding or coinciding with acute myeloid leukemia. Our institutional retrospective review (1999-2011) of MSs identified 12 cases of MS in which there was no known blood or bone marrow involvement at diagnosis. A panel of immunohistochemical stains and/or flow cytometry was reviewed; marker selection was subject to the pathologist's discretion. These tumors were consistently positive for CD117 (9/9), CD43 (7/7), myeloperoxidase (8/10), CD68 (4/5), and CD34 (5/9) by flow cytometry and/or immunohistochemistry. We also described a referral case, which had classic MS morphology and a myelomonocytic immunophenotype including positivity for CD45, lysozyme, and CD117 with supporting molecular information. Based on our institution's experience and review of the literature, we recommend that when the index of suspicion for MS is high, an immunohistochemical stain and/or flow cytometry panel should include CD43, lysozyme, CD117, CD68, CD33, Human Leukocyte Antigen DR (HLA-DR), and myeloperoxidase, in addition to thorough review of the patient's history, cytogenetic studies, and proper discussion with the clinician. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Myeloid sarcoma (MS) is an uncommon, extramedullary proliferation of immature myeloid cells that may be a diagnostic challenge if it precedes acute myeloid leukemia (AML). One facet of this challenge is illustrated, historically, by the variety of names MS has held, including granulocytic sarcoma, chloroma, extramedullary myeloid tumor, and myeloblastoma. Myeloid sarcoma may arise in a background of AML, myelodysplastic syndrome, myeloproliferative neoplasms, and mixed myelodysplastic-myeloproliferative neoplasms [1]. However, it is when MS presents de novo months or even years before the evolution of AML that creates the diagnostic dilemma [1,2]. It has been reported that as many as 27% of MS cases present de novo and that the mean interval between discovery of isolated MS and bone marrow involvement is 10 months [1,3]. The diagnosis of de novo MS is equivalent to the ⁎ Corresponding author at: Department of Anatomic Pathology, Moffitt Cancer Center, 12902 Magnolia Drive, Tampa, FL 33612. Tel.: +1 813 745 4940; fax: +1 813 745 5790. E-mail address: Marilyn.Bui@moffitt.org (M.M. Bui). http://dx.doi.org/10.1016/j.anndiagpath.2014.06.001 1092-9134/© 2014 Elsevier Inc. All rights reserved.
diagnosis of AML. Myeloid sarcoma can mimic other entities histologically, clinically, and radiographically [1,2,4]. The most common locations for involvement are the skin, soft tissue, bone, gastrointestinal tract, or lymph nodes [1,2]. Of note, MS with the t(8;21) (q22;q22) is associated with orbital manifestations in children [1,2,5]. Despite these predilections, MS may be found at any site in the body. Grossly, the tumor is often described as having a green hue, which is how it earned the name “chloroma”. As our understanding of molecular pathology deepens, pathologists may use a more standardized and sophisticated approach to suspected MS. However, these techniques may not be readily available to the general pathologist in community practice. Moreover, the pitfalls surrounding MS preceding AML may ensnare a pathologist who is not in a multidisciplinary setting or is not subspecialized. Although there are many excellent reviews on this subject in the literature, due to the rarity of this entity, we still encounter cases from the community, referred to our comprehensive cancer center, which were initially misdiagnosed. To more readily achieve the correct diagnosis and appropriate management for MS, we provide differential diagnoses, diagnostic pitfalls, and practical approaches for soft
254
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
tissue MS diagnosis when it precedes or coincides with AML based on our experience and a review of the literature. In addition, we provide a prototypical case that was referred to our institution for confirmation. 2. Our experience with MS 2.1. Materials and methods A retrospective review (1999-2011) of soft tissue tumors, including MS, alternatively termed granulocytic/monocytic sarcoma, chloroma, extramedullary myeloid tumor, and myeloblastoma was conducted using our pathology database following our institutional review board protocol. Patient data were retrieved for those with biopsy-confirmed soft tissue MS either preceding blood or marrow involvement or as a solitary lesion. Surgical specimens were fixed in 10% formaldehyde and embedded in paraffin. The paraffin blocks were then sectioned at 5 μm and stained with hematoxylin and eosin (H&E) per standard laboratory procedure. Immunohistochemical stains were ordered, when appropriate, and performed following the below procedure. All initial diagnoses were made by boardcertified hematopathologists and/or anatomical pathologists. Histologic review was performed to confirm the diagnosis for all cases. Immunohistochemical staining was performed using commercially available antibodies, a steam-induced epitope retrieval process, and a Ventana platform (Tucson, AZ) on formalin-fixed, paraffin-embedded sections following the supplier's protocol with appropriate positive and negative controls. Information on the immunohistochemical antibodies is summarized in Table 1. Flow cytometry was performed by a BD FACSCalibur flow cytometer platform (BD Biosciences, San Jose, CA). Clinical, radiologic, and pathologic data were compared with literature reviewed. 2.1.1. Cases reviewed at our institution Cases reviewed at our institution included patients who presented to our institution for initial diagnosis and management as well as patients who had prior diagnoses and were referred for confirmation and/or management. Of the 1200 soft tissue tumors reviewed at our institution, 12 patients (1%) were confirmed to have MS either without blood or marrow involvement at diagnosis (8/12) or as primary presenting lesion with no clinical history of hematopoietic disorder (4/12). There were 6 male patients and 6 female patients, giving a male-to-female ratio of 1:1. The patient's ages ranged from 2 to 67 years, with a mean of 50.8 years. A summary of pertinent clinical and pathologic features is presented in Table 2. The tumor locations included breast (2/12), mediastinum (2/12), extremities (2/12), neck (2/12), spine (1/12), gallbladder (1/12), appendix (1/12), and abdominal wall (1/12). Of 12 patients, 3 (25%) had classic morphological clues of MS on H&E, including effacement of tissue architecture by neutrophilic or promyelocytic myeloblasts with scant cytoplasm, round-oval nuclei, fine chromatin, and small nucleoli. The remaining 9 cases resembled other neoplasms with differential diagnoses including lymphoma, poorly differentiated carcinoma, melanoma, neuroendocrine carcinoma, Ewing sarcoma, neuroblastoma, and Langerhans cell histiocytosis. By immunohistochemistry and/or flow cytometry, the tumors were consistently positive for CD117 (9/9), CD68 (4/5), and CD43 (7/7), when these markers were interrogated. Myeloperoxidase (MPO) was strongly positive in 5 of 10 cases and weakly positive in 3 of 10 cases. CD34 was positive in 5 of 9 cases, and CD56 was strongly positive in 1 case. CD99 was positive in 1 of 3 cases. CD45 was weak or dim by immunohistochemistry or flow cytometry, respectively, in 28.6% (2/7) cases, the other 5 cases being positive. This included 1 case in which poorly differentiated carcinoma was in the differential diagnosis due to the tumor's morphological features. Our immunohistochemical and/or flow cytometric findings are summarized in Fig. 1. Cytogenetics and molecular studies were either not performed or were not available in
our database for all but 1 case in which the tumor had normal chromosomes and was negative for FLT3 and NPM1 mutations. 2.1.2. Prototypical case One example case referred to our institution involves an elderly lady with a history of abdominal pain. The patient was an 85-year-old Peruvian native with no significant family history. She was a lifetime nonsmoker. For 2 months, she had been experiencing intermittent, worsening, abdominal pain that necessitated a visit to her community emergency department. The initial computed tomographic scan was suggestive of acute appendicitis. However, an obstructive soft tissue mass was discovered, intraoperatively, in the area of the small bowel. The mass was resected and, grossly, featured an area of stricture and thickening of the small bowel wall. Histologically, there was infiltration by small- to medium-sized cells with scant, lightly basophilic cytoplasm and gray, round to oval, slightly irregular nuclei with open chromatin and an occasional folded nucleus. The initial
Table 1 Summary of immunohistochemical antibodies used and their relevant staining patterns Antibody
Clone
Dilution Supplier Staining pattern
CD34
QBEnd/10
RTU
VMSa
CD43
L60
RTU
VMSa
CD45
RP2/18
RTU
VMSa
CD56
1B6
RTU
LBb
CD68
KP-1
1:50100
VMSa
CD99
O13
1:500
CIc
CD117
YR145
1:800
CMd
MPO Polyclonal 1:4000 Lysozyme Polyclonal RTU
Dakoe BGf
S-100
VMSa
4C4.9
RTU
Abbreviation: RTU, ready to use. a Ventana Medical Systems, Tucson, AZ. b Leica Biosystems, Wetzlar, Germany. c Covance, Inc, Princeton, NJ. d Cell Marque Corporation, Rocklin, CA. e Dako, Inc, Carpinteria, CA. f BioGenex, Fremont, CA.
