ALBERT

Description: Abstract 659 Objects To evaluate the role of pre-emptive therapy with modified DLI on the prevention of relapse and improvement of survival after allo-HSCT. Methods The prospective study was performed to monitor MRD including WT1 and leukemia-associated aberrant immunophenotypes (LAIPs) in consecutive 814 standard-risk acute leukemia and MDS-RAEB patients after allo-HSCT, and then to initiate pre-emptive therapy with IL-2 (group B) or modified DLI (group C). Results The 3-year cumulative incidence of relapse, OS, DFS, and TRM were 18.1%, 66.0%, 61.6% and 19.7% for group A (without MRD); 68.0%, 23.9%, 20.8% and 11.2% for group B; as well as 29.8% (P=0.000), 55.4% (P=0.003), 52.5% (P=0.000), and 15.6% (P=0.208) for group C (Figure 1). In addition, the cumulative incidence of acute GVHD, grade 2 to 4 acute GVHD and grade 3 to 4 acute GVHD after MRD were 10.2%, 8.2% and 4.1% for group B; but 30.8% (P=0.017), 27.9% (P=0.017), and 8.4% (P=0.471) for group C. As well as, the 2-year cumulative incidence of chronic GVHD and chronic extensive GVHD after MRD were 37.3% and 30.5% for group B; as well as 42.9% (P=0.982) and 34.2% (P=0.858) for group C. Multivariate analysis identified pre-emptive therapy with modified DLI for the intervention of MRD (P=0.029, OR=2.016) as the only significant influence factor of superior survival. Furthermore, for the patients receiving modified DLI, multivariate analysis identified DLI-associated chronic GVHD (P=0.037, OR=6.158) and MRD conversion to negative within 3 months after DLI (P=0.042, OR=4.929) as the significant influence factors of lower relapse. Among the patients receiving modified DLI, the patients who had DLI-associated chronic GVHD and had MRD conversion to negative within 3 months after DLI, obtained the most superior survival (3-y relapse 0.0%, TRM 7.7%, OS 92.3%, and DFS 92.3%); however, the patients who had no DLI-associated chronic GVHD and had persistent MRD within 3 months after DLI, obtained the most inferior survival (3-y relapse 51.0% (P=0.045), TRM 20.4% (P=0.484), OS 28.6% (P=0.035), and DFS 27.4% (P=0.041)) (Figure 3). Moreover, apart from the patients who had MRD conversion to negative within 3 months after DLI and had DLI-associated chronic GVHD, repeated DLI possibly made other patients get better survival (P〉0.05) (Figure 4). Conclusions This study for the first time confirmed that as compared with the use of IL-2, pre-emptive with modified DLI was more an effective method to prevent relapse and probably a superior method to improve DFS in standard-risk acute leukemia and MDS-RAEB patients after HSCT. Furthermore, for the patients without DLI-associated chronic GVHD, repeated DLI should be performed in order to induce chronic GVHD, so as to possibly make the patients get better survival. In addition, this study confirmed again that MRD monitoring with WT1 and LAIPs was a reliable measure to detect relapse of acute leukemia after HSCT, because of its high specificity and sensitivity.Figure 1.The relapse, TRM, OS and DFS among the patients in all patients. (A) Cumulative incidence of relapse. (B) Cumulative incidence of transplant-related mortality (TRM). (C) Overall survival (OS). (D) Disease-free survival (DFS). The patients without MRD post-HSCT were assigned to group A (n=709); the patients with MRD post-HSCT and receiving IL-2 treatment were assigned to group B (n=49); the patients with MRD post-HSCT and receiving DLI were assigned to group C (n=56). 1* represents the P value of comparison between group B and C; 2* represents the P value of comparison between group A and C; and 3* represents the P value of comparison among three group.Figure 1. The relapse, TRM, OS and DFS among the patients in all patients. (A) Cumulative incidence of relapse. (B) Cumulative incidence of transplant-related mortality (TRM). (C) Overall survival (OS). (D) Disease-free survival (DFS). The patients without MRD post-HSCT were assigned to group A (n=709); the patients with MRD post-HSCT and receiving IL-2 treatment were assigned to group B (n=49); the patients with MRD post-HSCT and receiving DLI were assigned to group C (n=56). 