Case 4 -

Mixed Phenotype Acute Leukemia, B/myeloid, NOS Steven H. Kroft Medical College of Wisconsin Milwaukee, WI

Click on slide thumbnail images for an enlarged view.

If you have any difficulties viewing these slides, email the webmaster.

Clinical History: The patient was a previously healthy Hispanic male with no significant past medical history who presented with a 4-week history of fever, weight loss, and fatigue. Physical examination revealed bilateral posterior cervical, axillary, and inguinal adenopathy, but no hepatosplenomegaly. Pertinent Laboratory Data: CBC revealed WBC 390,000/uL, hemoglobin 9.5 g/dL, and platelets 72,000/uL. Case 4 - Figure 1 Peripheral blood demonstrating a dimorphic blast proliferation, one population resembling lymphoblasts, the other resembling myeloblasts. The dimorphism was most striking in the aspirate smears. Case 4 - Figure 2 Peripheral blood demonstrating a dimorphic blast proliferation, one population resembling lymphoblasts, the other resembling myeloblasts. The dimorphism was most striking in the aspirate smears. Case 4 - Figure 3 Bone marrow aspirate demonstrating a dimorphic blast proliferation, one population resembling lymphoblasts, the other resembling myeloblasts. The dimorphism was most striking in the aspirate smears. Case 4 - Figure 4 Bone marrow aspirate demonstrating a dimorphic blast proliferation, one population resembling lymphoblasts, the other resembling myeloblasts. The dimorphism was most striking in the aspirate smears. Case 4 - Figure 5 Bone marrow biopsy section revealing primitive-appearing blasts of varying size. Case 4 - Figure 6 A myeloperoxidase cytochemical stain is negative in the blasts. Case 4 - Figure 7 A non-specific (butyrate) esterase stain is negative in the blasts. non-specific (butyrate) esterase Case 4 - Figure 8 Flow cytometry (gated on low side scatter events over a wide range of CD45 expression) reveals co-expression of B-lymphoid and myeloid antigens. Many of the antigens show a heterogeneous expression pattern. See text for details. Case 4 - Figure 9 Flow cytometry (gated on low side scatter events over a wide range of CD45 expression) reveals co-expression of B-lymphoid and myeloid antigens. Many of the antigens show a heterogeneous expression pattern. See text for details. Case 4 - Figure 10 Re-analysis of flow cytometry data with cluster analysis and color-eventing reveals three distinct blast population. All have B-lymphoid and myeloid antigen expression, but differ in the degree to which these lineages are expressed. See text for details. Case 4 - Figure 11 Re-analysis of flow cytometry data with cluster analysis and color-eventing reveals three distinct blast population. All have B-lymphoid and myeloid antigen expression, but differ in the degree to which these lineages are expressed. See text for details. Case 4 - Figure 12 Day 14 marrow showing induction failure. The blasts are still dimorphic, but now the myeloblast-like population is dominant. Case 4 - Figure 13 Flow cytometry still shows three blast populations with similar immunophenotypic features as at diagnosis, but now a previously minor population with monocytic features is now dominant. See text for details. Introduction: Immunophenotyping, particularly by flow cytometry, has assumed a critical role in the diagnosis and classification of acute leukemias. However, as immunophenotyping techniques have become more sophisticated, with utilization of antibodies against a broad array of surface and intracellular antigens, the degree to which these malignancies deviate immunophenotypically from their presumed normal counterparts has become apparent. While these differences may be exploited for the purposes of diagnosis and follow-up, they also may cause difficulties in classification.. This case illustrates some of the complexities that may be encountered in the immunophenotypic diagnosis of acute leukemia. Pathological/Microscopic Findings and any Immunohistochemical or Other Studies: Wright-Giemsa stained blood and bone marrow smears demonstrated a dimorphic blast appearance, although this was more distinct in the marrow smears. The predominant population consisted of small blasts with regular nuclei, moderately condensed chromatin, small or inconspicuous nucleoli, and scanty, agraunular cytoplasm. These were cytologically most consistent with lymphoblasts. The minority population consisted of medium to large blasts with regular to mildly notched nuclei, open chromatin, several small but distinct nucleoli, and moderate amounts of pale basophilic, agranular cytoplasm. These blasts more closely resembled myeloblasts. A minority subpopulation of the larger blasts showed more prominent nuclear irregularity. A subset of both populations showed a "hand mirror" appearance, with