Context.-Immunophenotyping of bone marrow (BM) specimens with acute myelogenous leukemia (AML) may be performed by flow cytometric (FC) or immunohistochemical (IN) techniques. Some markers (CD34, CD15, and CD117) are available for both techniques. Myeloperoxidase (MPO) analysis may be performed by enzyme cytochemical (EC) or IH techniques.
Objective.-To determine the reliability of these markers and MPO by these techniques, we designed a study to compare the results of analyses of these markers and MPO by FC (CD34, CD15, and CD117), EC (MPO), and IH (CD34, CD15, CD117, and MPO) techniques.
Materials and Methods.-Twenty-nine AMLs formed the basis of the study. These AMIs all had been immunophenotyped previously by FC analysis; 27 also had had EC analysis performed. Of the AMLs, 29 had BM core biopsies and 26 had BM clots that could be evaluated. The paraffin blocks of the 29 BM core biopsies and 26 BM clots were
stained for CD34, CD1 17, MPO, and CDI 5. These results were compared with results by FC analysis (CD34, CD15, and CD117) and EC analysis (MPO).
Results.-Immunodetection of CD34 expression in AML had a similar sensitivity by FC and IH techniques. Immunodetection of CD1 5 and CD1 17 had a higher sensitivity by FC analysis than by IH analysis. Detection of MPO by IH analysis was more sensitive than by EC analysis. There was no correlation of French-American-British (FAB) subtype of AML with CD34 or CD1 17 expression. Expression of CD1 5 was associated with AMLs with a monocytic component. Myeloperoxidase reactivity by IH analysis was observed in AMLs originally FAB subtyped as MO.
Conclusions.-CD34 can be equally detected by FC and IH techniques. CD15 and CD117 are better detected by FC analysis and MPO is better detected by IH analysis.
(Arch Pathol Lab Med. 2001;125:1063-1069)
Immunophenotyping of bone marrow (BM) specimens is useful in establishing a diagnosis of acute myelocytic leukemia (AML) and important in predicting its prognosis. In particular, in a study by Geller et al, expression of CD34 in adult AML was significant in predicting the response to therapy.1,2 In CD34^sup +^ adult AML, the complete remission rate was 59%, whereas in CD34^sup -^ cases, the complete remission rate was 87%. Patients with CD34^sup +^ AML were one third as likely to enter complete remission as were those with CD34^sup -^ AML (P = .066).
Immunophenotyping of BM specimens may be performed by flow cytometric (FC) analysis (BM aspirate or core biopsy) or by immunohistochemical (IH) staining (frozen or paraffin-embedded clot and/or core biopsy).3-5 Although the antigens analyzed by many monoclonal antibodies used in diagnostic hematopathology are destroyed by routine tissue fixation, decalcification, and processing, a BM aspirate specimen for FC analysis may not be available due to reticulin fibrosis or a "packed" BM, and there are reports of the usefulness of IH staining in immunophenotyping acute leukemias.6-10 In addition, the repertoire of IH stains that can be performed on paraffin-embedded tissue for diagnostic hematopathology is expanding.
Since certain antibodies are available for immunophenotyping acute leukemia and establishing a diagnosis of AML by FC analysis and by IH staining, we wanted to correlate the results of expression of these markers (CD34, CD117, and CD15) by FC analysis and by paraffin IH (PIH) staining of the BM clot and core biopsy sections. In addition, we compared myeloperoxidase (MPO) reactivity by enzyme cytochemical (EC) and IH staining methods. These results should help establish the reliability of these markers and the MPO stain as determined by these immunophenotypic methods.