Membranous
Specificity
Vascular progenitor cells, endothelial cells, certain leukemic blasts, and certain soft tissue tumors Membranous, T cells, myeloid cells, cytoplasmic subset of B cells, Tand B-cell lymphomas Membranous T and B lymphocytes, monocytes, macrophages, mast cells, weakly on granulocytes Membranous Neurons, astrocytes, Schwann cells, NK cells, subset of activated T lymphocytes Cytoplasmic, Monocyte/ membranous macrophage lineage cells Membranous Ewing sarcoma, some primitive neuroectodermal tumors, and peripheral neuroepitheliomas Cytoplasmic, Interstitial cells of membranous Cajal, germ cells, bone marrow stem cells, melanocytes, breast epithelium, and mast cells Cytoplasmic Myeloid lineage cells Cytoplasmic Myeloid and monocyte/ macrophage lineage cells Cytoplasmic Melanoma, neuronally derived cells, monocyte/ macrophage lineage cells, clear cell sarcoma
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
255
Table 2 Summary of reviewed MS cases Case number
Age (y)
Sex
Preceding (P) or coinciding (C) with AML
Immunophenotype By flow cytometry
By immunohistochemistry
1
43
2
Location
M
C
N/A
60
F
P
N/A
3
2
M
Unknown
N/A
4
49
F
P
5
58
F
P
6
63
F
C
7
44
M
P
CD34− CD117+ HLA-DR+ CD33+ CD117+ CD45 dim HLA-DR+ CD4+ CD11b+ CD11c+ CD13+ CD34+ HLA-DR+ CD117 dim CD56 dim MPO dim TdT− N/A
CD34+ CD43+ CD68+ CD117+ MPO+ CD34− CD43+ CD45+ CD99+ CD117+ MPO+ S-100− CD34− CD45+ CD56− CD68+ CD99− CD117 weakly+ S-100− N/A
8
67
F
P
N/A
9
52
M
P
N/A
10
63
M
C
N/A
11
49
M
P
N/A
12
59
F
P
N/A
Prototypical Case
85
F
C
CD33+ CD34+ CD117+ HLA-DR+ CD45 dim
Mediastinum
Left neck, lymph node
Skin, right leg
Breast
MPO+
Multiple breast masses
N/A
Right breast Left forearm
CD34+ CD43+ CD45+ CD56 focally+ CD68− MPO− CD34+ CD43+ CD117+ MPO+ Cytokeratin− CD34− CD45+ CD68 focally+ Lysozyme+ MPO weakly+ S-100− CD34+ CD43+ CD99− CD117+ MPO− S-100− CD43+ CD45 weakly+ CD68+ Lysozyme+ MPO weakly+ CD43+ CD45+ CD117+ MPO+ Lysozyme+ CD117+ CD43+ MPO+ CD45 weakly+ CD68 weakly+ CD34 weakly+
Mediastinum
Gallbladder
Thoracic spine
Right abdominal wall
Appendix
Left neck
Small intestine
256
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
immunostain panel revealed blasts that were positive for lysozyme and CD117, weak for CD34 and CD45, and negative for CD68 and CD4. The contributing pathologist, who completed fellowship training in hematopathology, made the correct diagnosis: MS with a myelomonocytic immunophenotype. However, because of the rarity of this disease and the lack of concomitant AML, this case was referred to our institution for confirmation. The small bowel mass demonstrated classic MS histology, including sheets of myeloid blasts effacing the surrounding tissue and a predominance of eosinophils in the background (Fig. 2A). Immunohistochemical analysis at our institution demonstrated positivity for lysozyme, CD117 (Fig. 2B), CD43 (Fig. 2C), and MPO (Fig. 2D). The tumor cells were focally positive for CD45, CD68, and CD34. Flow cytometry revealed CD33-, CD34-, CD117-, and HLA-DR–positive myeloblasts with dim CD45 expression (Fig. 2E-H). Fluorescence in situ hybridization (FISH) was positive for inv(16) and trisomy 8 (Fig. 2I). The patient had no evidence of blood or bone marrow involvement by a hematologic malignancy at the time of diagnosis; however, a cutaneous papule was discovered on her leg that was found to be consistent with her abdominal MS. This mass was not sent to our institution for review. The patient received induction chemotherapy with idarubicin and cytosine arabinoside. This was followed by 2 cycles of high-dose cytosine arabinoside. Her treatment was complicated by a right subclavian vein thrombosis and an episode of cryptosporidiosis after induction. Fortunately, she is in complete remission 2 years after initial diagnosis.
3. Discussion Our findings corroborate and strengthen other comprehensive reviews of MS. Myeloid sarcoma may present with a multitude of morphological and immunophenotypical patterns. Furthermore, it has been demonstrated that there may be variability between the phenotypes of biopsied leukemias taken from different sites or at different times, either due to clonal evolution or treatment effect [2,6]. Historically, MS has been subdivided into either granulocytic or monoblastic, based on morphology [2]. Cytogenetics has made this distinction less relevant, and immunophenotyping has shown that MS may have features of any myeloid lineage, including multiple lineage expression within the same tumor [2]. Classic morphological features of MS include effacement of tissue architecture by neutrophilic or promyelocytic myeloblasts with
10 9
Number of cases
8 7 6 5 4 3 2
scant cytoplasm, round-oval nuclei, fine chromatin, and small nucleoli. These features in addition to a clinical history of tumor of unknown origin should raise one's index of suspicion for MS. In these cases, a portion of the specimen should be saved for flow cytometry, cytogenetics, FISH, and/or other molecular studies as indicated. The initial diagnosis of MS with classic morphologic features and immunoreactivity should not be difficult. However, at the time of initial H&E screening, deceptive morphological features of MS combined with the absence of a history of AML may lead the pathologist down the wrong diagnostic path. 4. Differential diagnoses and diagnostic pitfalls 4.1. Evaluation of immunomarkers in MS A number of immunomarkers have been recommended for the diagnosis of MS, yet some, if used inappropriately when presented with a tumor of unknown primary, may lead to a misdiagnosis. CD43 and lysozyme appear to be the most frequently expressed markers discussed in contemporary literature, but they lack specificity for MS [1-3,7-9]. In our review, CD43 and lysozyme were positive in all cases in which those stains were performed. CD68, another commonly attributed marker, is highly sensitive and poorly specific for MS. Klco et al [2] remark that the KP-1 clone of CD68, which our laboratory uses, has good sensitivity. One study demonstrated immunoreactivity in 11 (85%) of 13 cases [2,10]. CD68 was positive in 80% (4/5) of our reviewed cases. Pileri et al [1] have demonstrated that, despite its utility in identifying leukemic blasts, CD34 is inconsistently positive in MS, and Klco et al [2] suggest that HLA-DR may be a more sensitive marker. This fits because monocytic MS may not express CD34. CD45 is an immunostain that is commonly used to work up hematopoietic neoplasms when dealing with a tumor of unknown origin. One must use this marker with caution as MS may exhibit a blastic immunophenotype; thus, CD45 may be dim to negative, as in our prototypical referral case and in 2 of our review cases. Use of CD45 to rule in or rule out hematopoietic differentiation in a tumor of unknown primary may lead to a misdiagnosis, if used in isolation [1,2,7]. However, CD45 should still be used as a screening immunomarker for tumors of unknown origin with the caveat that, if the tumor is negative, but clinical suspicion for hematopoietic malignancy is high, then additional markers for MS should be obtained. CD43 expression, as noted above, is useful not only in supporting the diagnosis of MS but also in the establishment of hematolymphoid lineage in the case of a poorly differentiated malignancy of unknown origin. Regarding the later situation, CD43 may be of greater utility than CD45. CD33, CD117, and MPO are frequently recommended to rule in MS and to rule out non-Hodgkin lymphoma [1,2,7]. In our series, CD117 was positive in 100% (9/9) of cases in which the marker was assessed. However, the lack of CD117 staining does not rule out MS because MS with monocytic differentiation may be negative for the aforementioned markers [2]. A relatively new immunohistochemical stain for CD33 is available (clone PWS44; Novocastra Laboratories, Newcastle upon Tyne, UK) and has shown excellent correlation with flow cytometry for CD33, including intensity of staining [11]. This stain could be useful if fresh tissue for flow cytometry is unavailable. We propose a diagnostic algorithm for the approach to a tumor of unknown primary in which MS enters the differential (Fig. 3).