a unipolor cytoplasmic knob or projection. Very few maturing hematopoietic precursors were evident, and those present showed no overt dysplastic features (not illustrated). The bone marrow biopsy showed variably sized blasts with regular or mildly irregular nuclei, stippled chromatin, variably distinct nucleoli, and small to moderate amounts of cytoplasm. Cytochemical stains for myeloperoxidase and non-specific (butyrate) esterase were negative in the blasts, only highlighting occasional background maturing cells. Flow cytometry demonstrated heterogeneous expression of various antigens on the blasts. Because of highly heterogenous CD45 expression, the histograms provided were gated on all low side scatter events, and thus likely include mature lympohocyte as well as blasts. The quadrants are defined using an isotypic control with a similar scatter gate. Based on these histograms, the blasts have the following immunophenotype: CD10(partial +), CD19(+), CD22(+), intracellular CD79a(partial +), TdT(+), CD11b(partial+), CD13(+), CD15(-), CD33(+), CD36(partial dim+), CD64(partial+), myeloperoxidase(-). It is not clear from these histograms whether CD20 and CD14 are expressed on blast subsets or mature populations. Differential Diagnoses: The blasts in this case co-express B-lymphoid and myeloid antigens. Consequently, the differential diagnosis includes B-acute lymphoblastic leukemia, acute myeloid leukemia, and acute leukemia of ambiguous lineage. In order to resolve this, published schemes for the assignment of lineage in acute leukemia must be applied (see case discussion below). Final Diagnosis: Mixed phenotype acute leukemia, B/myeloid, NOS Case Discussion: It is now well appreciated that both acute lymphoblastic and acute myeloid leukemias very frequently demonstrate cross-lineage antigen expression. When this is present to a pronounced degree, it may result in ambiguities with respect to lineage assignment, and this has led to the concept of acute leukemias of ambiguous lineage (i.e., biphenotypic acute leukemia, bilineal acute leukemia, mixed phenotype acute leukemia, etc.). The diagnostic criteria for these leukemias have been problematic, and the clinical and therapeutic implications unclear. In order to clarify diagnostic criteria for acute leukemias of ambiguous lineage, scoring systems have been devised, the most widely used of which was that of the European Group for the Immunological Characterization of Leukemia (EGIL)(Bene, 1995)(see Table in slide 15 of attached Powerpoint presentation). The basis of the EGIL system is the application of antibodies to lineage-associated antigens for T, B, and myeloid lineages. Each antigen receives a point score (0.5 to 2) based on the perceived specificity of that particular antigen for a particular lineage. For example, CD22 (2 points) is considered to be more specific for B lineage than CD19 (1 point), since CD19 is expressed in some acute myeloid leukemias. In order to establish a particular lineage, the points for antigens associated with that lineage must total 2.5 or greater. Additionally, the EGIL criteria state that greater than 20% of blasts must express an antigen for it to be considered positive. Applying the EGIL criteria in the present case, the blasts express 6.5 points of lymphoid antigens (CD79a--2 points; CD22—2 points; CD19—1 point; CD10—1 point; TdT—0.5 points) and 2 points of myeloid antigens (CD33—1 point; CD13—1 point). While CD64 appears to be expressed by a blast subset, it turns out that fewer than 20% of the blasts exceed the isotypic threshold for this antigen, so it is not considered positive for CD64 by EGIL criteria. Thus, this case would be classified as B-acute lymphoblastic leukemia by EGIL criteria. Note, however, that CD117 and CD65s, two myeloid antigens included in the EGIL system, were not assessed in this case, so an acute leukemia of ambiguous lineage cannot be entirely excluded. Since the introduction of the EGIL system, several shortcomings have become apparent. Because cross-lineage antigen expression is so common in acute leukemias [in a recent series of 200 B-acute lymphoblastic leukemias (Seegmiller, 2009), expression of at least one myeloid antigen was seen in 86.5% of cases], any criteria for establishing mixed lineage are necessarily arbitrary; no gold standard exists to guide their development. The EGIL system does not include all B, T, and myeloid antigens in use in clinical flow cytometry laboratories, and conversely, include antigens that are not widely assessed (e.g., CD65s, cytoplasmic mu heavy chain). The system does not evolve with improved reagents, processing procedures, analysis techniques, and broader knowledge regarding the spectrum of antigens expressed in various malignancies. For example, the choice of a brighter