1* represents the P value of comparison between group B and C; 2* represents the P value of comparison between group A and C; and 3* represents the P value of comparison among three group.Figure 2.The relapse, TRM, OS and DFS among the patients receiving DLI (n=56). (A) Cumulative incidence of relapse. (B) Cumulative incidence of transplant-related mortality (TRM). (C) Overall survival (OS). (D) Disease-free survival (DFS). The patients were divided into four groups according to the results of MRD and chronic GVHD after DLI.Figure 2. The relapse, TRM, OS and DFS among the patients receiving DLI (n=56). (A) Cumulative incidence of relapse. (B) Cumulative incidence of transplant-related mortality (TRM). (C) Overall survival (OS). (D) Disease-free survival (DFS). The patients were divided into four groups according to the results of MRD and chronic GVHD after DLI.Figure 3.The effects of repeated DLI on the disease-free survival (DFS). (A) Patients had MRD conversion to negative within 3 months after DLI and had no DLI-associated chronic GVHD (n=15). (B) Patients had persistent MRD within 3 months after DLI, but had no DLI-associated chronic GVHD (n=19). (C) Patients had persistent MRD within 3 months after DLI and had DLI-associated chronic GVHD (n=9). The patients were divided into two groups according to the numbers of DLI that patients received.Figure 3. The effects of repeated DLI on the disease-free survival (DFS). (A) Patients had MRD conversion to negative within 3 months after DLI and had no DLI-associated chronic GVHD (n=15). (B) Patients had persistent MRD within 3 months after DLI, but had no DLI-associated chronic GVHD (n=19). (C) Patients had persistent MRD within 3 months after DLI and had DLI-associated chronic GVHD (n=9). The patients were divided into two groups according to the numbers of DLI that patients received. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction Leukemia relapse is the most common cause of treatment failure after allogeneic hematopoietic stem cell transplantation (allo-HSCT). The biological mechanism of relapse is still not completely clear. Allo-HSCT can be considered as a kind of immunotherapy directed toward malignant hematologic disease. Confirmation the correlation between clinical relapse and specific subset of T cells and further elucidating the immune mechanism behind this phenomenon would be beneficial to explore new methods of adoptive immunotherapy. Accumulative evidence showed that regulatory T cells (Tregs) might be involved in immune escape mechanism and associated with the failure of host to trigger efficient immunological antitumor response. Recently, a kind of non-traditional CD4+ CD25–CD69+ T cells was found to be involved in disease progression in tumor-bearing mouse models and cancer patients. In this study, we attempted to define whether this subset of T cells was related to leukemia relapse after allo-HSCT and also demonstrate the potential immune regulatory mechanism behind this phenomenon. Methods Twenty-nine patients with malignant hematological disease treated with non-T-cell-depleted allo-HSCT at the Peking University Institute of Hematology from November 2009 to April 2011 including patients undergoing hematological relapse (n=22) and with detectable minimal residual disease (MRD, n=7) were selected as the initial subjects in this study. The other fifty-six patients who received allo-HSCT from October 2009 to July 2010 were also enrolled for prospective study. The bone marrow of the initial 29 subjects was collected at the time of hematological relapse or detecting MRD. The MRD status of those 56 patients was examined at regular time points: +30 day (d), +60d, +90d, +180d, +270d, +360d. Bone marrow samples from patients at above time points were collected for the MRD examination and counting of CD4+CD25-CD69+ cells by flow cytometry (FCM). Results The frequency of CD4+CD25-CD69+ T cells in healthy donors’ bone marrow was 2.79% (range, 2.11-4.94%). However, the frequency of this subset of T cell was significantly increased in patients with detectable MRD (7.60%, range, 4.53-9.14%, P=0.008), or undergoing hematological relapse (12.96%, range, 8.62-20.49%, P0.05). The incidence of MRD+ or relapse in high frequency of CD4+CD25-CD69+ T cells group (〉7%) was distinctly higher than that of low frequency of CD4+CD25-CD69+ T cells group at +60d, +90d and +270d after transplant. However, our preliminary data indicated that CD4+CD25-CD69+ T cells might not display immune regulatory function via cytokine secretion. Conclusions This study gave the first clinical evidence of correlation between non-traditional CD4+CD25-CD69+ T cells and leukemia relapse after allo-HSCT and would be beneficial to explore new methods of adoptive immunotherapy. Further research related to regulatory mechanism behind this phenomenon would be necessary. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 4695 Prolonged thrombocytopenia is a frequent complication after allogeneic hematopoietic stem cell transplantation (allo-HSCT), but its pathogenesis has remained obscure. In present study, by using a flow cytometry technique, we determined the frequency of bone marrow megakaryocytes (MKs) and MKs ploidy distributions in allo-HSCT recipients with or without prolonged thrombocytopenia (n=32 and 27, respectively) and in 13 healthy volunteers. Along with that, the expression of c-Mpl in MKs was measured. The result showed a significant difference in MKs ploidy distributions between the 3 groups. Both the allo-HSCT patients with or without prolonged thrombocytopenia showed a significant shift to low ploidy cells (left shift) with 55.63±18.62% and 44.63±19.38% of their MKs in ploidy classes 2N to 8N, respectively. In contrast, the healthy volunteers had only 25.97±10.70% of their MKs in ploidy classes 2N to 8N. And the patients with prolonged thrombocytopenia showed a marked increase in cells ≤8N, a significant decrease in the 16N and ≥32N cells comparing to patients without prolonged thombocytopenia, which told us there were more less mature MKs in the allo-HSCT recipients with prolonged thrombocytopenia. We concluded that prolonged thrombocytopenia and slow platelet engraftment after allo-HSCT may be related to reduced ploidy and immaturation of MKs. Disclosures: No relevant conflicts of interest to declare.

Description: To investigate the effects of non-inherited maternal antigen (NIMA) on clinical outcomes and immune recovery, especially of regulatory T cells (Tregs), in patients who underwent unmanipulated haploidentical transplantation.A retrospective cohort (n=57) and a prospective cohort (n=88) were included. Reconstitution of immune subsets, including Tregs, was determined using multicolor flow cytometry. In the retrospective cohort, the cumulative incidence of grades II-IV acute GVHD in patients with NIMA-mismatched donors was significantly lower than that of cases with NIPA-mismatched donors (14.8% vs. 43.30%, P=0.018). Patients with higher percentages of CD4+CD25+CD45RA+ T cells (naive Tregs) within CD4+ T cells recovered on day 30 (≥1.55%) experienced a significantly lower incidence of grades II-IV acute GVHD than that of cases with lower percentages of naive Tregs (

Description: Background: Allogeneic hematopoietic stem cell transplantation (allo-HSCT) has been established as an effective treatment for patients with hematological malignancies. Disease relapse remains a major cause of transplant failure.T cell homeostasis is critical to determine the potency of the GVT effect. Cytotoxic T lymphocyte antigen-4 (CTLA-4 or CD152) is a T cell activation negative regulator. Recent studies have shown the association of the CTLA-4 polymorphisms with the outcome after HLA-identical sibling allogeneic HSCT. Patients and Methods: In this study, we focused on four CTLA-4 polymorphisms, and analyzed the impact of donor genotypes and haplotypes on the conditions of 154 acute leukemia patients after related HLA-haplotype-mismatched transplantation. The four SNP genotypes (-1661, -318, CT60 and +49) were determined by TaqMan SNP genotyping assays. Results: Recipients of donors with +49 GG showed significantly lower OS (69.1% vs. 85.6%, P=0.024) and higher incidence of III-IV aGVHD (10.0% vs. 2.1%, P=0.032) than those with GA + AA(Fig.1,Fig.2). Multivariate analyses showed that +49GG was an independent risk factor for OS (HR:0.457,95%CI=0.227-0.920,P=0.028). Patients receiving mDLI showed significantly lower OS with +49 GG donor than those with AG+AA (P=0.011).The haplotype analysis revealed only three haplotypes in the donor population -1661/-318/CT60/+49 i.e.,ACGG,ACAA and GTGA,the frequencies were 64.3%, 19.5%, and 16.2%, respectively.Donors with and without the ACGG/ACGG haplotype had the same effect on transplant outcome as those with +49 GG and +49 AG+AA. Conclusion: The CTLA-4 +49 GG and the haplotype ACGG/ACGG reduced the overall survival and increased the aGVHD after allo-HSCT from the related HLA-haplotype-mismatched donor,knowledge of the CTLA-4 polymorphism and haplotype may provide useful information for donor selection and individual application of immunosuppressive agents and immunotherapy. CONFLICT OF INTEREST The authors declare no conflict of interest. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 3065 Introduction: Current studies suggest that monitoring the minimal residual disease (MRD) of patients with malignant hematological disease after allogeneic hematopoietic stem cell transplantation (allo-HSCT) is helpful for screening patients at high risk of relapse. WT1 and leukemia-associated aberrant immune phenotypes (LAIPs) were the major MRD parameters used to predict leukemia relapse after allo-HSCT. The aim of this study was to first evaluate the clinical value of various positive MRD standards for accurately predicting relapse based on WT1 and flow cytometry (FCM) data in a large sample of adult patients with actue leukemia (AL) in a single transplant center. Methods: In total, 824 AL patients treated with non-T-cell-depleted allo-HSCT at the Peking University Institute of Hematology from January 1, 2006 to November 30, 2011 were enrolled in this study. BM samples from patients were obtained at regular time points: +1 month, +2 months, +3 months, +4.5 months, +6 months, +9 months, +12 months and every 6 months thereafter for the MRD investigation after HSCT. WT1 expression was evaluated using TaqMan-based RQ-PCR technology and LAIPs were detected using 4/7-color FCM. At least one positive MRD marker, including WT1 or LAIP, was detected before transplantation in 90.8% (748 of 824) of the subjects. Positive FCM could be found in all 748 patients. Both FCM+ and WT1+ were detected before transplantation in 58.7% (484/824) of subjects. We compared the sensitivity and specificity of diverse, multiple-criteria MRD forecasting standards including one WT1+, two consecutive WT1+, one FCM+, two consecutive FCM+, concurrent WT1+ & FCM+, concurrent WT11.0+ & FCM+, MRDco+ and MRDco1.0+, based on WT1 and FCM assays. Results: Higher sensitivity was achieved without a loss of specificity in total 824 patients as well as in 748 patients with WT1+ or FCM+ before transplant when MRDco+ was used as the positive MRD standard (Table 1). Similar results were observed, even in 484 patients with acute myeloid leukemia (AML, n=308) and acute lymphoblastic leukemia (ALL, n=176) who had both abnormal WT1 and LAIPs values before transplant (Table 1). A multivariate analysis that included age, sex, disease status (CR1/CR2), donor type, number of chemotherapy courses before CR1, and MRD status (including the different MRD standards mentioned above) after transplantation showed that MRDco+ was an independent risk factor to predict leukemia relapse after allo-HSCT; it had the most powerful contribution to leukemia recurrence after transplant in both AML (HR=12.54, 95% CI: 7.01–22.42, P

Description: Abstract 2347 The role of allogeneic hematopoietic stem cell transplantation (allo-HSCT) in the imatinib era in patients with chronic myelogenous leukemia (CML) in accelerated phase (AP) has not been evaluated due to few comparison study published. A prospective study was designed to compare the outcome of imatinib versus allo-HSCT for AP CML according to the World Health Organization (WHO) 2001 classification in our center (Registration Number: ChiCTR-TNC-10000955). In imatinib group, imatinib were given at an initial dose of 400 mg or 600 mg daily. In allo-HSCT group, the recommended treatment prior to transplantation was a short-term imatinib therapy less than 3 months. From April 2001 to September 2008, 132 patients were enrolled, 87 in imatinib group and 45 in allo-HSCT group, respectively, according to their intention. In allo-HSCT group, 19 patients performed transplant from a HLA-matched sibling donor, 23 from a HLA-mismatched sibling/haploidentical donor, and 3 from a HLA-mismatched unrelated donor. The end time of evaluation was April 2010. After a median follow-up of 45 months (range, 7–108 months) for 53 living patients on imatinib and 65 months (range, 18–108 months) for 38 living patients post allo-HSCT, imatinib was inferior to allo-HSCT in outcome, with the estimated 6-year event-free survival (EFS), overall survival (OS) and progression-free survival (PFS) rates of 39.2% vs 71.7% (P=0.035), 51.4% vs 83.3% (P=0.023), and 48.3% vs 95.2% (P=0.000), respectively. A multivariate analysis of the total population of 132 patients adding pretreatment characteristics