1 0 Lysozyme
CD68
CD43
CD33
Positive cases
MPO
CD117 HLA-DR
CD45
Total number of cases interrogated by given marker
Fig. 1. Immunohistochemical and flow cytometric results for frequently suggested MS markers [1,2].
4.1.1. Non-Hodgkin lymphoma T-cell lymphomas, in particular, may be difficult to rule out because MS may aberrantly express some T-cell markers such as CD2, CD4, and CD7 in addition to CD45 and CD43 [1,2]. Immunoreactivity for lysozyme, MPO, and CD68 should swing the diagnosis away from a T-cell neoplasm and toward that of MS [1,2]. Our series demonstrated 2 cases with aberrant expression of CD7 and 1 case with aberrant expression of CD4.
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
E A
F
B
G
C
257
D
I
H
J
Fig. 2. Photomicrographs of the small bowel MS from the prototypical referral case. Note the classic features (A, H&E, original magnification ×400) and the immunohistochemical staining patterns for CD117 (B, membranous,°×400), CD43 (C, membranous,°×400), and MPO (D, cytoplasmic, ×400). Flow cytometry results demonstrating a myeloblastic population that is dim for CD45 (E) and positive for CD34 (F and G), CD117 (F and H), CD33 (H), and HLA-DR (G). The separation of the red and green signals indicates inversion of chromosome 16 by FISH (I, ×1000) and the 3 red signals demonstrates trisomy 8 (J, ×1000) in our prototypical patient.
B-cell lymphomas can typically be excluded by the lack of traditional B-cell lineage markers, such as CD20. However, rare cases of MS with t(8;21) may demonstrate aberrant immunoreactivity for PAX-5 and CD79a [2,8]. Although terminal deoxynucleotidyl transferase (TdT) expression was not identified in our patient set, it is not an uncommon finding, especially in very immature MS and AML [1]. TdT expression may raise concern for T- or B-cell acute lymphoblastic leukemia, but lack of additional lymphoid lineage markers would aid in the exclusion of this entity. In addition, it is important to remember that TdT is occasionally expressed in AML, so it stands to reason that it could be observed in MS as well. The exclusion of non-Hodgkin lymphoma should be trivial if MS is considered in the differential diagnosis and the appropriate immunohistochemical or flow cytometric panel is performed. 4.1.2. Poorly differentiated carcinoma If MS presents with large immature cells, scant to moderate cytoplasm, nuclear pleomorphism and bizarre cells, poorly differentiated carcinoma may enter the differential diagnosis. Background fibrosis and necrosis, as seen in one of our review cases, can further obfuscate the identity of the neoplasm (Fig. 4A). In addition, rare reports of aberrant keratin expression in MS have been documented, which may lure the unsuspecting diagnostician off the correct path [1,2]. Implementation of an appropriate immunohistochemical panel should correctly characterize the neoplasm. Again, use of CD45 in isolation to identify an unknown neoplasm as hematopoietic may be a pitfall as this may be weak to negative in MS. 4.1.3. Melanoma Melanoma, often considered the “great mimicker,” can masquerade as MS. In our study as well as others, S-100 is universally negative in MS cells [1,2,7,8]. S-100 protein and other melanoma makers may
be indicated if the clinical suspicion remains high and the tumor is poorly differentiated. 4.1.4. Neuroendocrine tumor Neuroendocrine tumors may be in the differential, and in our series, CD56 expression was seen, at least focally, in 2 cases. CD56 is a neural cell adhesion molecule that is commonly expressed by neuroendocrine neoplasms but is also known to be expressed by natural killer cells and occasionally aberrantly expressed by neoplastic monocytic cells. A limited panel with CD56 positivity and cytological features similar to those of typical neuroendocrine tumors may be a diagnostic pitfall, but lack of cytokeratin, chromogranin, and synaptophysin staining coupled with expression of recommended MS markers will aid in the exclusion of neuroendocrine neoplasms. 4.1.5. Blue round cell tumor On occasion, small round blue cell morphology can be found in MS. In younger patients, entities such as Ewing sarcoma or neuroblastoma may enter the differential. CD99, the traditional Ewing sarcoma marker, may be seen in MS, but Ewing sarcoma should lack all immature myeloid markers [1]. One of our reviewed cases had CD99 immunoreactivity in addition to nonclassic MS morphology, including areas with small cells and crush artifact (Fig. 4B and C). If clinical suspicion is high, FISH may be performed to investigate for t(11; 22) EWS/FLI or other Ewing sarcoma– related translocations. Cytogenetic analysis has the additional benefit of potentially identifying changes associated with myeloid disorders even if a t(11; 22) translocation is not identified. 4.1.6. Germ cell tumor Although germ cell tumors tend to have more cytoplasm than classic MS, discohesion and positivity for CD117 may mislead the
258
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
N
N
N
Fig. 3. Proposed surgical pathology algorithm, using immunohistochemistry, for the approach to a tumor of unknown origin in which there is clinical suspicion for MS.
pathologist. This is particularly true if the tumor occurs in a site where germ cell tumors are common (ie, midline, testes). 4.1.7. Epithelioid sarcoma Epithelioid sarcoma typically presents on the distal upper extremities in which case it tends to exhibit a nodular pattern compose of eosinophilic, epithelioid, and spindled cells. Myeloid sarcoma with larger, more epithelioid cells could mimic this histomorphology, and epithelioid sarcoma may be immunoreactive for CD34. However, epithelioid sarcoma should demonstrate positivity for high- and low-molecular-weight cytokeratins. 4.1.8. Monocytic neoplasm Myeloid sarcoma with monocytic differentiation would demonstrate medium-sized cells with ample, somewhat basophilic cytoplasm containing occasional fine azurophilic granules. Their nuclei are typically round with lacy chromatin and a single prominent nucleolus
(Fig. 4D). These features may distract the diagnostician from the correct diagnosis as they can resemble other entities such as Langerhans cell histiocytosis. Differentiating these 2 entities may be difficult on H&E alone because the classic coffee bean–shaped nuclei of Langerhans cell histiocytosis and occasional eosinophilpredominant background can be seen in the monocytic variant of MS. Fortunately, Langerhans cell histiocytosis has coexpression of S-100 and CD1a and is characterized as having recognizable Birbeck granules by electron microscopy. Myeloid sarcoma tends to be CD1a and S-l00 negative allowing for a relative ease in distinction by immunophenotyping [1]. Although mastocytosis commonly presents with skin involvement, systemic mastocytosis must also be a consideration as the azurophilic granules of monocytic MS may be mistaken for mastocytic granules. Mastocytosis may also be immunoreactive for lysozyme, HLA-DR, CD33, and CD68, which may complicate the diagnosis if flow cytometry, alone, is used.
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
A
B
C
D
259
Fig. 4. Photomicrographs of nonclassical MS including epithelioid features with background fibrosis (A, H&E, original magnification ×600, case 8), round blue cell features (B, H&E, ×600, case 3), membranous immunoreactivity for CD99 (C, ×600, case 2), and monocytic features (D, H&E, ×600).