MATERIALS AND METHODS
Twenty-nine cases of AML were retrieved from the files of the Division of Hematopathology at St Louis University Health Sciences Center (St Louis, Mo). All cases (except 1 without spicules on BM aspiration or touch preparations of the BM core biopsy) had been subtyped according to the French-American-British (FAB) classification.11-14 Cases included 3 AMLs classified as MO, 2 Mls, 6 M2s (1 with eosinophilia), 2 M3s (1 M3, microgranular variant), 9 M4s (2 with eosinophilia), 2 MSs, 1 M6, and 3 M7s. Twenty-six cases had BM clot and core biopsy material that was adequate and able to be evaluated. The only difference in processing of the BM core and clot was decalcification of the trephine core biopsy. Three cases had no BM clot material to be evaluated but did have BM core biopsy material that was able to be evaluated. Two of these 3 cases did have BM aspirate smears and/ or touch preparations of the BM core biopsy with spicules that could be evaluated. Enzyme cytochemical stains were performed on BM aspirate smears in 27 of the 29 cases. The 2 cases in which EC stains were not performed were represented by AML subtypes M6 and M7 without spicules on the BM aspiration or on the touch preparations of the BM core biopsy. The EC stains were interpreted on fresh material immediately after the stains were performed. All cases had FC immunophenotyping performed on a BM aspirate or a suspension prepared from the BM core biopsy as previously described.15
The BM aspirate smears and touch preparations of the BM core biopsy had been routinely stained with Wright stain and for the presence of MPO, Sudan black B, alpha-naphthyl acetate esterase, alphanaphthyl butyrate esterase, and chloroacetate esterase (Sigma Chemical Co, St Louis, Mo) according to kit procedures. Only 2 cases (case 4, AML M7, and case 14, AML M6) did not have EC stains performed on BM aspirate smears. The BM clots and core biopsies were fixed in B-5 fixative, decalcified in rapid decalcification solution (Apex Engineering Products, Plainfield, Ill) for 1 hour, and placed in a vacuum infiltration processor (Tissue Tek, Elkhart, Ind) for overnight processing. Hematoxylin-eosinstained serial sections at 3 levels were prepared.
Flow Cytometric Immunophenotyping
The BM aspirates or cell suspensions of the BM core biopsy were analyzed on a flow cytometer (Cytoronabsolute, Ortho Diagnostic Systems, Raritan, NJ) for various antigens using standard techniques and the following commercially available monoclonal antibodies: CD1, CD15, HLA-DR (Ortho Diagnostic); CD2, CD10, CD13, CDA, CD19, CD20, CD24, CD33, CD34, CD56, CD64, CD117, and terminal deoxynucleotidyl transferase (TdT) (Coulter Immunology, Hialeah, Fla); CD3, CD4, CD7, and CD45 (Becton-Dickinson, San Jose, Calif); CD5 and CD8 (Gen Trak, Inc, Wayne, Pa); CD41 and CD42b (Dako Corporation, Carpinteria, Calif); K and X light chains (Kallestad Inc, Chaska, Mass); and intracytoplasmic CD3/MPO (cCD3/MPO) (Caltag, Burlingame, Calif). Dual staining of antibodies was performed as follows: CD3/CD4, CD8/CD56, CD19/CDS, CD20/HLA-DR, CD45/ CD10, CD13/CD14, CD33/CD34, CD15/CD117, CD2/CD24, CD41/CD42b, kappa/lambda, CD45/TdT, and CCD3/MPO. The others were singly labeled. Positivity of particular cells for a given marker was defined as greater than 20% of the cells having expression of the marker. Lymphocyte and monocyte regions were gated upon, defined by their forward and side light-scatter properties. At least 2000 events were counted for each antibody. The percentage of blasts was determined by multiplying the percentage of blasts in the lymphocyte region times the percentage of all cells in the lymphocyte region and adding the percentage of blasts in the monocyte region times the percentage of all cells in the monocyte region. This study was retrospective and MPO had not been analyzed by FC in the great majority (27/29) of cases.
The following immunoperoxidase stains were performed on the paraffin-embedded tissue blocks of the BM clot and core biopsy using the Dako LSAB-2 peroxidase system: CD34 (Q-Bend 10), CD117 (c-Kit; polyclonal), MPO (polyclonal) (Dako), and CD15 (Leu-Ml; clone MMA) (Becton-Dickinson). CD34 and MPO reactions were performed using antigen retrieval with Citra Buffer Retrieval Solution (Dako). Control tissue was tonsil for CD34 and MPO, gastric stromal tumor for CD117, and Hodgkin lymphoma for CD15. In 3 cases (cases 16, 23, and 29), PIH was not performed on BM clot material either because there were no BM elements on hematoxylin-eosin-stained sections of the BM clot or no BM aspirate was obtained for clot material.