over a dimmer fluorochrome (e.g., PE v. FITC) can significantly impact antigen assessment, as can the choice of one red cell lysis procedure over another. Such technical issues are particularly problematic given the 20% threshold criterion. Such thresholds, when applied to dimly or partially expressed antigens, are impossible to standardize across laboratories. Even if all processing procedures and reagents are identical, differing instrument settings (photomultiplier tube voltages and compensation settings) produce differing results for thresholds applied to dimly or partially expressed antigens. Thresholding also fails to capture the diverse qualitative patterns seen in hematological malignancies (strong v. weak expression, partial expression, subset positivity, etc.). The EGIL criteria also are primarily applicable to cases in which relatively uniform blast populations co-express antigens of more than one lineage (so-called "biphenotypic acute leukemia"), and fails to specifically address situations where more than one immunophenotypically distinct blast population is present ("bilineal acute leukemia"). Finally, the perception of many practitioners of flow cytometry is that with improved reagents and techniques, biphenotypia by EGIL criteria was becoming unsettlingly common. In the Seegmiller series (2009),12.5% of B-ALL cases satisfied strict criteria for biphenotypia. A commonly encountered scenario is that a case will show characteristic features of either ALL or AML (morphologically, immunophenotypically, and cytogenetically), but have the bare minimum of cross lineage antigen expression (often dimly expressed) for biphenotypia by EGIL criteria. Rather than call such cases biphenotypic, my practice was to diagnose them based on their dominant features, and add a comment that they satisfy minimal EGIL criteria for biphenotypia. In other words, there was a strong impression that the EGIL was forcing us to call too many things biphenotypic (although this judgment is, of course, based on arbitrary criteria). Partially in response to some of these concerns, the 2008 WHO classification provided a substantially different set of criteria for assigning a diagnosis of mixed phenotype acute leukemia (MPAL). The criteria are simplified, requiring assessment of fewer antigens, but more stringent, in that they require expression of antigens perceived to be highly lineage specific. For assignment of myeloid lineage, the WHO requires at least one of the following: 1) Myeloperoxidase expression, or 2) Monocytic differentiation, characterized by at least two of NSE, CD11c, CD14, CD64, and lysozyme. Additionally, while CD36 is not mentioned in the Table in the WHO monograph, it is mentioned in the body of the text. For T lineage, cytoplasmic or surface CD3 must be present. Finally, for designation of B lineage one of the following needs to be satisfied.: 1) Strong CD19 with at least one of CD79a, CD22, and CD10 strongly expressed, or 2) Weak CD19 with at least two of CD79a, CD22, and CD10 strongly expressed. Thus, criteria for B lineage appear to be satisfied in the present case. However, since myeloperoxidase is negative, unless monocytic differentiation can be demonstrated, the case would have to be considered B-ALL. There does appear to be some expression of CD36 and CD64, and possibly CD14 as well. But is there enough expression? Unlike the EGIL, the WHO 2008 does not provide specific percentages of blasts that need to be positive for a given antigen to be considered positive, so we appear to be left in an uncertain position regarding myeloid lineage. However, the WHO does provide us another potential avenue for assignment of myeloid lineage. Specifically, in addition to a "biphenotypic" pattern, whereby either myeloperoxidase or monocytic antigens are expressed on the same blast population that expresses lymphoid antigens, the WHO 2008 allows for the possibility that one of two or more immunophenotypically distinct blast populations would satisfy, by itself, criteria for AML ("bilineal pattern"). Importantly, it is not required that such a population itself constitute greater than 20% of marrow or blood cells. Consequently, it is important to revisit the flow cytometry in this case. It appears from the dot plots provided that the blasts are heterogeneous with respect to expression of a number of antigens, but whether there are truly discrete, reproducible sub-populations of blasts is not clear. However, if one applies color-eventing, based on an iterative analysis between multiple fluorescence and/or scatter plots in the individual tubes, as well as between tubes, one can discriminate three discrete blast populations in this case. The dominant population, constituting 67% of events, has low side scatter, variable forward scatter, dim