It is worth noting again that monocytic sarcoma may lack expression of some of the commonly described markers such as CD34 and CD117. However, they should still retain expression of CD68, CD43, CD33, and lysozyme, as in classic MS [2]. Per Lau et al [10], CD163, may be sensitive for monocytic sarcoma [2]. As mentioned above, CD56 or CD4 may be expressed and thus mimic a neuroendocrine tumor or T-cell lymphoma, respectively [2]. 4.1.9. Blastic plasmacytoid dendritic cell neoplasm Blastic plasmacytoid dendritic cell neoplasm, although rare, is a hematologic neoplasm that must be considered in the differential diagnosis of MS. It may be morphologically indistinguishable from MS as it frequently has medium to large atypical cells with fine chromatin, nucleoli, and irregular nuclear contours. Blastic plasmacytoid dendritic cell neoplasm most often affects the skin or soft tissue, similar to MS, so tumor location does not preclude the diagnosis of MS. Furthermore, 10% to 20% of blastic plasmacytoid dendritic cell neoplasms may develop myeloid neoplasms such as AML. In these cases, systematic investigation and close follow-up are warranted. Some immunostains for MS such as CD34, CD117, CD33, CD43, and CD163 have limited utility in differentiating these entities as there is significant overlap in expression. Myeloid sarcoma with monocytic differentiation is problematic as it is negative for CD34 and CD117 and positive for CD4, CD43, CD56, and CD68, which mimics blastic plasmacytoid dendritic cell neoplasm. Of note, the neoplastic cells of blastic plasmacytoid dendritic cell neoplasm are also partially positive for TdT and express CD2, CD7, and CD38. Two immunohistochemical panels have been suggested to aid differentiation [12]. The first larger panel includes plasmacytoid dendritic
cell markers (CD4, CD56, CD123, and TCL-1), lysozyme, MPO, and myxovirus A, a new research marker for an interferon α–inducible protein [12]. Myeloid sarcoma was associated with positive staining for lysozyme or MPO, as expected, and was negative staining for the other markers [12]. Blastic plasmacytoid dendritic cell neoplasm consistently had the opposite staining pattern [12]. This study also suggested a simplified 2-antibody panel with a scoring scheme involving CD56 and TCL-1. The panel was scored as an aggregate where a value of “1” was assigned to each positive result, “−1” to each negative result, and “0” to a marker that was not performed or equivocal. A score greater than or equal to zero was associated with blastic plasmacytoid dendritic cell neoplasm [12]. 5. Molecular and cytogenetic patterns in MS It has been estimated that 55% of MS may contain an cytogenetic abnormalities within the tumor cells and that de novo MS may lack cytogenetic abnormalities [7]. Interestingly, 1 study noted absence of cytogenetic abnormalities in all monoblastic MS cases, whereas only 1 (12.5%) of 8 nonmonoblastic cases had normal cytogenetics [7]. Although not well understood in MS, a study reported that NPM1 gene mutations occur in 50% to 60% of AMLs without cytogenetic abnormalities [13]. When they do occur, the most commonly reported cytogenetic abnormalities are trisomy 8 and inv(16) (Fig. 2I and J) [2,7,9], which occurred in our prototypical case. Trisomy 8 has been associated with a lower median survival by 1 author [14]. Where available, all of our review cases had normal cytogenetics, which matched with what Alexiev et al [7] has described regarding de novo MS. However, our example referral case was found to have inv(16) and trisomy 8 and did not have AML at presentation. Further study is required to elucidate the molecular anomalies observed in MS and their clinical impact.
260
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
6. Conclusions One must maintain a high index of suspicion for MS if a soft tissue tumor resembling non-Hodgkin lymphoma, poorly differentiated carcinoma, melanoma, or round blue cell tumor is encountered, as MS may precede or coincide with AML and may have nonclassic morphological features. When evaluating any tumor of unknown etiology, it is wise to save portions of fresh tumor for possible flow cytometric analysis. Based on our experience and a review of the literature, useful immunohistochemical markers for MS include CD43, lysozyme, CD68, CD117, and MPO. Per our review of the literature, CD163 may also be useful if the case has histomorphologic features suggestive of monocytic sarcoma [2]. A supplemental, minimal flow cytometry panel that includes HLA-DR, CD33, MPO, CD117, CD13, CD14, and CD45 should be performed if fresh tissue is available [1,2]. In our experience, special care must be taken when interpreting certain nonspecific immunomarkers because, if interpreted in a vacuum, they may lead the diagnostician astray. In these situations, thorough evaluation of the patient's clinical history and appropriate communication with the clinician is recommended. Utilization of additional ancillary testing, such as cytogenetic analysis and bone marrow biopsy, may also improve the diagnostic yield. Maintaining MS in the differential is important for the community pathologist as this entity is a mimicker, may have an aberrant immunophenotype, and may present de novo. This was true for our prototypical case example. The outside, “front-line,” pathologist recognized this entity and sent it to our institution for confirmation, using our subspecialty expertise. Achieving an early and accurate diagnosis is critical for timely and appropriate therapy and thus better patient outcomes. Acknowledgment The authors thank Ms Andrea Bumpus' assistance in submitting this abstract. The authors are also grateful to the Histology and Flow
Cytometry Laboratories at Moffitt Cancer Center for their excellent technical support. References [1] Pileri SA, Ascani S, Cox MC, Campidelli C, Bacci F, Piccioli M, et al. Myeloid sarcoma: clinico-pathologic, phenotypic and cytogenetic analysis of 92 adult patients. Leukemia 2007;21:340–50. [2] Klco JM, Welch JS, Nguyen TT, Hurley MY, Kreisel FH, Hassan A, et al. State of the art in myeloid sarcoma. Int J Lab Hematol 2011;33:555–65. [3] Neiman RS, Barcos M, Berard C, Bonner H, Mann R, Rydell RE, et al. Granulocytic sarcoma: a clinicopathologic study of 61 biopsied cases. Cancer 1981;48:1426–37. [4] Shah RS, Shin RK, Castellani RJ. Granulocytic sarcoma mimicking HSV encephalitis. Neurologist 2010;16:319–21. [5] Schwyzer R, Sherman GG, Cohn RJ, Poole JE, Willem P. Granulocytic sarcoma in children with acute myeloblastic leukemia and t(8;21). Med Pediatr Oncol 1998;31:144–9. [6] Macedo A, San Miguel JF, Vidriales MB, López-Berges MC, García-Marcos MA, Gonzalez M, et al. Phenotypic changes in acute myeloid leukaemia: implications in the detection of minimal residual disease. J Clin Pathol 1996;49:15–8. [7] Alexiev BA, Wang W, Ning Y, Chumsri S, Gojo I, Rodgers WH, et al. Myeloid sarcomas: a histologic, immunohistochemical, and cytogenetic study. Diagn Pathol 2007;2:42–9. [8] Campidelli C, Agostinelli C, Stitson R, Pileri SA. Myeloid sarcoma: extramedullary manifestation of myeloid disorders. Am J Clin Pathol 2009;132:426–37. [9] Deeb G, Baer MR, Gaile DP, Sait SN, Barcos M, Wetzler M, et al. Genomic profiling of myeloid sarcoma by array comparative genomic hybridization. Genes Chromosom Cancer 2005;44:373–83. [10] Lau SK, Chu PG, Weiss LM. CD163: a specific marker of macrophages in paraffinembedded tissue samples. Am J Clin Pathol 2004;122:794–801. [11] Hoyer JD, Grogg KL, Hanson CA, Gamez JD, Dogan A. CD33 detection by immunohistochemistry in paraffin-embedded tissues a new antibody shows excellent specificity and sensitivity for cells of myelomonocytic lineage. Am J Clin Pathol 2008;129:316–23. [12] Sangle NA, Schmidt RL, Patel JL, Jeffrey Medeiros L, Agarwal AM, Perkins SL, et al. Optimized immunohistochemical panel to differentiate myeloid sarcoma from blastic plasmacytoid dendritic cell neoplasm. Mod Pathol 2014:1–7. [13] Chen W, Rassidakis GZ, Medeiros LJ. Nucleophosmin gene mutations in acute myeloid leukemia. Arch Pathol Lab Med 2006;130:1687–92. [14] Tsimberidou AM, Kantarjian HM, Estey E, Cortes JE, Verstovsek S, Faderl S, et al. Outcome in patients with nonleukemic granulocytic sarcoma treated with chemotherapy with or without radiotherapy. Leukemia 2003;17:1100–3.