and therapy (imatinib versus allo-HSCT) indicated that longer CML disease duration, pretreatment anemia and higher percentage of peripheral blasts were independent adverse prognostic factors for OS and PFS in common, imatinib therapy was only associated with shorter PFS. In order to analyze whether therapy play an important role in survival differences among patients with or without common pretreatment poor prognostic factors for OS and PFS, we categorized the entire cohort into low-risk (neither factor, n=40), intermediate-risk (either factor, n=59) or poor-risk (at least two factors, n=33). Therapy had no influence on the outcome in low-risk patients, with the estimated 6-year EFS, OS and PFS rates of 80.9% vs 80.7% (P=0.898), 100% vs 81.2% (P=0.114), and 85.2% vs 95.2% (P=0.365), in imatinib group vs allo-HSCT group, respectively. EFS and OS showed no difference by therapy in intermediate-risk patients, with the estimated 6-year EFS and OS rates of 47.1% vs 61.9% (P=0.788), and 61.3% vs 81.3% (P=0.773), in imatinib group vs allo-HSCT group, respectively. However more patients developed a relapse in advanced phase with imatinib compared to those with allo-HSCT, the estimated 6-year PFS rates were 55.7% vs 92.9% (P=0.047), respectively. The superiority of allo-HSCT was extremely significant in poor-risk patients, with the estimated 5-year EFS, OS and PFS rates of 9.3% vs 66.7% (P=0.034), 17.7% vs 100% (P=0.008) and 18.8% vs 100% (P=0.006), in imatinib group vs allo-HSCT group, respectively. We conclude that allo-HSCT is the first-line option for all patients with CML in AP; it is superior to imatinib with evident survival advantage for poor/intermediate-risk patients. However, the outcome of imatinib and allo-HSCT were equally good in low-risk patients with CML in AP. For such patients, it may also be advised that imatinib remains the primary option by carefully monitoring of MRD, and allo-HSCT may be considered if there is evidence of imatinib resistance. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 1397 Background: Using neonatal NOD/SCID/IL2rγnull xenotransplantation model, we previously demonstrated that CD34+CD38+CD19+ cells as well as CD34+CD38−CD19+ cells have the capacities to initiate B-precursor ALL (B-ALL) in vivo and to self-renew, that is, leukemia initiating cells(LICs) are enriched in the CD34+CD19+phenotype in human B-ALL (Kong Y et al. Leukemia 2008; 22: 1207–1213). Nevertheless, in order to distinguish B-ALL initiating cells from their normal compartment, further markers have to be identified. CD58 (lymphocyte function–associated antigen 3), a cell surface adhesion molecule binding to CD2, plays a critical role in the attachment of cytotoxic T lymphocytes and non-specific killer cells to their targets. Recent data give evidence that diffuse large B cell lymphoma patients with poor expression of CD58 appear to be more aggressive by escaping from immune-surveillance of the host. However, the prognostic relevance of CD58 expression on B-ALL initiating cells is largely unknown. Objectives: To investigate the expression profile of CD58 on CD34+CD19+B-ALL initiating cells. To evaluate the prognostic significance of CD58 expression on LICs in human B-precursor ALL. Materials and methods: Using a cohort of 139 patients (including pediatric and adult patients) with CD34+ B-ALL, the expression profile of CD58 on LICs(CD34+CD19+ phenotype) were examined by multicolor flow cytometry (FCM) at diagnosis. A total of 1,050,000 events were routinely collected. More than 20% of LICs with CD58 expression were defined as CD58-positive LICs (CD34+CD19+CD58+ phenotype, abbreviated by CD58+LICs), all other cases were defined as CD58-negtive LICs (CD34+CD19+CD58−phenotype, abbreviated by CD58−LICs). Furthermore, the impact of CD58−LICs at diagnosis on the clinical outcome was prospectively investigated. The study was approved by the Ethics Committee of Peking University People's Hospital and written informed consent was obtained from all subjects. Results: Among the newly diagnosed B-ALL patients, 119 cases were detected with CD58+LICs and the remaining 20 cases with CD58−LICs. The expression of CD58 on LICs at diagnosis was inversely related to the age, WBC at diagnosis and NCCN risk group of the B-ALL patients (P=0.002, P=0.015 and P=0.009, respectively). CD58+LICs group had a higher complete remission(CR) rate after one course induction than CD58−LICs group (87.39% vs. 65.00%, P=0.019).Cumulative incidence of relapse (CIR) at 2-year in the CD58+LICs group was significantly lower than that in the CD58−LICs group (18.20% ± 0.15% vs. 72.00% ± 1.30%, P

