Search

To expand the current knowledge about azacitidine (Aza) and donor lymphocyte infusions (DLI) as salvage therapy for relapse after allogeneic hematopoietic stem cell transplantation (allo-HSCT) and to identify predictors for response and survival, we retrospectively analyzed data of 154 patients with acute myeloid leukemia (AML, n = 124), myelodysplastic (MDS, n = 28), or myeloproliferative syndrome (n = 2). All patients received a median number of 4 courses of Aza (range, 4 to 14) and DLI were administered to 105 patients (68%; median number of DLI, 2; range, 1 to 7). Complete and partial remission rates were 27% and 6%, respectively, resulting in an overall response rate of 33%. Multivariate analysis identified molecular-only relapse (hazard ratio [HR], 9.4; 95% confidence interval [CI], 2.0 to 43.5; P = .004) and diagnosis of MDS (HR, 4.1; 95% CI, 1.4 to 12.2; P = .011) as predictors for complete remission. Overall survival (OS) at 2 years was 29% ± 4%. Molecular-only relapse (HR, .14; 95% CI, .03 to .59; P = .007), diagnosis of MDS (HR, .33; 95% CI, .16 to .67; P = .002), and bone marrow blasts <13% (HR, .54; 95% CI, .32 to .91; P = .021) were associated with better OS. Accordingly, 2-year OS rate was higher in MDS patients (66% ± 10%, P = .001) and correlated with disease burden in patients with AML. In summary, Aza and DLI is an effective and well-tolerated treatment option for patients with relapse after allo-HSCT, in particular those with MDS or AML and low disease burden. The latter finding emphasizes the importance of stringent disease monitoring and early intervention.

The purpose of this analysis was to provide 6-year follow-up of the CLL3X trial, which studied reduced-intensity allogeneic hematopoietic stem cell transplantation (HSCT) in patients with poor-risk chronic lymphocytic leukemia (CLL), and to investigate the effect of TP53, SF3B1, and NOTCH1mutations on HSCT outcome. For 90 allografted patients, 6-year overall survival (OS) was 58% and 6-year event-free survival (EFS) was 38%. TP53, SF3B1, and NOTCH1mutations were found in 30%, 26%, and 14% of the trial population, respectively. By univariate and multivariate analyses, the mutational status of the TP53, SF3B1, and NOTCH1genes had no significant effect on OS and EFS. Studies of minimal residual disease confirmed durability of CLL eradication in mutated patients. We conclude that HSCT can provide long-term disease control in patients with poor-risk CLL independent of the presence of TP53, SF3B1, and NOTCH1mutations. The trial has been registered at the US National Cancer Institute as #EU-20554, NCT00281983.

The purpose of this analysis was to provide 6-year follow-up of the CLL3X trial, which studied reduced-intensity allogeneic hematopoietic stem cell transplantation (HSCT) in patients with poor-risk chronic lymphocytic leukemia (CLL), and to investigate the effect of TP53, SF3B1, and NOTCH1 mutations on HSCT outcome. For 90 allografted patients, 6-year overall survival (OS) was 58% and 6-year event-free survival (EFS) was 38%. TP53, SF3B1, and NOTCH1 mutations were found in 30%, 26%, and 14% of the trial population, respectively. By univariate and multivariate analyses, the mutational status of the TP53, SF3B1, and NOTCH1 genes had no significant effect on OS and EFS. Studies of minimal residual disease confirmed durability of CLL eradication in mutated patients. We conclude that HSCT can provide long-term disease control in patients with poor-risk CLL independent of the presence of TP53, SF3B1, and NOTCH1 mutations. The trial has been registered at the US National Cancer Institute as #EU-20554, NCT00281983.