Contents lists available at ScienceDirect
Annals of Diagnostic Pathology
Review Article
A practical approach to diagnose soft tissue myeloid sarcoma preceding or coinciding with acute myeloid leukemia Robert P. Seifert, MD a, William Bulkeley III, MD b, Ling Zhang, MD a, c, Manuel Menes, MD d, Marilyn M. Bui, MD, PhD a, e,⁎ a
Department of Pathology and Cell Biology, University of South Florida Morsani College of Medicine, Tampa, FL James A. Haley Veterans' Hospital, Department of Pathology and Laboratory Medicine, Tampa, FL Department of Hematopathology and Laboratory Medicine the H. Lee Moffitt Cancer Center, Tampa, FL d Department of Pathology, Delray Beach Medical Center, Delray Beach, FL e Department of Anatomic Pathology, the H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL b c
a r t i c l e
i n f o
Keywords: Myeloid sarcoma Granulocytic sarcoma Pitfalls Differential Acute myeloid leukemia
a b s t r a c t Myeloid sarcoma involving soft tissue is rare and may present a pathologic diagnostic challenge, particularly when it precedes or coincides with hematological malignancies. Furthermore, it may mimic non-Hodgkin lymphoma, poorly differentiated carcinoma, melanoma, or round blue cell tumors, which is a potential diagnostic pitfall. In addition to a retrospective review of myeloid sarcoma (MS) cases seen at our institution, we describe differential diagnoses, diagnostic pitfalls, and practical approaches to diagnosing soft tissue MS preceding or coinciding with acute myeloid leukemia. Our institutional retrospective review (1999-2011) of MSs identified 12 cases of MS in which there was no known blood or bone marrow involvement at diagnosis. A panel of immunohistochemical stains and/or flow cytometry was reviewed; marker selection was subject to the pathologist's discretion. These tumors were consistently positive for CD117 (9/9), CD43 (7/7), myeloperoxidase (8/10), CD68 (4/5), and CD34 (5/9) by flow cytometry and/or immunohistochemistry. We also described a referral case, which had classic MS morphology and a myelomonocytic immunophenotype including positivity for CD45, lysozyme, and CD117 with supporting molecular information. Based on our institution's experience and review of the literature, we recommend that when the index of suspicion for MS is high, an immunohistochemical stain and/or flow cytometry panel should include CD43, lysozyme, CD117, CD68, CD33, Human Leukocyte Antigen DR (HLA-DR), and myeloperoxidase, in addition to thorough review of the patient's history, cytogenetic studies, and proper discussion with the clinician. © 2014 Elsevier Inc. All rights reserved.
1. Introduction Myeloid sarcoma (MS) is an uncommon, extramedullary proliferation of immature myeloid cells that may be a diagnostic challenge if it precedes acute myeloid leukemia (AML). One facet of this challenge is illustrated, historically, by the variety of names MS has held, including granulocytic sarcoma, chloroma, extramedullary myeloid tumor, and myeloblastoma. Myeloid sarcoma may arise in a background of AML, myelodysplastic syndrome, myeloproliferative neoplasms, and mixed myelodysplastic-myeloproliferative neoplasms [1]. However, it is when MS presents de novo months or even years before the evolution of AML that creates the diagnostic dilemma [1,2]. It has been reported that as many as 27% of MS cases present de novo and that the mean interval between discovery of isolated MS and bone marrow involvement is 10 months [1,3]. The diagnosis of de novo MS is equivalent to the ⁎ Corresponding author at: Department of Anatomic Pathology, Moffitt Cancer Center, 12902 Magnolia Drive, Tampa, FL 33612. Tel.: +1 813 745 4940; fax: +1 813 745 5790. E-mail address: Marilyn.Bui@moffitt.org (M.M. Bui). http://dx.doi.org/10.1016/j.anndiagpath.2014.06.001 1092-9134/© 2014 Elsevier Inc. All rights reserved.
diagnosis of AML. Myeloid sarcoma can mimic other entities histologically, clinically, and radiographically [1,2,4]. The most common locations for involvement are the skin, soft tissue, bone, gastrointestinal tract, or lymph nodes [1,2]. Of note, MS with the t(8;21) (q22;q22) is associated with orbital manifestations in children [1,2,5]. Despite these predilections, MS may be found at any site in the body. Grossly, the tumor is often described as having a green hue, which is how it earned the name “chloroma”. As our understanding of molecular pathology deepens, pathologists may use a more standardized and sophisticated approach to suspected MS. However, these techniques may not be readily available to the general pathologist in community practice. Moreover, the pitfalls surrounding MS preceding AML may ensnare a pathologist who is not in a multidisciplinary setting or is not subspecialized. Although there are many excellent reviews on this subject in the literature, due to the rarity of this entity, we still encounter cases from the community, referred to our comprehensive cancer center, which were initially misdiagnosed. To more readily achieve the correct diagnosis and appropriate management for MS, we provide differential diagnoses, diagnostic pitfalls, and practical approaches for soft
254
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
tissue MS diagnosis when it precedes or coincides with AML based on our experience and a review of the literature. In addition, we provide a prototypical case that was referred to our institution for confirmation. 2. Our experience with MS 2.1. Materials and methods A retrospective review (1999-2011) of soft tissue tumors, including MS, alternatively termed granulocytic/monocytic sarcoma, chloroma, extramedullary myeloid tumor, and myeloblastoma was conducted using our pathology database following our institutional review board protocol. Patient data were retrieved for those with biopsy-confirmed soft tissue MS either preceding blood or marrow involvement or as a solitary lesion. Surgical specimens were fixed in 10% formaldehyde and embedded in paraffin. The paraffin blocks were then sectioned at 5 μm and stained with hematoxylin and eosin (H&E) per standard laboratory procedure. Immunohistochemical stains were ordered, when appropriate, and performed following the below procedure. All initial diagnoses were made by boardcertified hematopathologists and/or anatomical pathologists. Histologic review was performed to confirm the diagnosis for all cases. Immunohistochemical staining was performed using commercially available antibodies, a steam-induced epitope retrieval process, and a Ventana platform (Tucson, AZ) on formalin-fixed, paraffin-embedded sections following the supplier's protocol with appropriate positive and negative controls. Information on the immunohistochemical antibodies is summarized in Table 1. Flow cytometry was performed by a BD FACSCalibur flow cytometer platform (BD Biosciences, San Jose, CA). Clinical, radiologic, and pathologic data were compared with literature reviewed. 2.1.1. Cases reviewed at our institution Cases reviewed at our institution included patients who presented to our institution for initial diagnosis and management as well as patients who had prior diagnoses and were referred for confirmation and/or management. Of the 1200 soft tissue tumors reviewed at our institution, 12 patients (1%) were confirmed to have MS either without blood or marrow involvement at diagnosis (8/12) or as primary presenting lesion with no clinical history of hematopoietic disorder (4/12). There were 6 male patients and 6 female patients, giving a male-to-female ratio of 1:1. The patient's ages ranged from 2 to 67 years, with a mean of 50.8 years. A summary of pertinent clinical and pathologic features is presented in Table 2. The tumor locations included breast (2/12), mediastinum (2/12), extremities (2/12), neck (2/12), spine (1/12), gallbladder (1/12), appendix (1/12), and abdominal wall (1/12). Of 12 patients, 3 (25%) had classic morphological clues of MS on H&E, including effacement of tissue architecture by neutrophilic or promyelocytic myeloblasts with scant cytoplasm, round-oval nuclei, fine chromatin, and small nucleoli. The remaining 9 cases resembled other neoplasms with differential diagnoses including lymphoma, poorly differentiated carcinoma, melanoma, neuroendocrine carcinoma, Ewing sarcoma, neuroblastoma, and Langerhans cell histiocytosis. By immunohistochemistry and/or flow cytometry, the tumors were consistently positive for CD117 (9/9), CD68 (4/5), and CD43 (7/7), when these markers were interrogated. Myeloperoxidase (MPO) was strongly positive in 5 of 10 cases and weakly positive in 3 of 10 cases. CD34 was positive in 5 of 9 cases, and CD56 was strongly positive in 1 case. CD99 was positive in 1 of 3 cases. CD45 was weak or dim by immunohistochemistry or flow cytometry, respectively, in 28.6% (2/7) cases, the other 5 cases being positive. This included 1 case in which poorly differentiated carcinoma was in the differential diagnosis due to the tumor's morphological features. Our immunohistochemical and/or flow cytometric findings are summarized in Fig. 1. Cytogenetics and molecular studies were either not performed or were not available in
our database for all but 1 case in which the tumor had normal chromosomes and was negative for FLT3 and NPM1 mutations. 2.1.2. Prototypical case One example case referred to our institution involves an elderly lady with a history of abdominal pain. The patient was an 85-year-old Peruvian native with no significant family history. She was a lifetime nonsmoker. For 2 months, she had been experiencing intermittent, worsening, abdominal pain that necessitated a visit to her community emergency department. The initial computed tomographic scan was suggestive of acute appendicitis. However, an obstructive soft tissue mass was discovered, intraoperatively, in the area of the small bowel. The mass was resected and, grossly, featured an area of stricture and thickening of the small bowel wall. Histologically, there was infiltration by small- to medium-sized cells with scant, lightly basophilic cytoplasm and gray, round to oval, slightly irregular nuclei with open chromatin and an occasional folded nucleus. The initial
Table 1 Summary of immunohistochemical antibodies used and their relevant staining patterns Antibody
Clone
Dilution Supplier Staining pattern
CD34
QBEnd/10
RTU
VMSa
CD43
L60
RTU
VMSa
CD45
RP2/18
RTU
VMSa
CD56
1B6
RTU
LBb
CD68
KP-1
1:50100
VMSa
CD99
O13
1:500
CIc
CD117
YR145
1:800
CMd
MPO Polyclonal 1:4000 Lysozyme Polyclonal RTU
Dakoe BGf
S-100
VMSa
4C4.9
RTU
Abbreviation: RTU, ready to use. a Ventana Medical Systems, Tucson, AZ. b Leica Biosystems, Wetzlar, Germany. c Covance, Inc, Princeton, NJ. d Cell Marque Corporation, Rocklin, CA. e Dako, Inc, Carpinteria, CA. f BioGenex, Fremont, CA.