Description: Rosiglitazone is a peroxisome proliferator activated receptor-γ(PPAR-γ) agonist available to improve glucose metabolism in patients with Type 2 diabetes. Recent evidence suggests that therapeutic use of rosiglitazone have caused unwanted hematological side effects, such as anemia, leukopenia, thrombocytopenia, even pancytopenia. However, others reported that pretreatment with rosiglitazone for 5 days could protect against 5-Fu-induced myelotoxity which is FLT3 dependent. Thus, it is still unclear the exact effects of rosiglitazone treatment on homeostatic and stress hematopoiesis. Rosiglitazone is a potent stimulator of adipogenesis, as it can induce adipogenic differentiation both in vitro and in vivo. Recently, it was reported that adipocyte-rich bone marrow(BM) harbored a decreased frequency of progenitor cells. Furthermore, our laboratory and those of other investigators have demonstrated that BADGE, an inhibitor of PPAR-γ signaling pathway, can improved the hematopoietic recovery in response to stress. So we hypothesize that long-term treatment of rosiglitazone may inhibit the hematopoiesis by inducing BM adipogenesis. In this study, we first treated C57BL/6J mice with 37.5mg/kg rosiglitazone for six weeks to examine the long-term effect of rosiglitazone on hematopoiesis in vivo. We found that rosiglitazone treatment does not alter body weight or BM function of normal healthy mice. In response to 5-Fu induced hematopoietic stress, rosiglitazone group mice exhibited delayed recovery of WBC and BMMNC. Furthermore, the frequency and absolute number of GMP and CLP cells in rosiglitazone group was decreased at d14 after 5-Fu. Concomitantly, we found that CFU-GM counts of rosiglitazone group BM were significantly decreased when compared with control group mice, while CFU-GEMM counts revealed no difference. However, we didn't observed any alteration in absolute number of LSK cells between the two groups, although rosiglitazone group mice had increased percentage of LSK. These results indicated that rosiglitazone could maintain the HSC compartment while impair myeloid differentiation from myelosuppressive stress, which may explain the delayed hematopoietic recovery in rosiglitazone group. To answer the question whether the effects of rosiglitazone were directly mediated, in vitro cultures were used. We cultured d3 5-Fu BM cells (BM cells harvested 3 day after 5-FU administration) in the presence of rosiglitazone to detect whether rosiglitazone inhibited hematopoietic progenitor cells in stress condition. Cell proliferation of Lin- cell was unaffected at pharmacologically achievable concentrations of rosiglitazone. However, rosiglitazone at high concentrations (〉10μM) significantly inhibit cell proliferation of Lin- cell, which was in agreement with previous reports. To confirm our results, we choose two other TZD drugs to test whether cell proliferation was affected. Similarly, neither troglitazone nor pioglitazone exhibited an inhibition effect on cell proliferation at pharmacologically achievable concentrations. Furthermore, rosiglitazone treatment didn't decrease the numbers of CFU-GM. These data show that rosiglitazone can not directly inhibit cell proliferation or myeloid differentiation of HPC in stress condition. As long-term treatment of rosiglitazone caused a significant increase of BM adipocytes in mice, we speculated that impaired myeloid differentiation in rosiglitazone group mice may be due to enhanced adipogenesis. We found that rosiglitazone promoted the adipogenic differentiation of two stromal cell lines (C3H10T1/2,M2-10B4). In addition, rosiglitazone treated stromal cell lines significantly inhibited cell proliferation and myeloid differentiation of co-cultured Lin- cells. In conclusion, rosiglitazone induced BM adipogenesis can impair the myeloid differentiation of hematopoietic stem cell, which may contribute to a significant delay in hematopoietic recovery after chemotherapeutic stress. Further investigation of the molecular mechanism is warranted and may enable the development of therapies effective for suppressed hematopoiesis induced by TZDs. Disclosures No relevant conflicts of interest to declare.

