Immunophenotypic Profile of Myeloid Cells in Granulocytic Sarcoma ...

Hematopathology / IMMUNOPHENOTYPIC PROFILE IN GRANULOCYTIC SARCOMA

Immunophenotypic Profile of Myeloid Cells in Granulocytic Sarcoma by Immunohistochemistry Correlation With Blast Differentiation in Bone Marrow Chung-Che Chang, MD, PhD,1 Camellia Eshoa, MD,1 Bal Kampalath, MD,1 Vinod B. Shidham, MD, FIAC, MRCPath,1 and Sherrie Perkins, MD, PhD2 Key Words: Immunophenotypic profile; Granulocytic sarcoma; Lineage differentiation

Abstract The present study was designed to evaluate the lineage differentiation (particularly monocytic differentiation) of immature myeloid cells in granulocytic sarcoma (GS) by immunohistochemistry and correlate the results with lineage differentiation of blasts in the bone marrow and to determine the degree of maturation of the infiltrating myeloid cells in GS by immunohistochemistry using CD34 and HLA-DR. Immunohistochemical stains were performed on paraffin-embedded tissue from 17 GS lesions with lineage-associated markers: myeloperoxidase, CD68 (KP1), CD68 (PG-Ml), glycophorin A , factor VIII, and CD56; and with markers for blasts and immature myeloid cells: CD34 and HLA-DR. Our results show that positive staining with PG-M1, but not KP1, suggests monocytic differentiation of myeloid cells in GS and correlates with the monocytic differentiation of blasts in the bone marrow. Expression of CD56 is frequent in GS, especially when the marrow blasts have monocytic differentiation, and should not be interpreted as a primary natural-killer cell process. The immature myeloid cells in GS are frequently HLA-DR positive. However, CD34 positivity of the immature myeloid cells is relatively uncommon, except in cases with underlying myelodysplastic syndrome or chronic myelogenous leukemia.

© American Society of Clinical Pathologists

Granulocytic sarcoma (GS) is an uncommon extramedullary tumor composed of immature myeloid precursor cells. It usually occurs in association with acute myeloid leukemia, myeloproliferative disorders, or myelodysplastic syndrome, but it can occur without overt hematologic disease.1,2 A majority of patients without known hematologic disease at the time of GS diagnosis develop acute leukemia in 1 to 49 months (mean, 10 months).1,2 Most GSs occurring in patients with a known myeloproliferative disorder are associated with blast crisis. The most common sites of involvement by GS are bone, periosteum, soft tissue, lymph node, and skin.1,2 The histomorphologic diagnosis of GS sometimes can be challenging to pathologists, particularly in the absence of a known hematologic disorder. The morphologic features of the tumors vary from well differentiated, which includes all stages of myeloid differentiation, to poorly differentiated or blastic with little or no evidence of myeloid differentiation.3-6 Poorly differentiated GS, in particular, may resemble large cell non-Hodgkin lymphoma.3 Immunophenotypic features of GS remain poorly recognized and frequently are confused with other malignant neoplasms. Some studies have demonstrated that in cases of GS, the immunophenotypic patterns of myeloid cells in the skin infiltrates were concordant with those in the bone marrow.4,5 However, others have suggested that, in many cases, the infiltrating cells of GS have a different phenotype from those in the bone marrow.6 Some antibodies (eg, anti–S100 protein and MB2, a B-cell marker) produced staining in a few cases that could have misled to a diagnosis of histiocytic neoplasm, melanoma, or malignant lymphoma.6,7 The recent development of antibodies against fixativeresistant epitopes and of new antigen retrieval techniques has Am J Clin Pathol 2000114:807-811 807

Chang et al / IMMUNOPHENOTYPIC PROFILE IN GRANULOCYTIC SARCOMA

expanded the possibility of further defining the immunophenotypic profiles of the infiltrating immature myeloid cells in GS. A few recent studies have documented that accurately classifying acute myeloid leukemia according to the FrenchAmerican-British classification can be achieved by immunohistochemistry using marrow core biopsy.8-10 The present study was designed to evaluate the lineage differentiation (particularly monocytic differentiation) of the immature myeloid cells in GS by immunohistochemistry using a comprehensive panel of lineage-associated antibodies, to correlate the results with lineage differentiation of blasts in bone marrow, and to determine the degree of maturation of the infiltrating myeloid cells in GS by immunohistochemistry using CD34 and HLA-DR.