CD45, and is CD34(+), CD4(-), CD10(partial+), CD11b(partial+), CD13(+), CD14(-), CD15(-), CD19(+), CD20(-), CD22(partial+), CD33(partial+), CD34(+), CD36(partial dim+), CCD64(-), CD79a(+), MPO(-), TdT(+). The second population, constituting 2% of events, differs from the first in that it is smaller in size by forward light scatter characteristics, shows moderate expression of CD45, expresses CD20 and is brighter for CD19 and CD22, and shows less expression of CD13 and CD33. In other words, the second population more closely resembles a typical B-ALL. Finally, the third population is distinct from the dominant population in that it is of more uniform medium-to large size and shows slightly higher side scatter (light scatter characteristics similar to monocytes), is bright for CD45, brighter for CD13, CD33, and CD11b, positive for CD4 and CD64, predominantly dim positive for CD36, and dimmer for CD19. Additionally a minority subset of this population expresses CD14. These findings taken in aggregate indicate a mixed biphenotypic (cross-lineage antigens expressed on the same population of blasts) and bilineal (several discrete populations of blasts) pattern. Whether this case is considered an MPAL depends on the third population of blasts. Although blasts with monocytic differentiation typically express CD15 and show brighter CD36 expression in my experience, the combination of CD64, CD36, and the CD14(+) subset appears to satisfy WHO criteria for monocytic differentiation; the other immunotypic features (light scatter, CD45, and CD4) bolster the impression of monocytic differentiation. Thus, despite the fact that this population only represents 5% of events, we can apply the designation of MPAL to this case. In order to finally classify this case, a few more words about the WHO 2008 classification are required. This system further sub-divides acute leukemias of ambiguous lineage according to cytogenetic features and immunophenotype. The first category, acute undifferentiated leukemia, is a rare group in which insufficient lineage-associated markers are present to assign any lineage. Little is known about this entity. The next two groups are cytogenetically defined—based on whether they carry either a t(9;22)(q34;q11.2) or a t(v;11q23)—irrespective of immunophenotype. The next two groups are B/myeloid, NOS, and T/myeloid, NOS, in, immunophenotypically defined when the above cytogenetic abnormalities are absent. Finally, there are MPAL, NOS, rare types, cases which are not classifiable according to the above mentioned immunophenotypic scheme. It is notable that whether there is a biphenotypic or bilineal immunophenotypic pattern does not impact classification in the WHO 2008 scheme. The present case was found to contain an inv(5)(q13q33), a non-specific abnormality. Consequently, my final diagnosis in this case is "Mixed phenotype acute leukemia, B/myeloid, NOS." Following diagnosis of acute biphenotypic leukemia, the patient received cytarabine and idarubicin, followed by fludarabine, cytarabine, and G-CSF (FLAG regimen), an AML regimen. Re-marrow on day 14 revealed frank induction failure. Ultimately, the patient died 5 weeks from presentation. While the leukemic blasts maintained a biphasic appearance in the day 14 marrow, very interestingly the dominant population was now the larger, myeloblast-like population, with the lymphoblast-like subset now in the minority. By flow cytometry, three distinct blast populations were still discernable, immunophenotypically similar but not identical to the original three populations. Notably, the monocyte-like population now represented the dominant population (58%). The originally dominant population now comprised 33%, and the small ALL-like population now constituted only 0.13%. This evolution toward a more dominant myeloid appearance would seem to support the original designation of mixed lineage leukemia. Review of the Literature/Treatment Options: Data regarding the clinicopathologic features of MPAL are sparse, and are largely based on cases classified according to the EGIL criteria or similar schemes. Because of the more restrictive nature of the WHO 2008 criteria, it is expected that the frequency of this disease will decrease rather dramatically. In the Seegmiller series of B-ALL referred to earlier, only 1 of 25 cases that satisfied EGIL criteria for biphenotypia satisfied WHO 2008 criteria, although there was likely a selection bias. Consequently, not only will much of the existing data become inapplicable, accrual of new data on large numbers of cases will become extremely difficult. The discussion below applies to older data based on pre-WHO 2008 criteria. Biphenotypic acute leukemia, as defined by the EGIL, has been reported to comprise approximately 4% of acute