Membranous
Specificity
Vascular progenitor cells, endothelial cells, certain leukemic blasts, and certain soft tissue tumors Membranous, T cells, myeloid cells, cytoplasmic subset of B cells, Tand B-cell lymphomas Membranous T and B lymphocytes, monocytes, macrophages, mast cells, weakly on granulocytes Membranous Neurons, astrocytes, Schwann cells, NK cells, subset of activated T lymphocytes Cytoplasmic, Monocyte/ membranous macrophage lineage cells Membranous Ewing sarcoma, some primitive neuroectodermal tumors, and peripheral neuroepitheliomas Cytoplasmic, Interstitial cells of membranous Cajal, germ cells, bone marrow stem cells, melanocytes, breast epithelium, and mast cells Cytoplasmic Myeloid lineage cells Cytoplasmic Myeloid and monocyte/ macrophage lineage cells Cytoplasmic Melanoma, neuronally derived cells, monocyte/ macrophage lineage cells, clear cell sarcoma
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
255
Table 2 Summary of reviewed MS cases Case number
Age (y)
Sex
Preceding (P) or coinciding (C) with AML
Immunophenotype By flow cytometry
By immunohistochemistry
1
43
2
Location
M
C
N/A
60
F
P
N/A
3
2
M
Unknown
N/A
4
49
F
P
5
58
F
P
6
63
F
C
7
44
M
P
CD34− CD117+ HLA-DR+ CD33+ CD117+ CD45 dim HLA-DR+ CD4+ CD11b+ CD11c+ CD13+ CD34+ HLA-DR+ CD117 dim CD56 dim MPO dim TdT− N/A
CD34+ CD43+ CD68+ CD117+ MPO+ CD34− CD43+ CD45+ CD99+ CD117+ MPO+ S-100− CD34− CD45+ CD56− CD68+ CD99− CD117 weakly+ S-100− N/A
8
67
F
P
N/A
9
52
M
P
N/A
10
63
M
C
N/A
11
49
M
P
N/A
12
59
F
P
N/A
Prototypical Case
85
F
C
CD33+ CD34+ CD117+ HLA-DR+ CD45 dim
Mediastinum
Left neck, lymph node
Skin, right leg
Breast
MPO+
Multiple breast masses
N/A
Right breast Left forearm
CD34+ CD43+ CD45+ CD56 focally+ CD68− MPO− CD34+ CD43+ CD117+ MPO+ Cytokeratin− CD34− CD45+ CD68 focally+ Lysozyme+ MPO weakly+ S-100− CD34+ CD43+ CD99− CD117+ MPO− S-100− CD43+ CD45 weakly+ CD68+ Lysozyme+ MPO weakly+ CD43+ CD45+ CD117+ MPO+ Lysozyme+ CD117+ CD43+ MPO+ CD45 weakly+ CD68 weakly+ CD34 weakly+
Mediastinum
Gallbladder
Thoracic spine
Right abdominal wall
Appendix
Left neck
Small intestine
256
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
immunostain panel revealed blasts that were positive for lysozyme and CD117, weak for CD34 and CD45, and negative for CD68 and CD4. The contributing pathologist, who completed fellowship training in hematopathology, made the correct diagnosis: MS with a myelomonocytic immunophenotype. However, because of the rarity of this disease and the lack of concomitant AML, this case was referred to our institution for confirmation. The small bowel mass demonstrated classic MS histology, including sheets of myeloid blasts effacing the surrounding tissue and a predominance of eosinophils in the background (Fig. 2A). Immunohistochemical analysis at our institution demonstrated positivity for lysozyme, CD117 (Fig. 2B), CD43 (Fig. 2C), and MPO (Fig. 2D). The tumor cells were focally positive for CD45, CD68, and CD34. Flow cytometry revealed CD33-, CD34-, CD117-, and HLA-DR–positive myeloblasts with dim CD45 expression (Fig. 2E-H). Fluorescence in situ hybridization (FISH) was positive for inv(16) and trisomy 8 (Fig. 2I). The patient had no evidence of blood or bone marrow involvement by a hematologic malignancy at the time of diagnosis; however, a cutaneous papule was discovered on her leg that was found to be consistent with her abdominal MS. This mass was not sent to our institution for review. The patient received induction chemotherapy with idarubicin and cytosine arabinoside. This was followed by 2 cycles of high-dose cytosine arabinoside. Her treatment was complicated by a right subclavian vein thrombosis and an episode of cryptosporidiosis after induction. Fortunately, she is in complete remission 2 years after initial diagnosis.
3. Discussion Our findings corroborate and strengthen other comprehensive reviews of MS. Myeloid sarcoma may present with a multitude of morphological and immunophenotypical patterns. Furthermore, it has been demonstrated that there may be variability between the phenotypes of biopsied leukemias taken from different sites or at different times, either due to clonal evolution or treatment effect [2,6]. Historically, MS has been subdivided into either granulocytic or monoblastic, based on morphology [2]. Cytogenetics has made this distinction less relevant, and immunophenotyping has shown that MS may have features of any myeloid lineage, including multiple lineage expression within the same tumor [2]. Classic morphological features of MS include effacement of tissue architecture by neutrophilic or promyelocytic myeloblasts with
10 9
Number of cases
8 7 6 5 4 3 2
scant cytoplasm, round-oval nuclei, fine chromatin, and small nucleoli. These features in addition to a clinical history of tumor of unknown origin should raise one's index of suspicion for MS. In these cases, a portion of the specimen should be saved for flow cytometry, cytogenetics, FISH, and/or other molecular studies as indicated. The initial diagnosis of MS with classic morphologic features and immunoreactivity should not be difficult. However, at the time of initial H&E screening, deceptive morphological features of MS combined with the absence of a history of AML may lead the pathologist down the wrong diagnostic path. 4. Differential diagnoses and diagnostic pitfalls 4.1. Evaluation of immunomarkers in MS A number of immunomarkers have been recommended for the diagnosis of MS, yet some, if used inappropriately when presented with a tumor of unknown primary, may lead to a misdiagnosis. CD43 and lysozyme appear to be the most frequently expressed markers discussed in contemporary literature, but they lack specificity for MS [1-3,7-9]. In our review, CD43 and lysozyme were positive in all cases in which those stains were performed. CD68, another commonly attributed marker, is highly sensitive and poorly specific for MS. Klco et al [2] remark that the KP-1 clone of CD68, which our laboratory uses, has good sensitivity. One study demonstrated immunoreactivity in 11 (85%) of 13 cases [2,10]. CD68 was positive in 80% (4/5) of our reviewed cases. Pileri et al [1] have demonstrated that, despite its utility in identifying leukemic blasts, CD34 is inconsistently positive in MS, and Klco et al [2] suggest that HLA-DR may be a more sensitive marker. This fits because monocytic MS may not express CD34. CD45 is an immunostain that is commonly used to work up hematopoietic neoplasms when dealing with a tumor of unknown origin. One must use this marker with caution as MS may exhibit a blastic immunophenotype; thus, CD45 may be dim to negative, as in our prototypical referral case and in 2 of our review cases. Use of CD45 to rule in or rule out hematopoietic differentiation in a tumor of unknown primary may lead to a misdiagnosis, if used in isolation [1,2,7]. However, CD45 should still be used as a screening immunomarker for tumors of unknown origin with the caveat that, if the tumor is negative, but clinical suspicion for hematopoietic malignancy is high, then additional markers for MS should be obtained. CD43 expression, as noted above, is useful not only in supporting the diagnosis of MS but also in the establishment of hematolymphoid lineage in the case of a poorly differentiated malignancy of unknown origin. Regarding the later situation, CD43 may be of greater utility than CD45. CD33, CD117, and MPO are frequently recommended to rule in MS and to rule out non-Hodgkin lymphoma [1,2,7]. In our series, CD117 was positive in 100% (9/9) of cases in which the marker was assessed. However, the lack of CD117 staining does not rule out MS because MS with monocytic differentiation may be negative for the aforementioned markers [2]. A relatively new immunohistochemical stain for CD33 is available (clone PWS44; Novocastra Laboratories, Newcastle upon Tyne, UK) and has shown excellent correlation with flow cytometry for CD33, including intensity of staining [11]. This stain could be useful if fresh tissue for flow cytometry is unavailable. We propose a diagnostic algorithm for the approach to a tumor of unknown primary in which MS enters the differential (Fig. 3).
