Acute Leukemias of Ambiguous Origin
The issue of MPO expression in the diagnosis of MPAL has been discussed in the literature. Previous French-American-British group guidelines, based on cytochemistry, used 3% of MPO-positive blasts in BM smears as sufficient to call a leukemia MPO positive. A threshold of 10% expression has been used by the EGIL group. However, discordant cases are found when both methods are compared, leading to the conclusion that the 10% threshold may be conservative but not very sensitive. It should be emphasized that MPO expression can be a difficult test to establish by FCM immunophenotypic analysis since differences in permeabilization reagents and various antibodies have been reported. Thus, MPO expression in leukemic blasts has to be compared with internal negative and positive controls. In a study comparing MPO expression detected by FCM and cytochemistry in cases of AML and ALL, a 13% threshold was found to be relevant using an isotype control as a background reference (sensitivity, 95.1%; specificity, 91.7%). If residual normal lymphocytes were used as reference, a threshold of 28% had to be applied, yielding an improved 97.4% sensitivity and 96.1% specificity in distinguishing between ALL and AML. The WHO criteria for MPAL do not indicate any lower limit for MPO expression in leukemic blasts, but it seems reasonable to use published thresholds. Since MPAL cases often show more than one population of blasts, MPO could be present in only a minor population, which has to be identified as the myeloid component. Another issue is that AML with minimal differentiation and no MPO expression could be involved in MPAL. Therefore, as stated in the WHO criteria, when there are two or more distinct populations of leukemic cells, one with coexpression of a number of myeloid markers, no MPO and no lymphoid markers also can be accepted to define the myeloid component of MPAL.
Rarely, cases of otherwise typical B-ALL/lymphoma can express MPO by FCM or IHC (case 259) or least often by cytochemistry. These cases are best classified and treated as B-ALL/lymphoma. In contrast, most MPAL cases show expression of other myeloid-associated markers and also are characterized by the presence of two or more subpopulations of blasts with slightly different immunophenotypes. In some cases of MPAL, MPO may be expressed only in a small subset of blasts, with or without lymphoid markers (biphenotypic or bilineal presentation).
Cytoplasmic CD3 (cCD3) is considered the most specific marker for the T-cell lineage and has been used to diagnose T-cell malignancies since the 1980s. FCM immunophenotyping methods to detect cCD3 were established in the 1990s when permeabilization reagents became available. As in the case of MPO, expression of cCD3 in leukemic blasts has to be determined in comparison with internal negative and positive controls (B cells, monocytes, granulocytes, and normal T lymphocytes, respectively). At least a fraction of blasts should express cCD3 at the level of normal T cells, and weak expression in a minor fraction of blasts is insufficient to diagnose MPAL. Many AML cases show aberrant expression of other T-cell–associated markers, such as CD2, CD4, CD5, or CD7, or NK-cell–associated markers, such as CD56. Thus, the correct interpretation of cCD3 staining is important for final diagnosis. If a BM biopsy specimen is available, the expression of cCD3 may be confirmed by IHC. In case 261, positive for CD2, CD4, and CD56, review of the FCM files revealed no sufficient cCD3 expression, and no confirmatory IHC could be performed. Thus, considering the presence of a complex karyotype, a consensus diagnosis of AML with myelodysplasia-related changes was rendered Table 5 . Similarly, in case 349 ( Table 5 ), expression of cCD3 was very weak and could not be confirmed by IHC. Also, in case 303 ( Table 2 ), heterogeneous expression of cCD3 was not confirmed by IHC, although a fraction of blasts that showed adequately high expression to qualify as MPAL was observed by FCM, together with CD5 and CD7. No cCD3 expression was seen at relapse in this patient.
What Are Some of the Important Technical and Interpretation Issues in the Diagnosis of MPAL?
MPO
The issue of MPO expression in the diagnosis of MPAL has been discussed in the literature. Previous French-American-British group guidelines, based on cytochemistry, used 3% of MPO-positive blasts in BM smears as sufficient to call a leukemia MPO positive. A threshold of 10% expression has been used by the EGIL group. However, discordant cases are found when both methods are compared, leading to the conclusion that the 10% threshold may be conservative but not very sensitive. It should be emphasized that MPO expression can be a difficult test to establish by FCM immunophenotypic analysis since differences in permeabilization reagents and various antibodies have been reported. Thus, MPO expression in leukemic blasts has to be compared with internal negative and positive controls. In a study comparing MPO expression detected by FCM and cytochemistry in cases of AML and ALL, a 13% threshold was found to be relevant using an isotype control as a background reference (sensitivity, 95.1%; specificity, 91.7%). If residual normal lymphocytes were used as reference, a threshold of 28% had to be applied, yielding an improved 97.4% sensitivity and 96.1% specificity in distinguishing between ALL and AML. The WHO criteria for MPAL do not indicate any lower limit for MPO expression in leukemic blasts, but it seems reasonable to use published thresholds. Since MPAL cases often show more than one population of blasts, MPO could be present in only a minor population, which has to be identified as the myeloid component. Another issue is that AML with minimal differentiation and no MPO expression could be involved in MPAL. Therefore, as stated in the WHO criteria, when there are two or more distinct populations of leukemic cells, one with coexpression of a number of myeloid markers, no MPO and no lymphoid markers also can be accepted to define the myeloid component of MPAL.
Rarely, cases of otherwise typical B-ALL/lymphoma can express MPO by FCM or IHC (case 259) or least often by cytochemistry. These cases are best classified and treated as B-ALL/lymphoma. In contrast, most MPAL cases show expression of other myeloid-associated markers and also are characterized by the presence of two or more subpopulations of blasts with slightly different immunophenotypes. In some cases of MPAL, MPO may be expressed only in a small subset of blasts, with or without lymphoid markers (biphenotypic or bilineal presentation).
Cytoplasmic CD3
Cytoplasmic CD3 (cCD3) is considered the most specific marker for the T-cell lineage and has been used to diagnose T-cell malignancies since the 1980s. FCM immunophenotyping methods to detect cCD3 were established in the 1990s when permeabilization reagents became available. As in the case of MPO, expression of cCD3 in leukemic blasts has to be determined in comparison with internal negative and positive controls (B cells, monocytes, granulocytes, and normal T lymphocytes, respectively). At least a fraction of blasts should express cCD3 at the level of normal T cells, and weak expression in a minor fraction of blasts is insufficient to diagnose MPAL. Many AML cases show aberrant expression of other T-cell–associated markers, such as CD2, CD4, CD5, or CD7, or NK-cell–associated markers, such as CD56. Thus, the correct interpretation of cCD3 staining is important for final diagnosis. If a BM biopsy specimen is available, the expression of cCD3 may be confirmed by IHC. In case 261, positive for CD2, CD4, and CD56, review of the FCM files revealed no sufficient cCD3 expression, and no confirmatory IHC could be performed. Thus, considering the presence of a complex karyotype, a consensus diagnosis of AML with myelodysplasia-related changes was rendered Table 5 . Similarly, in case 349 ( Table 5 ), expression of cCD3 was very weak and could not be confirmed by IHC. Also, in case 303 ( Table 2 ), heterogeneous expression of cCD3 was not confirmed by IHC, although a fraction of blasts that showed adequately high expression to qualify as MPAL was observed by FCM, together with CD5 and CD7. No cCD3 expression was seen at relapse in this patient.