The discordant results of the PIH analysis of CD34, CD15, CD117, and MPO on the BM clot and core biopsy sections and the EC analysis of MPO and FC analysis of CD34, CD15, and CD117 are demonstrated in the Table. Only 2 cases (cases 6 and 27, both AML M4) revealed completely concordant results of all antigens and the enzyme tested by all methods.
CD34 by FC and PIH Methods
By FC analysis, 19 (66%) of 29 AMLs revealed CD34 expression by the entire population or a subset of the blasts. Paraffin IH analysis of the BM core biopsy in the 29 cases revealed 19 cases (66%) with immunoreactivity. The 26 cases with BM clot material that could be evaluated revealed 16 cases (62%) with immunoreactivity. There was complete concordance in the PIH results of those cases with both BM clot and core material that could be evaluated. There were 2 cases with discordant results by FC analysis versus PIH. One case was represented by an AML M5a that was CD34^sup -^ by PIH on BM clot and core biopsy, but CD34^sup +^ by FC analysis. The CD34 expression by FC analysis was of a low intensity (Figure 1). The other case was represented by a microgranular variant of AML M3, which revealed less than 5% CD34^sup +^ blasts by PIH analysis of BM clot and core and was CD34^sup -^ by FC analysis. There was no correlation of CD34 expression with a specific FAB subtype of AML. In general, CD34 negativity occurred in the majority of FAB subtypes M3, M4, and M5.
CD15 Analysis by FC and PIH Methods
By FC analysis, 13 (45%) of 29 AMLs revealed CD15 expression by the entire population or a subset of the blasts. Paraffin IH analysis of the BM core biopsy in the 29 cases revealed 7 cases (24%) with immunoreactivity. The 26 cases with BM clot material that could be evaluated revealed 8 cases (31%) with immunoreactivity. There was 96% concordance in the PIH results of those cases with BM clot and core material available. The one discordant case revealed less than 3% of blasts with moderate immunoreactivity on the BM clot section and a negative result on the BM core biopsy. There were 6 cases with discordant results by FC versus PIH analysis. In 5 of these 6 cases, the blasts were completely negative for CD15 by PIH analysis (4 cases, BM clot and core; 1 case, BM core only could be evaluated) and were CD15^sup +^ by FC analysis (4 cases, subset of blasts positive; 1 case, entire population of blasts positive). Review of the flow cytograms in these 5 cases revealed CD15 expression of a low intensity in 1 case, of a low to intermediate intensity in 3 cases, and of an intermediate to strong intensity in 1 case. The sixth discordant case revealed less than 3% of blasts with moderate immunoreactivity on the BM clot, CD15- blasts by PIH analysis of the BM core, and CD15 expression by the entire population of blasts by FC analysis. Review of the flow cytogram in this case revealed CD15 expression of low to intermediate intensity (Figure 2). CD15 positivity was observed in 45% of AMLs with a monocytic component. Only 1 AML without a monocytic component revealed a subset of blasts that were CD15^sup +^ by FC analysis.
CD117 Analysis by FC and PIH Methods
By FC analysis, 24 (83%) of 29 AMLs revealed CD117 expression by the entire population or a subset of the blasts., Paraffin IH analysis of the BM core biopsy in the 29 cases revealed 18 cases (62%) with immunoreactivity. The 26 cases with BM clot material that could be evaluated revealed only 7 cases (27%) with immunoreactivity. There was marked (62%) discordance in the PIH results of those cases with both BM clot and core material available. In 9 (35%) of these 26 cases, CD117 revealed immunoreactivity by PIH analysis of the BM core and was negative by PIH analysis of the BM clot. In an additional 27% (7/26 cases), the PIH reactivity for CD117 was much more intense on the BM core than on the BM clot (Figure 3). There were no cases that were CD117- by PIH analysis of the BM core with PIH reactivity on the BM clot. There were 6 cases with discordant results by FC versus PIH analysis; all of these cases were CD117^sup +^ by FC analysis and were CD117by PIH analysis of the BM clot and core. Review of the flow cytograms of these 6 cases revealed that the CD117 expression was of a low intensity in 2 cases (Figure 4), of a low to intermediate intensity in 3 cases, and of an intermediate intensity in 1 case. There were no cases of CD117 immunoreactivity by PIH analysis of the BM clot and/or core that were CD117^sup -^ by FC analysis. There was no correlation of CD117 expression with the FAB subtype of AML.