1 0 Lysozyme
CD68
CD43
CD33
Positive cases
MPO
CD117 HLA-DR
CD45
Total number of cases interrogated by given marker
Fig. 1. Immunohistochemical and flow cytometric results for frequently suggested MS markers [1,2].
4.1.1. Non-Hodgkin lymphoma T-cell lymphomas, in particular, may be difficult to rule out because MS may aberrantly express some T-cell markers such as CD2, CD4, and CD7 in addition to CD45 and CD43 [1,2]. Immunoreactivity for lysozyme, MPO, and CD68 should swing the diagnosis away from a T-cell neoplasm and toward that of MS [1,2]. Our series demonstrated 2 cases with aberrant expression of CD7 and 1 case with aberrant expression of CD4.
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
E A
F
B
G
C
257
D
I
H
J
Fig. 2. Photomicrographs of the small bowel MS from the prototypical referral case. Note the classic features (A, H&E, original magnification ×400) and the immunohistochemical staining patterns for CD117 (B, membranous,°×400), CD43 (C, membranous,°×400), and MPO (D, cytoplasmic, ×400). Flow cytometry results demonstrating a myeloblastic population that is dim for CD45 (E) and positive for CD34 (F and G), CD117 (F and H), CD33 (H), and HLA-DR (G). The separation of the red and green signals indicates inversion of chromosome 16 by FISH (I, ×1000) and the 3 red signals demonstrates trisomy 8 (J, ×1000) in our prototypical patient.
B-cell lymphomas can typically be excluded by the lack of traditional B-cell lineage markers, such as CD20. However, rare cases of MS with t(8;21) may demonstrate aberrant immunoreactivity for PAX-5 and CD79a [2,8]. Although terminal deoxynucleotidyl transferase (TdT) expression was not identified in our patient set, it is not an uncommon finding, especially in very immature MS and AML [1]. TdT expression may raise concern for T- or B-cell acute lymphoblastic leukemia, but lack of additional lymphoid lineage markers would aid in the exclusion of this entity. In addition, it is important to remember that TdT is occasionally expressed in AML, so it stands to reason that it could be observed in MS as well. The exclusion of non-Hodgkin lymphoma should be trivial if MS is considered in the differential diagnosis and the appropriate immunohistochemical or flow cytometric panel is performed. 4.1.2. Poorly differentiated carcinoma If MS presents with large immature cells, scant to moderate cytoplasm, nuclear pleomorphism and bizarre cells, poorly differentiated carcinoma may enter the differential diagnosis. Background fibrosis and necrosis, as seen in one of our review cases, can further obfuscate the identity of the neoplasm (Fig. 4A). In addition, rare reports of aberrant keratin expression in MS have been documented, which may lure the unsuspecting diagnostician off the correct path [1,2]. Implementation of an appropriate immunohistochemical panel should correctly characterize the neoplasm. Again, use of CD45 in isolation to identify an unknown neoplasm as hematopoietic may be a pitfall as this may be weak to negative in MS. 4.1.3. Melanoma Melanoma, often considered the “great mimicker,” can masquerade as MS. In our study as well as others, S-100 is universally negative in MS cells [1,2,7,8]. S-100 protein and other melanoma makers may
be indicated if the clinical suspicion remains high and the tumor is poorly differentiated. 4.1.4. Neuroendocrine tumor Neuroendocrine tumors may be in the differential, and in our series, CD56 expression was seen, at least focally, in 2 cases. CD56 is a neural cell adhesion molecule that is commonly expressed by neuroendocrine neoplasms but is also known to be expressed by natural killer cells and occasionally aberrantly expressed by neoplastic monocytic cells. A limited panel with CD56 positivity and cytological features similar to those of typical neuroendocrine tumors may be a diagnostic pitfall, but lack of cytokeratin, chromogranin, and synaptophysin staining coupled with expression of recommended MS markers will aid in the exclusion of neuroendocrine neoplasms. 4.1.5. Blue round cell tumor On occasion, small round blue cell morphology can be found in MS. In younger patients, entities such as Ewing sarcoma or neuroblastoma may enter the differential. CD99, the traditional Ewing sarcoma marker, may be seen in MS, but Ewing sarcoma should lack all immature myeloid markers [1]. One of our reviewed cases had CD99 immunoreactivity in addition to nonclassic MS morphology, including areas with small cells and crush artifact (Fig. 4B and C). If clinical suspicion is high, FISH may be performed to investigate for t(11; 22) EWS/FLI or other Ewing sarcoma– related translocations. Cytogenetic analysis has the additional benefit of potentially identifying changes associated with myeloid disorders even if a t(11; 22) translocation is not identified. 4.1.6. Germ cell tumor Although germ cell tumors tend to have more cytoplasm than classic MS, discohesion and positivity for CD117 may mislead the
258
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
N
N
N
Fig. 3. Proposed surgical pathology algorithm, using immunohistochemistry, for the approach to a tumor of unknown origin in which there is clinical suspicion for MS.
pathologist. This is particularly true if the tumor occurs in a site where germ cell tumors are common (ie, midline, testes). 4.1.7. Epithelioid sarcoma Epithelioid sarcoma typically presents on the distal upper extremities in which case it tends to exhibit a nodular pattern compose of eosinophilic, epithelioid, and spindled cells. Myeloid sarcoma with larger, more epithelioid cells could mimic this histomorphology, and epithelioid sarcoma may be immunoreactive for CD34. However, epithelioid sarcoma should demonstrate positivity for high- and low-molecular-weight cytokeratins. 4.1.8. Monocytic neoplasm Myeloid sarcoma with monocytic differentiation would demonstrate medium-sized cells with ample, somewhat basophilic cytoplasm containing occasional fine azurophilic granules. Their nuclei are typically round with lacy chromatin and a single prominent nucleolus
(Fig. 4D). These features may distract the diagnostician from the correct diagnosis as they can resemble other entities such as Langerhans cell histiocytosis. Differentiating these 2 entities may be difficult on H&E alone because the classic coffee bean–shaped nuclei of Langerhans cell histiocytosis and occasional eosinophilpredominant background can be seen in the monocytic variant of MS. Fortunately, Langerhans cell histiocytosis has coexpression of S-100 and CD1a and is characterized as having recognizable Birbeck granules by electron microscopy. Myeloid sarcoma tends to be CD1a and S-l00 negative allowing for a relative ease in distinction by immunophenotyping [1]. Although mastocytosis commonly presents with skin involvement, systemic mastocytosis must also be a consideration as the azurophilic granules of monocytic MS may be mistaken for mastocytic granules. Mastocytosis may also be immunoreactive for lysozyme, HLA-DR, CD33, and CD68, which may complicate the diagnosis if flow cytometry, alone, is used.
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
A
B
C
D
259
Fig. 4. Photomicrographs of nonclassical MS including epithelioid features with background fibrosis (A, H&E, original magnification ×600, case 8), round blue cell features (B, H&E, ×600, case 3), membranous immunoreactivity for CD99 (C, ×600, case 2), and monocytic features (D, H&E, ×600).
