Prognostic Significance of Immunophenotypic and Karyotypic Features of Philadelphia positive B-Lymphoblastic Leukemia in the Era of Tyrosine Kinase Inhibitors

Abstract

Background

Philadelphia (Ph) positive B-lymphoblastic leukemia exhibits immunophenotypic, karyotypic, and molecular genetic heterogeneity. The prognostic significance of these parameters was assessed in the context of intensive tyrosine kinase inhibitor (TKI)-based chemotherapy.

Methods

We studied 65 cases of adult Ph-positive ALL treated with TKI-based therapy, correlated their clinicopathologic heterogeneity with patient outcome, and compared the findings with 60 cases of adult diploid B-ALL treated with similar chemotherapy without TKI.

Results

Ph-positive ALL was associated with older age (p=0.01) and common-B immunophenotype characterized by higher a frequency of co-expression of CD13 (p=0.004), CD66c (p=0.007), and CD25 (p<0.001) with a lower frequency of CD15 expression (p<0.001). Conventional karyotypic analyses showed the Ph chromosome was the sole abnormality in 19 cases (30%), present with other aberrancies in 43 cases (65%), and absent (detectable only by fluorescence in situ hybridization [FISH] or quantitative RT-PCR) in 3 cases (5%). BCR-ABL was confirmed in all cases by FISH or RT-PCR [p190 in 49 (77%) and p210 in 15 (23%), respectively]. Supernumerical Ph correlated with a higher incidence of CD20 expression (p<0.001), whereas p210 correlated with aberrant CD25 expression (p=0.05). Outcomes were not influenced by the degree of karyotypic complexity (including presence or absence of supernumerical Ph); CD20 expression, or myeloid antigen expression (CD13, CD33, CD66c). CD25 expression was associated with an inferior survival in univariate analysis (p=0.051), but not by multivariable analysis (p=0.092).

Conclusions

In the context of intensive TKI-based chemotherapy, the immunophenotypic, karyotypic, and molecular heterogeneity of Ph-positive ALL no longer influences outcome.

INTRODUCTION

B-lymphoblastic leukemia is characterized by proliferation of lymphoblasts committed to B-cell lineage arrested at an early stage of B-cell maturation. Aside from age, karyotypic abnormalities are highly associated with prognosis and can be used stratify patients into standard or high risk groups.^1,2 In adults with B-lymphoblastic leukemia, the most common cytogenetic abnormality is the Philadelphia chromosome (Ph), accounting for approximately 20% – 30% of adult ALL cases overall, with the incidence rising to over 50% in patients aged 50 years or older.³ Several studies have confirmed the adverse prognostic influence of this karyotype in the pre-tyrosine kinase inhibitor (TKI) era, with long-term disease-free survival (DFS) rates rarely exceeding 20% in the absence of allogeneic stem cell transplantation (SCT).^4–7 The Ph chromosome is the result of the reciprocal translocation t(9;22)(q34;q11), which creates a novel fusion transcript, BCR-ABL.^8–10 This translocation ultimately results in a constitutively active tyrosine kinase protein. The different breakpoints in BCR on chromosome 22 result in proteins of varying sizes, such as the 190 kDa (p190) protein observed exclusively in Ph-positive ALL or the 210 kDa (p210) comprising 20% – 40% of Ph-positive ALL and nearly all cases of chronic myelogenous leukemia (CML).¹¹ Ph-positive ALL is also characterized by significant karyotypic heterogeneity, with monosomy 7, supernumerary Ph with +der(22)t(9;22), trisomy 8, and deletion 9p as the most frequently identified aberrancies.^7,12,13 Several studies have shown that the latter 3 have correlated with an especially adverse prognosis. Approximately 15% of ALL patients with the Ph chromosome will also have the favorable feature of high hyperdiploidy (modal chromosome number 51-67 chromosomes).^2,7

Virtually all cases of B-lymphoblastic leukemia demonstrate immunophenotypic aberrancies that can be routinely detected using flow cytometric methodology. These aberrant immunophenotypes often correlate with specific chromosomal abnormalities, thereby influencing prognosis.^14,15 Ph-positive ALL in particular has typically been associated with a high frequency of myeloid antigen (e.g., CD13, CD33, CD66c),^16–18 and CD25 (interleukin-2 receptor alpha chain)¹⁹ expression. Prior reports detailing the significance of immunophenotype in Ph-positive ALL have been limited by small numbers and/or comparisons to subtypes of Ph-negative B-lymphoblastic leukemia with gene rearrangements characteristically associated with certain phenotypic aberrancies. For example, the MLL gene is associated with lack of CD10 expression and aberrant CD15 expression.²⁰ TEL-AML1 frequently expresses CD13 and CD33 and less so CD15; and E2A-PBX1 seldom expresses CD13 or CD33.²¹