Myeloperoxidase Analysis by EC and PlH Methods
By EC analysis, 21 (78%) of 27 AMLs revealed reactivity for MPO within the entire population or a subset of the blasts. Paraffin IH analysis of the BM core biopsy in 29 cases revealed 26 cases (90%) with immunoreactivity. The 26 cases with BM clot material available for evaluation revealed 24 cases (92%) with immunoreactivity. There was complete concordance in the PIH results of those cases with both BM clot and core material available. There were 6 cases with discordant results by EC versus PIH methods. In 4 of these 6 cases, the blasts were MPO- by EC, and either all the blasts (2 cases) or a subset of the blasts (2 cases) were immunoreactive by PIH analysis of the BM clot and core biopsy (2 cases) or BM core only (2 cases). In the other 2 discordant cases (AML MO), less than 3% blasts were MPO^sup +^ by EC, and either all of the blasts (1 case; Figure 5) or a subset (>3%) of the blasts (1 case) were immunoreactive by PIH analysis of the BM clot and core biopsy. No cases were MPO+ by EC analysis and MPO- by PIH analysis. As explained earlier, only 2 cases (cases 3 and 11) had MPO analyzed by FC analysis. In case 3, MPO was positive by IH, EC, and FC analyses; in case 11, MPO was negative by all 3 methods.
As previously discussed, immunophenotyping of BM specimens may be performed by FC or IH methods. Flow cytometric immunophenotyping may be performed on BM aspirates or core biopsies as previously described." Immunohistochemical immunophenotyping may be performed on BM clot or core biopsy material. Since appropriate BM material may not be available for FC analysis in occasional cases of AML or extramedullary presentation as granulocytic sarcoma and some antibodies are available for both FC and IH analysis, we compared the results of these antibodies (CD34, CD15, and CD117) by both methods. In particular, CD117 is a relatively new antibody available for PIH analysis, as is MPO. Thus, we also wanted to compare results of MPO reactivity by EC and IH methods. In this discussion, our comparative results of CD34, CD15, CD117, and MPO by FC or EC and IH methods will be addressed individually. Our results will be compared with findings reported in the current literature.
In a study by Hanson et al in 1992,(16) analysis of CD34 expression in acute leukemia was compared by FC and PIH analysis of BM core and/or clot sections. This study revealed that all cases that were CD34^sup +^ by FC analysis were also CD34^sup +^ by PIH analysis. Manaloor et al17 also demonstrated a high concordance for positive and negative results for CD34 by both methods. However, in both studies, there was no comparison of PIH results in BM core versus clot sections. In contrast to these results, a study by Arber et al7 in 1996 concluded that the IH detection of CD34 appeared to be less reliable than the FC method. Their study included 14 acute leukemias that were CD34^sup +^ by FC analysis but negative by IH analysis. There were only 2 cases (both AMLs) that were CD34- by FC analysis and CD34^sup +^ by IH analysis. Our present study supports the study by Hanson et al. There were only 2 cases with discordant results by FC versus IH methods (1 case was CD34^sup +^ by FC analysis and CD34^sup -^ by PIH analysis, and 1 was CD34- by FC analysis and
CD15 has been described as preferentially detecting mature myeloid cells.9 Contrasting results have been reported in the literature in regard to CD15 immunoreactivity of myeloblasts by PIH analysis. In the study by Van der Walk et al,3 CD15 did not stain myeloblasts. However, in another study by Horny et al,6 which evaluated immunoreactivity of CD15 in AML by PIH analysis of BM core biopsies, CD15 stained blasts in 20% of cases. There is no study describing FC analysis of CD15 in AML. Our present study revealed CD15 immunoreactivity of blasts by PIH analysis of BM core sections (24%) and BM clot sections (31%) of AML cases. A higher percentage (45%) of AML cases were CD15^sup +^ by FC analysis. Thus, the FC detection of CD15 appeared to be more reliable than the IH method. In addition, our study revealed an association of CD15^sup +^ blasts in AML cases with a monocytic component.