Materials and Methods Patients We retrospectively studied 17 cases of GS (males, 13; females, 4; age range, 2 months to 70 years; median age, 42 years) with preceding, concurrent, or subsequent diagnosis of acute myeloid leukemia (AML), chronic myeloid leukemia (CML), or myelodysplastic syndrome (MDS). All materials, including histology sections, immunohistochemical and flow cytometry studies, and histochemical stains, from the GS were reviewed, and the diagnoses were confirmed. The French-American-British classification subtypes of the associated bone marrow diagnoses in these cases, as determined by enzyme cytochemistry with or without flow cytometry, were as follows: AML without monocytic differentiation, 4 (M1, 1; M2, 2; M3, 1); AML with monocytic differentiation, 7 (M4, 3; M5a, 3; M5b, 1); CML, 3 (chronic phase, 2; blast crisis with marrow blasts showing monocytic differentiation, 1); and MDS, 3 (refractory anemia with excess of blasts, 2; chronic myelomonocytic leukemia, 1). All 3 cases with CML had positive Philadelphia chromosomes by cytogenetic analysis, and the single case of M3 had t(15;17)(q22;q12-21). The sites of GS presentation included the following: lymph node, 6; skin, 3; paraspinal area, 3; brain, 1; lung, 1; ovary, 1; orbit, 1; and scrotum, 1. The clinical information and marrow diagnoses are summarized in ❚Table 1❚. Immunohistochemical Stains We performed immunohistochemical staining on paraffin-embedded tissue sections of GS lesions with antibodies against lineage-associated (but not necessarily specific) markers: myeloperoxidase (MPO; DAKO, Carpinteria, CA) for myeloid lineage, CD68 (clone, KPI, DAKO) and CD68 (clone, PG-M1, DAKO) for monocytic lineage, 808

Am J Clin Pathol 2000;114:807-811

glycophorin A (clone, JC159, DAKO) for erythroid lineage, factor VIII (DAKO) for megakaryocytic lineage, and CD56 (clone, 1136, Novocastra, Newcastle upon Tyne, England) for natural-killer cell lineage. Markers for blasts and immature myeloid cells (CD34, clone, QBEND/10, DAKO; and HLA-DR, clone, LN3, Biogenex, San Ramon, CA) were used to evaluate the degree of maturation of the infiltrating myeloid cells. In some cases, not all markers were used owing to insufficient materials. Paraffin section immunohistochemistry was performed using the avidin-biotin-peroxidase technique with or without antigen epitope enhancement. The diaminobenzidine reaction was used as the final detection step. ❚Table 2❚ shows the characteristics of antibodies used and the methods of antigen epitope enhancement for each antibody. Following antigen epitope enhancement, the staining was performed using a DAKO automated immunostainer. Interpretation of Staining Results The expression of each marker was reviewed systematically by two of us (C.-C.C. and C.E.), independently without the knowledge of clinical presentation or marrow diagnosis. The intensity of staining was graded as follows: strongly positive if positive staining could be recognized readily at low power (2× objective), moderately positive if positive staining could be recognized at intermediate power (10× objective) but not at low power, weakly positive if recognized only at high power (40× objective), and negative if staining could not be recognized even at high power. This grading system led to agreement between the 2 observers for more than 95% of staining results. For discrepant results, the 2 investigators reevaluated the slide and eventually reached consensus.

Results The results of the immunohistochemical staining for each marker are shown in Table 1. For lineage differentiation, MPO and KP1 were sensitive markers for determining myeloid origin in GS and were positive in 88% (14/16) and 94% (16/17) of cases, respectively. KP1 seems to be a better marker than MPO for myeloid origin since KP1 was moderately or strongly positive in 16 of 17 cases, while MPO was moderately or strongly positive in 11 of 16 cases (P = .059, chi-square). In addition, MPO was strongly positive in 2 of 4 GS lesions from patients with marrow diagnosis of M5. In GS lesions, PG-M1 was more frequently positive (6/7 [86%]) in cases with marrow blasts demonstrating monocytic differentiation (including M4, M5a, M5b, chronic myelomonocytic leukemia, and CML in blast crisis with monocytic differentiation) than in the cases (2/8 [25%]) with marrow blasts that showed no evidence of monocytic © American Society of Clinical Pathologists

Hematopathology / ORIGINAL ARTICLE

❚Table 1❚ Clinical Information and Results of Immunohistochemical Stains in Granulocytic Sarcoma Lesions* Case No./ Sex/ Age (y)†