The majority of reports detailing the outcome of Ph-positive ALL with respect to immunophenotypic, karyotypic and molecular subtypes were performed prior to use of intensive TKI-based chemotherapy regimens. In this study, various pathophysiological parameters were assessed in a cohort of uniformly treated adults with Ph-positive ALL, and compared to a cohort of adults with Ph-negative diploid karyotype B-lymphoblastic leukemia.

MATERIALS AND METHODS

Study Group

The diagnosis of B-lymphoblastic leukemia was rendered according to criteria established by the World Health Organization.²² Cases with ambiguous lineage or lymphoid blast phase of CML were excluded. Subjects referred to the University of Texas MD Anderson Cancer Center (MDACC) from May 1, 2004 through July 31, 2009 were reviewed in accordance with protocols approved by the Institutional Review Board at MDACC. All cases with active Ph-positive ALL that had at least one diagnostic bone marrow aspirate sample submitted for flow cytometric immunophenotyping (FCI), conventional cytogenetic analysis, fluorescence in situ hybridization (FISH), and reverse transcription-polymerase chain reaction (RT-PCR) were included.

Frontline chemotherapy was administered according to the hyper-CVAD regimen (hyperfractionated cyclophosphamide, vincristine, doxorubicin, dexamethasone alternating with high dose methotrexate and cytarabine) in adults with Ph negative (confirmed by negative FISH for BCR-ABL on bone marrow smear) diploid karyotype B-lymphoblastic leukemia as previously described in detail.²³ Ph-positive ALL patients were treated with the hyper-CVAD regimen concurrently with either imatinib mesylate or dasatinib as previously reported.^24–26 Allogeneic stem cell transplantation (SCT) was performed in first CR for all Ph-positive ALL cases as feasible.

Flow Cytometric Immunophenotyping

FCI was performed using a panel of antibodies designed for acute leukemia further analyzed by an extended panel designed for ALL. The acute leukemia screening panel included CD2, CD3 (surface and cytoplasmic), CD5, CD7, CD10, CD13, CD14, CD15, CD19, CD20, CD33, CD34, CD38, CD41, CD45, CD56, CD64, CD117, HLA-DR, terminal deoxynucleotidyl transferase (TdT), and myeloperoxidase (MPO). The ALL extended panel included CD1, CD4, CD8, cytoplasmic CD22, CD79a, cytoplasmic IgM, surface kappa and lambda, CD25, CD58, CD66c, and CD81. The minimal residual disease (MRD) panel included CD10, CD13, CD15, CD19, CD20, CD22, CD25, CD33, CD38, CD58, and CD81. Bone marrow aspirate specimens were collected using standard techniques and placed in EDTA tubes. After incubation with monoclonal antibodies for 10 min at 4°C, erythrocytes were lysed with ammonium chloride (OrthoLyse, BD Biosciences, San Diego, CA) using a standard lyse/wash technique. For the detection of cytoplasmic antigens, fixation and permeabilization steps were taken before staining using the InterPrep kit (Beckman-Coulter, Fullerton, CA) for TdT, and the Fix & Perm kit (Caltag Laboratories, Burlingame, CA) for MPO. All other antibodies were obtained from BD Biosciences (San Jose, CA). Data were acquired on FACSCalibur cytometers using CellQuest software (BD Biosciences) for the initial screening assay and the MRD panel of markers was acquired on FACSCalibur or FACSCanto II instruments (BD Biosciences) and analyzed using CellQuest, FlowJo (Treestar, Ashland, OR) or FCS Express software (De Novo Software, Los Angeles, CA). CD19/CD34/SSC was used to define the blast population, and isotype-matched monoclonal antibodies with irrelevant specificities were used as negative controls. An arbitrary cutoff of 20% or more analyzed events brighter than the control stain was required for an antigen to be considered positive for CD10, CD13, CD15, CD20, CD33, CD34, CD66c or cytoplasmic IgM. In addition, for comparison purposes, expression levels of CD10, CD20, CD34 and cytoplasmic IgM were semi-quantitatively graded as negative (<20%), partial (20% – 75%), or positive (>75%). CD25 expression was reported as a percentage of positive lymphoblasts. A cutoff of 30% was calculated according to the optimal cutoff values obtained with receiver-operating characteristic (ROC) curves to better distinguish the Ph-positive from the diploid cases. Antigen expression was verified by reviewing raw data (dot plots). Immunophenotypic classifications were grouped according to the criteria of the European Group for the Immunological Characterization of Leukemia (EGIL).²⁷