Since CD117 is a relatively new antibody available for PIH analysis of AMLs, there are no previous data in the literature regarding the immunoreactivity of CD117 in paraffin-embedded BM specimens. Our study revealed 83% of AMLs were CD117^sup +^ by FC analysis and a lesser percentage were CD117^sup +^ by PIH analysis (62% in core sections and only 27% in clot sections). Thus, the IH detection of CD117 appeared to be less reliable than the FC method, especially IH detection in clot sections. This latter phenomenon was most likely due to the decalcification and fixation processes. This discordance was quite marked (62%) in those cases with both BM clot and core material available. Thus, if a BM aspirate is not available for FC analysis, results of the BM core material are more reliable. In agreement with a previous study, there was no correlation of CD34 immunoreactivity with CD117 immunoreactivity and no correlation of CD34 immunoreactivity or CD117 immunoreactivity with the FAB subtype of AML. However, it is important to detect these antigens (CD34 and CD117) by PIH in acute leukemias since BM aspirates may not be available for FC analysis in some cases of acute leukemia due to fibrosis or a packed BM. As noted previously, CD34 expression in adult AML is associated with a worse prognosis and CD117, to our knowledge, has not been identified in acute leukemias of nonmyelogenous origin.
Myeloperoxidase is also a relatively new antibody available for PIH analysis of AMLs. Previous studies have demonstrated that PIH analysis of MPO is more sensitive than EC analysis of MPO.7,18 These studies have demonstrated MPO immunoreactivity by PIH analysis of BM clot and core biopsy sections of AML M0s with MPO negativity by EC analysis. To our knowledge, there has been no description of results of MPO immunoreactivity by PIH analysis of BM clot versus BM core sections. Our study supports the previous studies in that there was a higher percentage of MPO immunoreactivity in the AMLs by PIH analysis (90%, BM core; 92%, BM clot) than by EC analysis (78%), and there were 2 of 3 AML cases originally FAB subtyped as MO with MPO immunoreactivity by PIH analysis. Thus, MPO detection by PIH may alter the FAB subtype of the AML and alter the prognosis. There was complete concordance in the PIH results of MPO by BM clot versus BM core sections.
In summary, our study demonstrates the following findings:
1. Immunodetection of CD34 expression in AML seems to have a similar sensitivity by FC analysis and PIH analysis of BM clot and core sections. CD34 expression by FC analysis and negativity by PIH analysis may be due to low surface expression of CD34. CD34 immunoreactivity by PIH analysis and negativity by FC analysis may be due to sampling or a low percentage of CD34^sup +^ cells.
2. CD15 may be expressed by blasts in AML. Immunodetection of CD15 expression in AML has a higher sensitivity by FC analysis (45%) than by PIH analysis of BM clot (31%) and core (24%) sections. CD15 expression in AML is associated with those cases with a monocytic component.
3. Immunodetection of CD117 expression in AML has a higher sensitivity by FC analysis (83%) than by PIH analysis of BM clot (27%) and core (62%) sections. The sensitivity of detection of CD117 immunoreactivity by PIH analysis of BM clot sections is much less than that of BM core sections.
4. There was no correlation of CD34 or CD117 expression with the FAB subtype of AML.
5. Detection of MPO by PIH analysis of BM clot and core sections is more sensitive (92% and 90%, respectively) than by the EC method (78%). Cases of AML originally FAB subtyped as M0 may be MPO^sup +^ by PIH analysis.