Bone Marrow Diagnosis

Site

MPO

KP1

PG-M1

Factor VIII

Glycophorin A

CD56

CD34

HLA-DR

1/M/47 2/M/43 3/F/36 4/M/69 5/M/36 6/M/69 7/M/59 8/M/8 9/M/55 10/F/26 11/M/53 12/M/2 mo 13/F/70 14/M/67 15/M/62 16/M/17 mo 17/F/21

CML, BC CML, CP CML, CP CMMoL RAEB RAEB M1 M2 M2 M3 M4 M4 M4 M5a M5a M5a M5b

LN LN SPI Skin SPI Skin LN Other Other Other LN Other LN Other LN Skin SPI

+++ +++ +++ + +++ ++ ++ + +++ ND +++ + ++ +++ +++ – –

+++ ++ +++ +++ +++ +++ +++ +++ +++ ++ +++ – +++ +++ +++ +++ +++

++ – – ND – – – ++ ++ – + ++ +++ +++ ++ ND –

– – – ND ++ – – – – – – – – – – ND –

– – – ND – – – – – ND – – – – – ND ND

– ND – ++ ND – – – – ND ++ – – + +++ ND ND

++ +++ ++ – +++ ++ – – – – +++ – – – – ND –

++ ND +++ ND +++ ++ +++ +++ + – +++ – +++ +++ +++ ND +++

BC, blast crisis; CML, chronic myeloid leukemia; CMMoL, chronic myelomonocytic leukemia; CP, chronic phase; KP1, CD68; LN, lymph node; MPO, myeloperoxidase; ND, not done; PG-M1, CD68; RAEB, refractory anemia with excess of blasts; SPI, paraspinal mass. * Staining intensity: –, negative; +, weak; ++, moderate; +++, strong. For cases 7-17, diagnoses are the French-American-British subtypes. † Unless otherwise indicated.

❚Table 2❚ Characteristics of Antibodies Used Antibody

Clone

Manufacturer

Dilution

MPO CD68 CD68 Factor VIII

NA* KP1 PG-M1 NA*

DAKO, Carpinteria, CA DAKO DAKO DAKO

1:10,000 1:500 1:1,000 1:12,000

Glycophorin A CD56 CD34 HLA-DR

JC159 1B6 QBEND/10 LN3

DAKO Novocastra, Newcastle upon Tyne, England DAKO Biogenex, San Ramon, CA

1:200 1:100 1:50 1:2,000

* †

Method for Epitope Enhancement ND Heat† Heat Heat Proteinase K ND Heat Heat Heat

Not applicable, polyclonal. With DAKO Target Retrieval Solution at pH 6.0 for 30 minutes at 97°C.

differentiation (P < .02, chi-square). In contrast, KP1 expression was seen in almost all GS lesions regardless of marrow blasts with or without monocytic differentiation. For example, case 10, without marrow blasts demonstrating monocytic differentiation, showed moderate staining for KPl but negative staining for PG-M1 ❚Image 1❚. Factor VIII was positive in only 1 GS lesion, in a case with marrow that showed refractory anemia with excess of blasts. This case had very dysplastic megakaryocytes and megakaryocytic blasts in marrow and stained positively with factor VIII. Glycophorin A was negative in all GSs, probably because no blasts in any case showed erythroid differentiation in bone marrow samples. CD56 was positive in GS lesions only in cases with marrow blasts that showed monocytic differentiation. Among the GS in cases with marrow showing AML with monocytic differentiation, CD56 was positive in 3 (60%) of © American Society of Clinical Pathologists

5 lesions stained, 2 at the nodal site and 1 in lung. CD56 was also positive in 1 GS of skin with marrow that showed chronic myelomonocytic leukemia. We further evaluated the degree of maturation of the myeloid cells in GS by the expression of CD34 and HLADR. In general, CD34 expression was seen relatively infrequently in GS (6/16 [38%]). However, CD34 expression was observed more frequently in GS cases associated with CML or MDS than in those associated with AML (5/6 [83%] vs 1/10 [10%]; P = .003, chi-square). In contrast, HLA-DR was expressed in all but 2 of 14 cases tested (1 M3 and 1 M4).