leukemias, although the exact percentage has varied considerably across studies. Clinically, these leukemias present without distinct clinical features. Immunophenotypically, about two thirds show a B/myeloid pattern, with the majority of the remainder being T/myeloid. T/B and triphenotypic cases occur, but are very unusual. A biphenotypic pattern is much more common than a bilineal pattern, but this does not impact classification. Interestingly, as in the present case, the blasts often show dimorphic cytology, and this does not correlate with whether the immunophenotype shows a biphenotypic or bilineal pattern. The most common cytogenetic abnormality is the t(9;22), seen in about 35-40% of cases, followed by 11q23 rearrangements in 5-10% of cases. Complex karyotypes are not uncommon. In adults, the prognosis of biphenotypic acute leukemia is poor, with overall survival of only 17% at two years and 8% at 4 years (Legrand, 1998; Killick, 1999), worse than age-matched ALL and AML cases. However, children appear to fare much better, with 75% overall survival at 2 years (Killick, 1999), with no difference in survival compared to age matched AML and ALL cases. Based on limited data, there is no evidence that immunophenotypic features impact prognosis. However patients with a t(9;22) appear to fare the worst. Whether or not biphenotypia, per se, is a poor prognostic indicator independent of adverse cytogenetics is not clear. The optimal therapy for MPAL remains to be determined; randomized trials do not exist do the rarity of this disease. Often either AML or ALL therapy is used, depending on which immunophenotypic pattern is dominant. Some oncologists employ regimens that combine features of ALL and AML regimens. Killick et al (2009) found that this resulted in a high rate of early treatment deaths. Conclusion(s): Whether the WHO 2008 scheme is an improvement, i.e., whether it allows us to better define discrete clinicopathologic entities among the acute leukemias and development of better therapy, remains to be seen. The WHO criteria are clearly more stringent, but are ultimately arbitrary in the same way that the EGIL criteria are arbitrary. While the WHO has gotten rid of difficult to standardize quantitative thresholds, it has added difficult to standardize assessments of "weak" and "strong" antigen expression on the B-cell side. While such assessments are consonant with the intuitive sense that stronger antigen expression imparts more lineage specificity, this is a difficult premise to test in the setting of ambiguous lineage leukemias. We thus await new clinical and pathological investigation using the WHO 2008 criteria to further enhance our understanding of this difficult area. References: Anonymous. The value of c-kit in the diagnosis of biphenotypic acute leukemia. EGIL (European Group for the Immunological Classification of Leukaemias). Leukemia. 1998;12:2038. Bene MC, Castoldi G, Knapp W, et al. Proposals for the immunological classification of acute leukemias. European Group for the Immunological Characterization of Leukemias (EGIL). Leukemia. 1995;9:1783-1786. Borowitz MJ, Bene MC, Harris NL, Porwit A, Matutes E. Acute leukemias of ambiguos lineage. In: Swerdlow SH, Campo E, Harris NL, et al., eds. WHO Classification of Tumours of Haematopoietic and Lymphoid Tissues. Lyon: IARC; 2008:150-155. Carbonell F, Swansbury J, Min T, et al. Cytogenetic findings in acute biphenotypic leukaemia. Leukemia. 1996;10:1283-1287. Killick S, Matutes E, Powles RL, et al. Outcome of biphenotypic acute leukemia. Haematologica. 1999;84:699-706. Legrand O, Perrot JY, Simonin G, et al. Adult biphenotypic acute leukaemia: an entity with poor prognosis which is related to unfavourable cytogenetics and P-glycoprotein over- expression. Br J Haematol. 1998;100:147-155. Matutes E, Morilla R, Farahat N, et al. Definition of acute biphenotypic leukemia. Haematologica. 1997;82:64-66. Owaidah TM, Al Beihany A, Iqbal MA, Elkum N, Roberts GT. Cytogenetics, molecular and ultrastructural characteristics of biphenotypic acute leukemia identified by the EGIL scoring system. Leukemia. 2006;20:620-626. Rubio MT, Dhedin N, Boucheix C, et al. Adult T-biphenotypic acute leukaemia: clinical and biological features and outcome. Br J Haematol. 2003;123:842-849. Seegmiller AC, Kroft SH, Karandikar NJ, McKenna RW. Characterization of immunophenotypic aberrancies in 200 cases of B acute lymphoblastic leukemia. Am J Clin Pathol. 2009;132:940-949. Tiribelli M, Damiani D, Masolini P, Candoni A, Calistri E, Fanin R. Biological and clinical features of T-biphenotypic acute leukaemia: report from a single centre. Br J Haematol. 2004;125:814-815. Weir EG, Ali Ansari-Lari M, Batista DA, et al. Acute bilineal leukemia: a rare disease with poor outcome. Leukemia. 2007;21:2264-2270.