Description: Previous studies have suggested that IKZF1 gene deletions are associated with an adverse prognosis in acute lymphoblastic leukemia (ALL). However, the prognostic value of IKZF1 deletions in adult common B-cell ALL (Com-B ALL) has not been well-defined, especially in the context of different post-remission therapies. The objectives of this study were to evaluate the prognostic role of IKZF1 deletions in adult Com-B ALL. Overall, 162 untreated adult Com-B ALL patients were recruited between April 2006 and April 2013. Patients received one or two cycles of induction therapy and, if in remission, were allowed to select either ongoing systemic chemotherapy or hematopoietic stem cell transplantation (HSCT). BCR-ABL positive patients were additionally treated with tyrosine kinase inhibitors (TKIs). Deletions in the IKZF1 gene were detected using multiplex RQ-PCR, multiplex fluorescent PCR, multiplex ligation-dependent probe amplification (MLPA) and sequence analysis. High-risk (HR) was defined as having any of the following factors: central nervous system involvement at diagnosis; age ¡Ý50 years; WBC count ¡Ý30x109/L; hypodiploidy; mixed lineage leukemia (MLL) gene rearrangements; BCR-ABL rearrangements; t(1;19) translocations; or a complicated karyotype (¡Ý5 chromosomal abnormalities); failure to achieve remission after two cycles of induction therapy. The study end points included relapse-free survival (RFS), disease-free survival (DFS) and overall survival (OS), which was analyzed using the Kaplan-Meier method. The cumulative incidence of relapse (CIR) was estimated, adjusting for competing risks of death, and compared by Gray's test. IKZF1 deletions were detected in 79 (48.8%) out of 162 adult Com-B ALL patients. Patients with an IKZF1 deletion (n=79) had a significantly inferior prognosis than those with wild-type IKZF1 (n=83) (Fig.1). The prognosis of patients with IKZF1 deletions with or without the BCR-ABL rearrangements was significantly inferior compared to those of patients with neither IKZF1 deletions nor BCR-ABL rearrangements. The power of IKZF1 deletion status as a prognostic factor remained in both non-HR group (5-yr OS: 52.5%¡À13.1% for IKZF1 deletions (n=25) vs. 72.2%¡À8.8% for wild-type IKZF1 (n=42), P=0.048) and HR group (5-yr OS: 36.7%¡À8.2% for IKZF1 deletions (n=54) vs. 69.3%¡À8.3% for wild-type IKZF1 (n=41), P=0.015). Among patients who received chemotherapy as post-remission therapy (n=52), IKZF1 deletions were associated with an unfavorable OS (P=0.003). In this group, the median OS time for IKZF1-deletion patients (n=28) was 10.0¡À1.0 months, while the median OS time for IKZF1-wild-type carriers (n=24) was not achieved. Among those that received HSCT as post-remission therapy (n=94), IKZF1 status did not affect the prognosis. For patients with IKZF1 deletions, HSCT (n=40) was significantly superior to chemotherapy (n=28) as a post-remission therapy. In the wild-type IKZF1 HR group, patients that received HSCT (n=24) achieved a longer DFS and RFS while the OS did not significantly differ from that of patients who received chemotherapy alone (n=12). In the wild-type IKZF1 non-HR group, there was no significant difference between patients that received HSCT (n=30) and those who received chemotherapy (n=12) as a post-remission therapy. Comparing survival curves of HSCT to chemotherapy are shown in Fig.2. Multivariate analysis confirmed the negative impact of IKZF1 deletions on OS (RR 2.3; 95% CI, 1.3-4.3; P=0.007), DFS (RR 1.8; 95% CI, 1.1-3.0; P=0.029), RFS (RR 2.1; 95% CI, 1.3-3.9; P=0.018) and favorable impact of HSCT on OS (RR 0.2; 95% CI, 1.0-0.5; P=0.000), DFS (RR 0.2; 95% CI, 0.1-0.4; P=0.000), RFS (RR 0.2; 95% CI, 0.1-0.4; P=0.000). These results indicated that IKZF1 status served as an independent negative prognostic marker in adult Com-B ALL and surpassed conventional HR features in the era of TKIs. Patients may benefit from therapy stratified by IKZF1 deletions and conventional HR features. Fig. 1 Fig. 1. Survival curves related to status of IKZF1 in Com-B-ALL cohort. Abbreviation: IKZF1 WT, IKZF1 wide type; IKZF1 DEL, IKZF1 deletion. Fig. 2 Fig. 2. Comparing survival curves of HSCT to chemotherapy in IKZF1-deleted group, wild-type IKZF1 HR group and wild-type IKZF1 non-HR group. Disclosures No relevant conflicts of interest to declare.