Discussion The morphologic diagnosis of granulocytic sarcoma on paraffin-embedded tissue sections is difficult and sometimes impossible by using routine H&E sections because of the Am J Clin Pathol 2000;114:807-811 809

Chang et al / IMMUNOPHENOTYPIC PROFILE IN GRANULOCYTIC SARCOMA

A

B

❚Image 1❚ Immature myeloid cells in granulocytic sarcoma with strong KP1 and negative PG-M1 staining (the corresponding marrow blasts did not show monocytic differentiation). A, Infiltrate of immature myeloid cells (H&E, ×100). B, The majority of immature myeloid cells show moderate staining intensity. Histiocytes show very strong staining of dendritic processes (KP1, ×100). C, Negative staining in immature myeloid cells and strongly positive staining in dendritic processes of histiocytes (PG-M1, ×150).

C

primitive morphologic features of the neoplastic cells. Studies have been done to improve the accuracy of diagnosis of GS by using immunohistochemistry.3-7,11-13 These studies have focused on using immunohistochemistry to diagnose GS as an entity. The present study extends these findings and attempts to determine whether the lineage differentiation, particularly monocytic differentiation, and degree of maturation of immature myeloid cells in GS can be determined by using immunohistochemistry. Our attempt to examine the lineage differentiation of immature myeloid cells in GS and to correlate it with blast differentiation in bone marrow demonstrates promising results. Our findings suggest that PG-M1 was more specific than KP1 for correlating with monocytic differentiation of blasts in bone marrow. Importantly, our study demonstrates that KP1 was positive in most of the GS and that KP1 expression in GS cannot be used as a marker that correlates 810

Am J Clin Pathol 2000;114:807-811

with monocytic differentiation of blasts in bone marrow. Our findings agree with the results of previous studies by Pileri et al8 and Manaloor et al.14 They reported that in bone marrow core biopsy, PG-M1 was more reliable than KP1 for monocytic differentiation by immunohistochemistry. KP1, however, is a good marker and even appears more sensitive than MPO for documenting myeloid origin. In addition, although MPO is negative in the marrow blasts in M5 using cytochemical stain, 2 of the M5 cases showed strong MPO expression in the GS lesions by immunohistochemistry. Therefore, expression of MPO in immature myeloid cells of GS by immunohistochemistry cannot be used to exclude the possibility of marrow blasts being classified as M5. For megakaryocytic and erythroid differentiation, the immunophenotypic profile of immature myeloid cells in GS also seems to correlate with the findings for marrow blasts. The only GS case (case 5) that had MDS with marrow blasts expressing factor VIII showed that factor VIII was also positive in the myeloid cells in GS, suggesting megakaryocytic differentiation. Finally, all GS lesions studied were negative for glycophorin A staining. This correlated with the finding that no GS cases included in the study had marrow blasts demonstrating erythroid differentiation. Furthermore, our results show that expression of CD56 in GS strongly correlates with monocytic differentiation in © American Society of Clinical Pathologists

Hematopathology / ORIGINAL ARTICLE

the bone marrow. Although expression of CD56 has been shown relatively commonly in GS in previous studies,2,15,16 these studies did not report the correlation between CD56 expression and monocytic differentiation in bone marrow as shown in the present study. The expression of the adhesion molecules, such as CD56, may increase the ability of circulating blasts to home to the extramedullary sites. Our results indicate that CD34 is expressed relatively infrequently in myeloid cells of GS and are in agreement with previous studies.3,12 In addition, we observed that the myeloid cells in GS cases with CML or MDS were more likely to express CD34 than were those with AML. This most likely reflects the concept that the blasts in patients with CML or MDS are more primitive than those in patients with AML.17,18 It is important to recognize that the absence of CD34 expression in GS does not exclude the diagnosis of GS. The immature myeloid cells in GS are frequently HLADR positive, in contrast with CD34. The myeloid cells in all but 2 GS lesions expressed HLA-DR. One of 2 cases that did not express HLA-DR had a marrow diagnosis of acute promyelocytic leukemia (M3). However, since there is only 1 M3 case included in the present study, further study is required to evaluate whether absence of HLA-DR expression in GS favors M3 as the marrow diagnosis. Our results reveal that positive staining with PG-M1, but not KP1, suggests monocytic differentiation of myeloid cells in GS and correlates with the monocytic differentiation of the blasts in bone marrow. Expression of CD56 is relatively frequent in GS with monocytic differentiation and should not be interpreted as a primary natural killer cell process. The immature myeloid cells in GS are frequently HLA-DR positive. However, CD34 positivity of the immature myeloid cells is relatively uncommon, except in cases with underlying MDS or chronic myeloid leukemia. From the Departments of Pathology, 1Medical College of Wisconsin, Milwaukee, and 2University of Utah, Salt Lake City. Address reprint requests to Dr Chang: Dept of Pathology, Medical College of Wisconsin, 9200 W Wisconsin Ave, Milwaukee, WI 53006.

