Patients aged ≥18 years diagnosed with ALL or Ph-pos biphenotypic acute leukemia (BAL) in the Asan Medical Center, Seoul, Korea, were included in this retrospective analysis. Patients were excluded from the analysis if they had at least one of the following features: diagnosis with L3 ALL (Burkitt leukemia) or chronic myeloid leukemia (CML) with lymphoid blast crisis, no results from CD19/CD20/CD22 cell surface marker analysis, or no treatment with vincristine, prednisone, and daunorubicin plus L-asparaginase (VPDL) [15] or modified VPDL-based chemotherapy [41617].

Patients were assigned to the high clinical risk group (CRG), according to the definition of the MRC UKALL XII/ECOG E2993 trial [6], if they met one or more of the following criteria: 1) age ≥35 years, 2) white blood cell (WBC) count at diagnosis ≥30×10⁶/L (for B-cell) and ≥100×10⁶/L (for T-cell), 3) presence of t(9;22) or BCR-ABL1 transcript, and 4) presence of t(4;11) or MLL-AF4 transcript. The institutional review board of Asan Medical Center approved this study (AMC-2015-0891).

Ethylenediamine tetraacetic acid (EDTA)-anticoagulated bone marrow (BM) aspirates were obtained from 230 patients at the time of diagnosis. IPM expression of an ALL marker panel (CD45, CD34, terminal deoxynucleotidyl transferase [TdT], CD19, CD10, CD20, cytoplasmic CD22, CD2, CD3, cytoplasmic CD3, CD5, CD7, CD13, CD33, myeloperoxidase, surface immunoglobulin M [IgM], and cytoplasmic IgM) was assessed within 24 hours of sample collection. BM aspirate (100 µL) was incubated with 8 µL of each fluorescence-conjugated monoclonal antibody for 20 minutes at room temperature. Nuclear (TdT) and cytoplasmic antigens (cytoplasmic CD22, CD3, IgM, and myeloperoxidase) were incubated with specific antibodies after a permeabilization process, using permeability reagents, and erythrocytes were lysed using lysing solution. Isotypic antibodies were used as negative controls in separate tubes. After washing with 2 mL of phosphate-buffered saline (PBS) and fixing in 500 µL of 1% phosphate-buffered saline (PBS)-paraformaldehyde, flow cytometry was performed using the FACSCanto II flow cytometry system (BD Biosciences, San Jose, CA, USA) and analyzed using FACSDiva software (BD Biosciences). Twenty thousand nucleated cells were acquired per tube, and leukemic blasts were isolated using CD45/side scatter gating. The expression level (%) of each IPM was estimated by comparison with the isotypic control, and positivity was defined as ≥20% leukemic cells staining positive for an IPM among total leukemic blasts. All antibodies were purchased from Becton Dickinson (BD Biosciences).

All patients received induction chemotherapy with VPDL (for Ph-negative ALL) or VPD plus BCR-ABL1 TKIs (for Ph-pos ALL). Those who were diagnosed and received induction chemotherapy after June 2005 were enrolled in the prospective study by the Adult ALL Working Party of the Korean Society of Hematology (June 2005 to May 2010) and were treated with a modified VPDL regimen, to evaluate the feasibility of daunorubicin dose escalation during induction for patients <65 years of age [16].

BCR-ABL1 TKIs were administered to patients with Ph-pos ALL from 2001, and patients enrolled in two prospective studies to evaluate the effectiveness of TKIs were included in this analysis. In the first study, conducted from 2005 until 2008, patients received imatinib (600 mg per day) plus multiagent chemotherapy [17], and in the second study (2009 until 2011), patients were treated with nilotinib (400 mg twice a day) plus multiagent chemotherapy [4]. BCR-ABL1 TKIs were changed to the second-line agent, including dasatinib, when patients experienced intolerance to or failure with the first-line agent.

After achieving HCR, patients either received a course of consolidation chemotherapy or, if it was feasible and a suitable donor was available, were considered for allo-HCT. Allo-HCT could be performed at any time after the achievement of HCR. Acceptable donors included matched siblings, fully or partially matched unrelated donors, and haploidentical familial (familial mismatched) donors. Patients undergoing allo-HCT received either myeloablative conditioning (busulfan plus cyclophosphamide or fludarabine; patients <55 years old) or reduced-intensity conditioning, consisting of a combination of busulfan, fludarabine, and antithymocyte globulin (patients ≥55 years old, or with comorbidities). Total body irradiation was not used for conditioning. Stem cells were collected either from BM or from granulocyte colony-stimulating factor-mobilized peripheral blood (PB).

Positivity for each IPM was defined as expression on ≥20% of blasts. Patients were dichotomized according to positivity for each IPM.

Characteristics of each group were summarized using descriptive statistics, and between-group differences were assessed by the t-test for normally distributed clinical characteristics, the Mann-Whitney U-test for non-normally distributed features, and Fisher's exact test for nominal variables.