1. Geller RB, Zahurak M, Hurwitz CA, et al. Prognostic importance of immunophenotyping in adults with acute myelocytic leukaemia: the significance of the stem-cell glycoprotein CD34 (MylO). Haematology. 1990;76:340-347.
2. Wells SJ, Bray RA, Stempora LL, et al. CD 117/CD34 expression in leukemic blasts. I Clin Pathol. 1996;106:192-195.
3. Van Der Walk P, Mullink H, Huijgens PC, et al. Immunohistochemistry in bone marrow diagnosis. Am I Surg Pathol. 1989;13:97-106.
4. Shin SS, Sheibani K, Kezirian J, et al. Immunoarchitecture of normal human bone marrow: a study of frozen and fixed tissue sections. Hum Pathol. 1992;23: 686-694.
5. Shin SS, Sheibani K, Kezirian J, et al. Immunohistologic studies of bone marrow biopsies on frozen sections: an analysis of 42 cases. Hum Pathol. 1993; 24:30-36.
6. Horny H-P, 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.
7. Arber DA, Jenkins KA. Paraffin section immunophenotyping of acute leukemias in bone marrow specimens. Am] Clin Pathol. 1996;106:462-468.
8. Chuang S-S, Li C-Y. Useful panel of antibodies for the classification of acute leukemia by immunohistochemical methods in bone marrow trephine biopsy specimens. Am J Clin Pathol. 1997;107:410-418.
9. Horny H-P, Campbell M, Steinke B, et al. Acute myeloid leukemia: immunohistologic findings in paraffin-embedded bone marrow biopsy specimens. Hum PathoL 1990;21:648-655.
10. Kurec AS, Cruz VE, Barrett D, et al. Immunophenotyping of acute leukemias using paraffin-embedded tissue sections. Am) Clin Pathol. 1990;93:502509.
11. Bennett JM, Catovsky D, Daniel M-T, et al. Proposals for the classification of the acute leukaemias: French-American-British (FAB) Cooperative Group. Br] HaematoL 1976;33:451-458.
12. Bennett IM, Catovsky D, Daniel M-T, et al. Proposed revised criteria for the classification of acute myeloid leukemia: a report of the French-AmericanBritish Cooperative Group. Ann Intern Med. 1985;103:626-629.
13. Bennett JM, Catovsky D, Daniel M-T, et al. Proposal for the recognition of minimally differentiated acute myeloid leukaemia (AML-MO). Br J Haematol. 1991;78:325-329.
14. Bennett JM, Catovsky D, Daniel M-T, et al. Criteria for the diagnosis of acute leukemia of megakaryocyte lineage (M7): a report in the French-AmericanBritish Cooperative Group. Ann Intern Med. 1985;103:460-462.
15. Dunphy CH, Dunphy FR, Visconti JL. Flow cytometric immunophenotyping of bone marrow core biopsies: report of 8 patients with previously undiagnosed hematological malignancy and failed bone marrow aspiration. Arch Pathol Lab Med. 1999;123:206-212.
16. Hanson CA, Ross CW, Schnitzer B. Anti-CD34 immunoperoxidase staining in paraffin sections of acute leukemia: comparison with flow cytometric immunophenotyping. Hum Pathol. 1992;23:26-32.
17. 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 I Clin Pathol. 2000;113: 814-822.
18. Kotylo PK, Seo I-S, Smith FO, et al. Flow cytometric immunophenotypic characterization of pediatric and adult minimally differentiated acute myeloid leukemia (AML-MOi. Am I Clin Pathol. 2000;113:193-200.
Accepted for publication April 9, 2001.
From the Division of Hematopathology, Department of Pathology, St Louis University Health Sciences Center, St Louis, Mo (Drs Dunphy, Evans, and Gardner); and the Department of Pathology, University of South Alabama, Mobile, Ala (Dr Polski).
Reprints: Cherie H. Dunphy, MD, Department of Pathology and Laboratory Medicine, CB#7525, University of North Carolina, Chapel Hill, NC 27599-7525 (e-mail: email@example.com).
Copyright College of American Pathologists Aug 2001
Provided by ProQuest Information and Learning Company. All rights Reserved