Conventional Karyotyping and Fluorescence in Situ Hybridization (FISH)

Conventional chromosomal analysis was performed on G-banded metaphase cells prepared from unstimulated bone marrow aspirate cultures using standard techniques. Twenty metaphases were analyzed and the results were reported using the International System for Human Cytogenetic Nomenclature.²⁸

FISH for BCR-ABL was performed on interphase nuclei, using the Vysis LSI BCR-ABL ES dual-color translocation probe (Abbott Molecular, Des Plaines, IL). The cutoff to define a positive result for BCR-ABL was 1.5%.

Real-Time Quantitative RT-PCR Assay

Levels of BCR-ABL fusion transcripts were quantified in a multiplex RT-PCR assay that simultaneously detected b2a2, b3a2, and e1a2 transcripts. RNA was extracted from bone marrow samples using Trizol reagent (Gibco-BRL, Gaithersburg, MD) according to the manufacturer’s instructions. Reverse transcription was performed on total RNA (1 μg), using random hexamers and SuperScript II reverse transcriptase (Gibco-BRL). The resulting complementary DNA was subjected to PCR to amplify BCR-ABL fusion transcripts in an ABI PRISM 7700 Sequence Detector (Perkin Elmer/Applied Biosystems, Foster City, CA) using primers and conditions as previously described.²⁹

Statistical Analysis

The Mann-Whitney test was used for numerical comparison between two cohorts. Fisher exact and chi-square tests were applied for categorical variables. Survival was estimated by the Kaplan–Meier method from the date of diagnosis until death from any cause (censored at last follow-up).³⁰ Disease-specific survival (DSS) was measured from diagnosis until death deemed a direct result of the B-lymphoblastic leukemia. Survival curves were compared via the log-rank test.³¹ Multivariable analysis was performed by the Cox proportional regression model to examine the relationship between survival and age, leukocyte count, CD20 expression, and CD25 expression.³²

RESULTS

Study Group

Sixty-five adults with newly diagnosed Ph-positive ALL were included in this analysis. Forty-four (68%) patients were initially diagnosed and treated at MDACC whereas 21 (34%) were referred after minimal prior therapy. A contemporaneous cohort of 60 adults with newly diagnosed diploid karyotype B-lymphoblastic leukemia was identified. The absence of BCR-ABL1 was confirmed by FISH and/or quantitative RT-PCR in all cases and the absence of MLL gene rearrangements was verified by FISH in 7 cases with insufficient metaphases for complete karyotyping. This group included 38 (63%) newly diagnosed cases and 22 (37%) who were minimally treated prior to referral. Allogeneic SCT was performed in first CR after hyper-CVAD and TKI (either imatinib or dasatinib) in 9 (14%) of the Ph-positive ALL cases, similar to the historical rate of 23% after hyper-CVAD in the pre-TKI era.²⁴ In terms of clinical features, Ph-positive ALL cases were significantly older than their diploid B-lymphoblastic counterparts (median age 53 versus 39 years, p=0.001) (Tables 1, 2). Median leukocyte count was higher in the Ph-positive subset (16 × 10⁹/L versus 11 × 10⁹/L, p=0.041) compared with the diploid B-lymphoblastic leukemia cases.

Table 1.