References 1. Neiman RS, Barcos M, Berard C, et al. Granulocytic sarcoma: a clinicopathologic study of 61 biopsied cases. Cancer. 1981;48:1426-1437. 2. Byrd JC, Edenfield WJ, Dow NS, et al. Extramedullary myeloid cell tumors in myelodysplastic-syndromes: not a true indication of impending acute myeloid leukemia. Leuk Lymphoma. 1996;21:153-159. 3. Traweek ST, Arber DA, Rappaport H, et al. Extramedullary myeloid cell tumors: an immunohistochemical and morphologic study of 28 cases. Am J Surg Pathol. 1993;17:1011-1019.

© American Society of Clinical Pathologists

4. Ratnam KV, Su WP, Ziesmer SC, et al. Value of immunohistochemistry in the diagnosis of leukemia cutis: study of 54 cases using paraffin-section markers. J Cutan Pathol. 1992;19:193-200. 5. Hudock J, Chatten J, Miettinen M. Immunohistochemical evaluation of myeloid leukemia infiltrates (granulocytic sarcomas) in formaldehyde-fixed, paraffin-embedded tissue. Am J Clin Pathol. 1994;102:55-60. 6. Kaiserling E, Horny HP, Geerts ML, et al. Skin involvement in myelogenous leukemia: morphologic and immunophenotypic heterogeneity of skin infiltrates. Mod Pathol. 1994;7:771-779. 7. Menasce LP, Banerjee SS, Beckett E, et al. Extra-medullary myeloid tumour (granulocytic sarcoma) is often misdiagnosed: a study of 26 cases. Histopathology. 1999;34:391-398. 8. Pileri SA, Ascani S, Milani M, et al. Acute leukaemia immunophenotyping in bone-marrow routine sections. Br J Haematol. 1999;105:394-401. 9. Arber DA, Jenkins KA. Paraffin section immunophenotyping of acute leukemias in bone marrow specimens. Am J Clin Pathol. 1996;106:462-468. 10. Horny HP, Wehrmann M, Steinke B, et al. Assessment of the value of immunohistochemistry in the subtyping of acute leukemia on routinely processed bone marrow biopsy specimens with particular reference to macrophage-associated antibodies. Hum Pathol. 1994;25:810-814. 11. Fellbaum C, Hansmann ML. Immunohistochemical differential diagnosis of granulocytic sarcomas and malignant lymphomas on formalin-fixed material. Virchows Arch A Pathol Anat Histopathol. 1990;416:351-355. 12. Ritter JH, Goldstein NS, Argenyi Z, et al. Granulocytic sarcoma: an immunohistologic comparison with peripheral T-cell lymphoma in paraffin sections. J Cutan Pathol. 1994;21:207-201. 13. Quintanilla-Martinez L, Zukerberg LR, Ferry JA, et al. Extramedullary tumors of lymphoid or myeloid blasts: the role of immunohistology in diagnosis and classification. Am J Clin Pathol. 1995;104:431-443. 14. Manaloor EJ, Neiman RS, Heilman DK, et al. Immunohistochemistry can be used to subtype acute myeloid leukemia in routinely processed bone marrow biopsy specimens: comparison with flow cytometry. Am J Clin Pathol. 2000;113:814-822. 15. Kuwabara H, Nagai M, Yamaoka G, et al. Specific skin manifestations in CD56 positive acute myeloid leukemia. J Cutan Pathol. 1999;26:1-5. 16. Kaddu S, Beham-Schmid C, Zenahlik P, et al. CD56+ blastic transformation of chronic myeloid leukemia involving the skin. J Cutan Pathol. 1999;26:497-503. 17. Kanter-Lewensohn L, Hellstrom-Lindberg E, Kock Y, et al. Analysis of CD34-positive cells in bone marrow from patients with myelodysplastic syndromes and acute myeloid leukemia and in normal individuals: a comparison between FACS analysis and immunohistochemistry. Eur J Haematol. 1996;56:124-129. 18. Khalidi HS, Brynes RK, Medeiros LJ, et al. The immunophenotype of blast transformation of chronic myelogenous leukemia: a high frequency of mixed lineage phenotype in “lymphoid” blasts and a comparison of morphologic, immunophenotypic, and molecular findings. Mod Pathol. 1998;11:1211-1221.

Am J Clin Pathol 2000;114:807-811 811