Hematologic relapse (HREL) was defined as the presence of >5% lymphoblasts in the total mononuclear cells in BM aspirate, or the infiltration of lymphoblasts in extramedullary organs. Hematologic relapse-free survival (HRFS) time was measured from the date of HCR to the date of non-relapse mortality (NRM) or HREL. Overall survival (OS) was measured from the diagnosis of ALL to the date of last follow-up or death.

Differences in HCR rates between subgroups defined by positivity for IPMs were compared by logistic regression. Differences in HRFS and OS rates were compared using the Kaplan-Meier method and the log-rank test, and by Cox proportional hazards analysis. Variables with P values<0.1 in univariate analysis were included in multivariate Cox proportional hazards analysis. All statistical analyses were performed with IBM SPSS 21 (IBM Corp., Armonk, NY, USA).

Baseline characteristics and treatment outcomes of 230 patients who were diagnosed with ALL or Ph-pos BAL from 1994 until 2011 at Asan Medical Center, Seoul, Korea, and met the inclusion criteria are provided in Table 1. Twenty-eight percent of patients were diagnosed with Ph-pos ALL and 76% were classified into the high CRG.

All patients received VPDL- or modified VPDL-based chemotherapeutic cycles, and BCR-ABL1 TKIs were combined with cytotoxic agents for 37 (58%) of 64 patients with Ph-pos ALL. The BCR-ABL1 TKIs used were imatinib (N=23), nilotinib (N=13), and dasatinib (N=1).

The HCR rate for all patients was 92%, and the rate did not differ between patients with Ph-negative and Ph-pos ALL. A total of 124 patients (54%) underwent allo-HCT in first HCR (N=99), second HCR (N=9), and relapsed/refractory (N=16) status, respectively.

After a median of 45.4 months of follow-up of surviving patients (range, 1.4-203.0), the 2-year and 5-year HRFS rates were 48% and 36%, respectively, and the median HRFS time was 20.2 months (95% confidence interval [CI], 9.5-30.9) among patients who achieved HCR. The 2-year and 5-year OS rates of all patients included in this analysis were 57% and 39%, respectively, and the median OS time was 30.8 months (95% CI, 23.3-38.3).

Table 2 shows the expression rates of major ALL IPMs in samples from patients included in this study. The majority of patient samples were positive for CD19 and CD22 (87.0% and 63.9%, respectively). Among 226 and 220 evaluable patients, 45.6% and 80.5% were CD20- and TdT-positive, respectively.

None of the established prognostic factors or IPMs were associated with the HCR rate (Tables 3, 4). In contrast, positivity for CD20, CD13, CD34, and TdT was associated with the HRFS rate, and positivity for CD20 and CD13 was strongly associated with the OS rate (Table 4). When the outcomes were analyzed among patients who achieved HCR, 5-year HRFS (40% vs. 32%, P=0.002) and OS rates (46% vs. 31%, P=0.002) were significantly higher among those who received allo-HCT than among those who did not.

We assessed the prognostic value of IPMs by multivariate analysis including variables from the univariate analysis with P values<0.1. Positivity for CD20 and negativity for TdT were both significantly associated with HRFS and OS rates, as were CRG and performance of allo-HCT (Table 5).

We categorized patients according to CRG status and CD20 and TdT positivity, which were all significantly associated with outcome variables on multivariate analysis. Initially, patients were categorized into four subgroups according to CRG and CD20 positivity: group 1, standard CRG and CD20-negative; group 2, high CRG and CD20-negative; group 3, standard CRG and CD20-positive; and group 4, high CRG and CD20-positive. There were significant differences between the groups in median values for HRFS, which were "not reached", 23.1 months, 15.5 months, and 14.7 months for groups 1-4, respectively (P=0.001). Median values for OS also differed significantly at "not reached", 28.4 months, 33.9 months, and 16.6 months for subgroups 1-4, respectively (P=0.001; Fig. 1A, B).

Next, patients were categorized into three subgroups according to positivity for CD20 and TdT: group 1, CD20-negative and TdT-positive; group 2, CD20-negative and TdT-negative or CD20-positive and TdT-positive; and group 3, CD20-positive and TdT-negative. There were also significant differences between these groups in median HRFS (65.8 months, 16.8 months, and 6.0 months, respectively; P<0.001) and median OS (36.1 mo, 25.4 mo, and 9.0 mo, respectively; P<0.001) (Fig. 1C, D).

The prognostic significance of stratification by positivity for CD20 and TdT was evaluated and validated for each patient cohort (Table 6).

Stratification by positivity for CD20 and TdT demonstrated discriminative power in both groups based on age at diagnosis (<35 years and ≥35 years) and in high and standard CRG patients. Positivity for CD20 and TdT were also prognostic in Ph-negative and B-cell patient cohorts, whereas these markers were not significantly associated with prognosis in patients with Ph-pos and T-cell ALL.