Clinical, Immunophenotypical, Karyotypic, and Molecular features of 65 adults with Ph-positive B-lymphoblastic Leukemia

Parameter	No. (%)
Median age (yrs)	53 (range, 18 – 85)
Median leukocyte count (× 109/L)	16 (range, 0.4 – 279)
Bcr-abl fusion transcripts by RT-PCR (n=64)
p190	49 (77)
p210	14 (22)
• Both b3a2 and b2a2	3
Others (e1a3)	1 (1)
Karyotype
• Diploid, BCR-ABL by FISH only	6 (9)
• t(9;22)(q34;q11.2) as sole abnormality	17 (26)
• t(9;22)(q34;q11.2) plus others	42 (65)
• Monosomy 7	12
• Deletion 9p	2
• Trisomy 8	1
FISH for supernumerical Ph chromosome (n=63)	23 (37)
Median survival (months)	27 (range, 1 – 54)
Alive	24 (37)
Dead	41 (63)

Table 2.

Immunophenotypical comparison of adult cases of Ph-positive B-lymphoblastic leukemia versus diploid karyotype B-lymphoblastic leukemia

Parameter	Diploid karyotype (n=60)	Ph-positive ALL (n=65)	P-value
Median age (yrs)	39 (range, 18 – 79)	53 (range, 18 – 85)	0.001
Gender (M/F)	38/22	33/32	0.207
Marker expression, no/total (%)
CD10 negative	9/51 (18)	6/58 (10)	0.414
CD13	21/60 (35)	40/65 (62)	0.004
CD15	14/51 (27)	2/58 (3)	<0.001
CD20 (negative/partial/positive)	27(39)/24(34)/19(27)	26(35)/27(36)/22(29)	0.848
CD25*	13/50 (26)	35/58 (60)	<0.001
CD33	32/59 (54)	41/63 (65)	0.363
CD34 negative	15/60 (25)	2/64 (3)	<0.001
CD66c	19/43 (44)	40/54 (74)	0.007
Cytoplasmic IgM (negative/partial/positive)	20(44)/14(31)/11(25)	30(55)/14(25)/11(20)	0.552
Disease specific-survival (mos)	47 (range, 1 – 107)	27 (range, 1 – 54)	0.026

^*As measured by 30% cutpoint

Cytogenetic and Molecular Features of Ph-positive ALL

The cytogenetic and molecular characteristics of the Ph-positive ALL cohort are shown in Table 1. The t(9;22)(q34;q11.2) was detected by conventional cytogenetic analysis in 59 (91%) of the cases. The Ph was the sole abnormality in 19 (32%) cases and was associated with other abnormalities in 43 (73%) cases. Additional karyotypic abnormalities of interest included monosomy 7 (n=12, 20%); del(9)(p21) (n=2, 3%) and trisomy 8 (n=1, 2%) (Table 1). FISH was performed on 63 cases. Supernumerical Ph was found in 23 (37%) of the cases. Of the 64 evaluable cases, RT-PCR showed p190 BCR-ABL transcripts (e1a2) in 49 (77%) cases and p210 (b2a and/or b3a) BCR-ABL transcripts in 14 (22%). One case harbored e1a3 fusion transcripts. In 3 (5%), the conventional chromosomal analysis showed diploid karyotype with BCR-ABL detectable only by FISH and RT-PCR; possibly related to corticosteroid therapy in one of the 3 cases.

Flow Cytometric Immunophenotyping Studies

All of the cases, whether Ph-positive or diploid in karyotype, expressed CD19 in conjunction with at least one other B-lineage marker (CD22, CD79a, and/or cytoplasmic IgM), and were negative for cytoplasmic CD3 and MPO. Compared with the diploid cases, Ph-positive lymphoblasts showed similar rates of CD10 and CD20 expression, but less frequent rates of decreased or negative CD34 expression (Table 2). In Ph-positive ALL, the cytoplasmic IgM was less frequently expressed whereas TdT (data not shown) was often more brightly expressed. Overall, Ph-positive cases generally exhibited a common-B-ALL immunophenotype (CD10 positive, cytoplasmic IgM negative), whereas the immunophenotypic classifications of the diploid karyotype cases were more heterogeneous and included pro-B (CD10 negative), common-B, precursor-B, and in one case mature-B (surface light chain positive) categorizations. Notably, the Ph-positive cases had a higher frequency of CD13 (62% versus 35%, p=0.05) and CD66c (74% versus 44%, p=0.007) expression compared with the diploid cases and less frequent CD15 expression (3% versus 27%, p<0.001) (Table 2). Using the 30% cutoff, CD25 was more frequently expressed (60% versus 26%, p<0.001) in the Ph-positive subset. Comparisons were not performed for expression of the surface antigens CD81, CD58, and CD38 since data was not available for all cases owing to relatively recent implementation of these markers in our FCI assay.