It is worth noting again that monocytic sarcoma may lack expression of some of the commonly described markers such as CD34 and CD117. However, they should still retain expression of CD68, CD43, CD33, and lysozyme, as in classic MS [2]. Per Lau et al [10], CD163, may be sensitive for monocytic sarcoma [2]. As mentioned above, CD56 or CD4 may be expressed and thus mimic a neuroendocrine tumor or T-cell lymphoma, respectively [2]. 4.1.9. Blastic plasmacytoid dendritic cell neoplasm Blastic plasmacytoid dendritic cell neoplasm, although rare, is a hematologic neoplasm that must be considered in the differential diagnosis of MS. It may be morphologically indistinguishable from MS as it frequently has medium to large atypical cells with fine chromatin, nucleoli, and irregular nuclear contours. Blastic plasmacytoid dendritic cell neoplasm most often affects the skin or soft tissue, similar to MS, so tumor location does not preclude the diagnosis of MS. Furthermore, 10% to 20% of blastic plasmacytoid dendritic cell neoplasms may develop myeloid neoplasms such as AML. In these cases, systematic investigation and close follow-up are warranted. Some immunostains for MS such as CD34, CD117, CD33, CD43, and CD163 have limited utility in differentiating these entities as there is significant overlap in expression. Myeloid sarcoma with monocytic differentiation is problematic as it is negative for CD34 and CD117 and positive for CD4, CD43, CD56, and CD68, which mimics blastic plasmacytoid dendritic cell neoplasm. Of note, the neoplastic cells of blastic plasmacytoid dendritic cell neoplasm are also partially positive for TdT and express CD2, CD7, and CD38. Two immunohistochemical panels have been suggested to aid differentiation [12]. The first larger panel includes plasmacytoid dendritic
cell markers (CD4, CD56, CD123, and TCL-1), lysozyme, MPO, and myxovirus A, a new research marker for an interferon α–inducible protein [12]. Myeloid sarcoma was associated with positive staining for lysozyme or MPO, as expected, and was negative staining for the other markers [12]. Blastic plasmacytoid dendritic cell neoplasm consistently had the opposite staining pattern [12]. This study also suggested a simplified 2-antibody panel with a scoring scheme involving CD56 and TCL-1. The panel was scored as an aggregate where a value of “1” was assigned to each positive result, “−1” to each negative result, and “0” to a marker that was not performed or equivocal. A score greater than or equal to zero was associated with blastic plasmacytoid dendritic cell neoplasm [12]. 5. Molecular and cytogenetic patterns in MS It has been estimated that 55% of MS may contain an cytogenetic abnormalities within the tumor cells and that de novo MS may lack cytogenetic abnormalities [7]. Interestingly, 1 study noted absence of cytogenetic abnormalities in all monoblastic MS cases, whereas only 1 (12.5%) of 8 nonmonoblastic cases had normal cytogenetics [7]. Although not well understood in MS, a study reported that NPM1 gene mutations occur in 50% to 60% of AMLs without cytogenetic abnormalities [13]. When they do occur, the most commonly reported cytogenetic abnormalities are trisomy 8 and inv(16) (Fig. 2I and J) [2,7,9], which occurred in our prototypical case. Trisomy 8 has been associated with a lower median survival by 1 author [14]. Where available, all of our review cases had normal cytogenetics, which matched with what Alexiev et al [7] has described regarding de novo MS. However, our example referral case was found to have inv(16) and trisomy 8 and did not have AML at presentation. Further study is required to elucidate the molecular anomalies observed in MS and their clinical impact.
260
R.P. Seifert et al. / Annals of Diagnostic Pathology 18 (2014) 253–260
6. Conclusions One must maintain a high index of suspicion for MS if a soft tissue tumor resembling non-Hodgkin lymphoma, poorly differentiated carcinoma, melanoma, or round blue cell tumor is encountered, as MS may precede or coincide with AML and may have nonclassic morphological features. When evaluating any tumor of unknown etiology, it is wise to save portions of fresh tumor for possible flow cytometric analysis. Based on our experience and a review of the literature, useful immunohistochemical markers for MS include CD43, lysozyme, CD68, CD117, and MPO. Per our review of the literature, CD163 may also be useful if the case has histomorphologic features suggestive of monocytic sarcoma [2]. A supplemental, minimal flow cytometry panel that includes HLA-DR, CD33, MPO, CD117, CD13, CD14, and CD45 should be performed if fresh tissue is available [1,2]. In our experience, special care must be taken when interpreting certain nonspecific immunomarkers because, if interpreted in a vacuum, they may lead the diagnostician astray. In these situations, thorough evaluation of the patient's clinical history and appropriate communication with the clinician is recommended. Utilization of additional ancillary testing, such as cytogenetic analysis and bone marrow biopsy, may also improve the diagnostic yield. Maintaining MS in the differential is important for the community pathologist as this entity is a mimicker, may have an aberrant immunophenotype, and may present de novo. This was true for our prototypical case example. The outside, “front-line,” pathologist recognized this entity and sent it to our institution for confirmation, using our subspecialty expertise. Achieving an early and accurate diagnosis is critical for timely and appropriate therapy and thus better patient outcomes. Acknowledgment The authors thank Ms Andrea Bumpus' assistance in submitting this abstract. The authors are also grateful to the Histology and Flow
Cytometry Laboratories at Moffitt Cancer Center for their excellent technical support. References [1] Pileri SA, Ascani S, Cox MC, Campidelli C, Bacci F, Piccioli M, et al. Myeloid sarcoma: clinico-pathologic, phenotypic and cytogenetic analysis of 92 adult patients. Leukemia 2007;21:340–50. [2] Klco JM, Welch JS, Nguyen TT, Hurley MY, Kreisel FH, Hassan A, et al. State of the art in myeloid sarcoma. Int J Lab Hematol 2011;33:555–65. [3] Neiman RS, Barcos M, Berard C, Bonner H, Mann R, Rydell RE, et al. Granulocytic sarcoma: a clinicopathologic study of 61 biopsied cases. Cancer 1981;48:1426–37. [4] Shah RS, Shin RK, Castellani RJ. Granulocytic sarcoma mimicking HSV encephalitis. Neurologist 2010;16:319–21. [5] Schwyzer R, Sherman GG, Cohn RJ, Poole JE, Willem P. Granulocytic sarcoma in children with acute myeloblastic leukemia and t(8;21). Med Pediatr Oncol 1998;31:144–9. [6] Macedo A, San Miguel JF, Vidriales MB, López-Berges MC, García-Marcos MA, Gonzalez M, et al. Phenotypic changes in acute myeloid leukaemia: implications in the detection of minimal residual disease. J Clin Pathol 1996;49:15–8. [7] Alexiev BA, Wang W, Ning Y, Chumsri S, Gojo I, Rodgers WH, et al. Myeloid sarcomas: a histologic, immunohistochemical, and cytogenetic study. Diagn Pathol 2007;2:42–9. [8] Campidelli C, Agostinelli C, Stitson R, Pileri SA. Myeloid sarcoma: extramedullary manifestation of myeloid disorders. Am J Clin Pathol 2009;132:426–37. [9] Deeb G, Baer MR, Gaile DP, Sait SN, Barcos M, Wetzler M, et al. Genomic profiling of myeloid sarcoma by array comparative genomic hybridization. Genes Chromosom Cancer 2005;44:373–83. [10] Lau SK, Chu PG, Weiss LM. CD163: a specific marker of macrophages in paraffinembedded tissue samples. Am J Clin Pathol 2004;122:794–801. [11] Hoyer JD, Grogg KL, Hanson CA, Gamez JD, Dogan A. CD33 detection by immunohistochemistry in paraffin-embedded tissues a new antibody shows excellent specificity and sensitivity for cells of myelomonocytic lineage. Am J Clin Pathol 2008;129:316–23. [12] Sangle NA, Schmidt RL, Patel JL, Jeffrey Medeiros L, Agarwal AM, Perkins SL, et al. Optimized immunohistochemical panel to differentiate myeloid sarcoma from blastic plasmacytoid dendritic cell neoplasm. Mod Pathol 2014:1–7. [13] Chen W, Rassidakis GZ, Medeiros LJ. Nucleophosmin gene mutations in acute myeloid leukemia. Arch Pathol Lab Med 2006;130:1687–92. [14] Tsimberidou AM, Kantarjian HM, Estey E, Cortes JE, Verstovsek S, Faderl S, et al. Outcome in patients with nonleukemic granulocytic sarcoma treated with chemotherapy with or without radiotherapy. Leukemia 2003;17:1100–3.