Within the Ph-positive ALL subgroup, CD25 was more frequently expressed in cases with p210 compared with p190 (10 of 12 (83%) versus 24 of 43 (56%), p=0.05). There were no other significant differences between p210 and p190 with respect to expression of CD13, CD15, CD20, CD33, CD66c or cytoplasmic IgM. Cases with the supernumerical Ph chromosome showed a higher frequency of CD20 expression (p<0.001); but there were no differences in expression of the other surface antigens within the Ph cytogenetic subgroups.

Survival Comparisons

With a median overall follow-up time of 21 months (range, 1 – 107+ months), the median DSS of patients with Ph-positive ALL was 27 months (range, 1 – 54 months) compared with 47 months (range 1 – 107) for diploid karyotype B-lymphoblastic leukemia, p=0.026. Within the Ph-positive ALL group, there were no differences in survival between cases with p190 versus p210 fusion transcripts; Ph chromosome as the sole abnormality versus Ph plus other abnormalities; or presence/absence of the supernumerical Ph chromosome (Table 3). DSS was not correlated with expression (or lack thereof) with respect to the surface antigens CD13, CD33, CD66c or CD20. There was no difference in survival by CD25 expression with the standard cut point of 20% (p=0.123), but worse DSS was noted with CD25 expression when analyzed by the 30% cutpoint (p=0.051). However, CD25 expression was not significant by multivariable analysis after accounting for age, leukocyte count, and CD20 expression (p=0.092). Too few cases expressed CD15 for a valid comparative analysis.

Table 3.

Outcome by specific clinicopathological features of Ph-positive B-lymphoblastic leukemia treated with Tyrosine kinase inhibitor (TKI)-based chemotherapy

Parameter	No.	Median disease specific survival (months, range)	P-value
BCR-ABL			0.625
p190	49	26 (2–54)
p210	14	28 (6–23)
Karyotype			0.963
Sole Ph	17	27 (2–54)
Ph + others	42	26 (1–70)
Supernumerical Ph (n=63)			0.406
Yes	23	24 (1–41)
No	40	28 (2–70)
CD13			0.965
Negative	25	30 (1–70)
Positive	40	27 (2–54)
CD33			0.536
Negative	22	30 (1–70)
Positive	41	27 (2–48)
CD66c			0.714
Negative	14	24 (2–48)
Positive	40	26 (1–70)
CD20			0.256
Negative	30	27 (2–48)
Positive (including partial)	25	29 (1–70)
CD25*			0.030
Negative	23	30 (7–70)
Positive	35	19 (1–43)

^*As measured by 30% cutpoint

DISCUSSION

Our clinicopathological analysis of the features of Ph-positive ALL treated in the TKI era confirmed prior observations that this disease tends to be characterized by older age. The vast majority of the Ph-positive ALL cases were relatively uniform expression of a common-B immunophenotype. In contrast, the diploid karyotype B-lymphoblastic leukemia cases exhibited variable stages of maturation, including pro-B, common-B, precursor-B and rarely mature B immunophenotypes. Of the four myeloid antigens analyzed, the Ph-positive ALL cases expressed CD13 and CD66c more frequently, CD33 at similar levels, and CD15 less frequently than their diploid karyotype counterparts. CD33 expression by B lymphoblasts, especially at a low intensity, is relatively frequent, and can be acquired at relapse in both Ph-positive and negative B-lymphoblastic leukemia.¹⁴ However, in our experience, CD13 and CD33 tended to be expressed with high intensity and brightness in Ph-positive ALL. The prognostic value of myeloid antigen expression in pediatric and adult B-lymphoblastic leukemia has been controversial largely because of the relatively small numbers of patients in adult ALL series and differences in the treatment protocols, further confounded by the lack of segregation by presence or absence of the Ph chromosome.^33,34 In our analysis of a selected group of adults with uniformly treated B-lymphoblastic leukemia, the expression of myeloid antigens did not influence outcome.

Paietta and colleagues¹⁹ previously reported that expression of CD25 (IL-2 receptor alpha chain) appeared to be a surrogate marker for the presence of BCR-ABL. Nakase and colleagues³⁵ had previously demonstrated that CD25 expression as measured by the more sensitive methodology of antigen binding capacity (using mean fluorescence intensity) was an independent adverse predictor of outcome in adult B-lymphoblastic leukemia regardless of Ph chromosome status. In our study, CD25 was expressed at a much higher frequency in Ph-positive ALL compared with diploid karyotype B-lymphoblastic leukemia. CD25 expression (defined as at least 30%) correlated with a shorter disease-related survival in the univariate analysis, but not in the multivariable analysis after accounting for age, leukocyte count, and CD20 expression.

In our study, the incidence of CD20 expression was not different between cases of Ph-positive ALL and diploid karyotype B-lymphoblastic leukemia, nor did its expression have prognostic influence in the Ph-positive subset. We have previously demonstrated that CD20 expression (cutpoint 20%) was an independent adverse prognostic factor in adult B-lymphoblastic leukemia (inclusive of Ph-positive and negative cases), but the significance was mainly confined to the younger subset.³⁶ Similar findings of an association of CD20 expression with an increased incidence of relapse was also observed in the Group for Research on Adult Acute Lymphoblastic Leukemia (GRAALL) 2003 trial which applied a pediatric regimen to younger adults with de novo Ph negative B-lymphoblastic leukemia.³⁷ The incorporation of rituximab into frontline therapy for Ph-negative CD20-positive B-lymphoblastic leukemia appears to improve outcome, particularly in the younger age groups.^23,38 Similar approaches with respect to monoclonal antibody therapy are now being applied to the Ph-positive ALL subsets, although whether CD20 expression exerts a similar adverse influence on this subtype is less clear.³⁹

Our study confirmed the molecular genetic heterogeneity in Ph-positive ALL as previously reported.^2,4,6 In a risk factor analysis of 80 adults with Ph-positive ALL treated per the Japan Adult Leukemia Study Group (JALSG) ALL202 study, multivariable analysis revealed that the presence of secondary chromosome aberrations (at a frequency of 67%) in addition to t(9;22) at diagnosis constituted an independent predictive value for relapse-free survival with an increased risk of treatment failure by 2.8-fold. In our series, approximately 65% of our cases harbored additional chromosomal abnormalities in addition to the Ph chromosome without clinical consequence. Although we identified a few associations such as the higher incidence of CD25 or CD20 expression with p210 BCR-ABL or supernumerical Ph, respectively, neither of these correlated with outcome. Factors which could account for the different findings with respect to aberrant karyotype could be related to differences in therapy between JALSG ALL202 and hyper-CVAD plus TKI regimens, particularly with respect to a more continuous exposure to TKIs in the latter.⁴⁰

In summary, we illustrated the significant immunophenotypic, karyotypic, and molecular genetic heterogeneity of de novo Ph-positive ALL and correlated these parameters with outcome in a uniformly treated population. Prior to the incorporation of TKIs into frontline chemotherapy for Ph-positive ALL, several of these clinicopathological features had prognostic relevance, although they were somewhat overshadowed by the dismal outcomes observed despite application of intensive chemotherapy regimens. The incorporation of TKIs into frontline chemotherapy has significantly improved the outcome for Ph-positive ALL; however, disease recurrence continues to be problematic.^{24–26,40,41} Unfortunately, despite a systematic reevaluation of these disease features in the era of TKI-based chemotherapy, none emerged as potential candidates to allow risk-stratification, particularly with respect to application of allogeneic SCT in first complete remission. With respect to prognosis, the nonmodifiable host features such as age continues to predominate, particularly owing to association with poorer tolerance of intensive chemotherapy and higher incidence for development of resistance via ABL tyrosine kinase domain mutations.^42,43 Molecular signatures as established by gene expression profiling may allow the identification of relevant prognostic features of Ph-positive ALL and potentially allow early intervention to avert development of resistance.

CONDENSED ABSTRACT.

The prognostic significance of immunophenotypic, karyotypic, and molecular genetic features of previously established relevance in de novo adult Philadelphia (Ph)-positive B-lymphoblastic leukemia were reassessed in the context of intensive tyrosine kinase inhibitor (TKI)-based chemotherapy, and compared with a contemporaneous cohort of adults with diploid karyotype B-lymphoblastic leukemia treated with similar chemotherapy without TKI. For Ph-positive ALL, CD25 expression was initially associated with an inferior survival. However, after adjusting for age, leukocyte count, and CD20 expression, it no longer retained prognostic significance. Age remains the dominating factor in predicting outcome for adults with Ph-positive ALL.