The purpose of this prospective multicenter phase 2 trial was to investigate the long-term outcome of reduced-intensity conditioning allogeneic stem cell transplantation (alloSCT) in patients with poor-risk chronic lymphocytic leukemia. Conditioning was fludarabine/ cyclophosphamide-based. Longitudinal quantitative monitoring of minimal residual disease (MRD) was performed centrally by MRD-flow or real-time quantitative polymerase chain reaction. One hundred eligible patients were enrolled, and 90 patients proceeded to alloSCT. With a median follow-up of 46 months (7-102 months), 4-year nonrelapse mortality, event-free survival (EFS) and overall survival (OS) were 23%, 42%, and 65%, respectively. Of 52 patients with MRD monitoring available, 27 (52%) were alive and MRD negative at 12 months after transplant. Four-year EFS of this subset was 89% with all event-free patients except for 2 being MRD negative at the most recent assessment. EFS was similar for all genetic subsets, including 17p deletion (17p−). In multivariate analyses, uncontrolled disease at alloSCT and in vivo T-cell depletion with alemtuzumab, but not 17p−, previous purine analogue refractoriness, or donor source (human leukocyte antigen-identical siblings or unrelated donors) had an adverse impact on EFS and OS. In conclusion, alloSCT for poor-risk chronic lymphocytic leukemia can result in long-term MRD-negative survival in up to one-half of the patients independent of the underlying genomic risk profile. This trial is registered at http://clinicaltrials.gov as NCT00281983.

The purpose of this prospective multicenter phase 2 trial was to investigate the long-term outcome of reduced-intensity conditioning allogeneic stem cell transplantation (alloSCT) in patients with poor-risk chronic lymphocytic leukemia. Conditioning was fludarabine/ cyclophosphamide-based. Longitudinal quantitative monitoring of minimal residual disease (MRD) was performed centrally by MRD-flow or real-time quantitative polymerase chain reaction. One hundred eligible patients were enrolled, and 90 patients proceeded to alloSCT. With a median follow-up of 46 months (7-102 months), 4-year nonrelapse mortality, event-free survival (EFS) and overall survival (OS) were 23%, 42%, and 65%, respectively. Of 52 patients with MRD monitoring available, 27 (52%) were alive and MRD negative at 12 months after transplant. Four-year EFS of this subset was 89% with all event-free patients except for 2 being MRD negative at the most recent assessment. EFS was similar for all genetic subsets, including 17p deletion (17p−). In multivariate analyses, uncontrolled disease at alloSCT and in vivo T-cell depletion with alemtuzumab, but not 17p−, previous purine analogue refractoriness, or donor source (human leukocyte antigen-identical siblings or unrelated donors) had an adverse impact on EFS and OS. In conclusion, alloSCT for poor-risk chronic lymphocytic leukemia can result in long-term MRD-negative survival in up to one-half of the patients independent of the underlying genomic risk profile. This trial is registered at http://clinicaltrials.govas NCT00281983.

Forty patients were enrolled in this phase 2 study combining radioimmunotherapy (RIT) using yttrium-90-ibritumomab-tiuxetan (15 MBq [0.4 mCi]/kg) with reduced-intensity conditioning (RIC) using fludarabine (90 mg/m2) and 2 Gy total body irradiation followed by allogeneic hematopoietic cell transplantation (HCT) from related (n = 13) or unrelated (n = 27) donors for the treatment of advanced non-Hodgkin lymphoma. Diagnoses were follicular lymphoma (n = 17), chronic lymphocytic leukemia (n = 13), mantle cell lymphoma (n = 8), marginal zone lymphoma (n = 1), and lymphoplasmacytic lymphoma (n = 1). Median age was 55 years (range, 34-68 years). All patients were high risk with refractory disease or relapse after preceding autologous HCT. No additional toxicities attributable to RIT were observed. Engraftment was rapid and sustained. Incidences of acute graft-versus-host disease 2-4 and chronic graft-versus-host disease were 43% and 53%, respectively. Kaplan-Meier–estimated nonrelapse mortality was 45% at 2 years. Twenty-two of 40 patients (55%) are alive, resulting in a Kaplan-Meier–estimated 2-year survival of 51% for all, 67% for follicular lymphoma, 49% for chronic lymphocytic leukemia, and 37% for mantle cell lymphoma patients. The combined use of RIT with RIC is feasible with acceptable toxicity, even in elderly and heavily pretreated patients. This study is registered at www.clinicaltrials.govas #NCT00302757.

Forty patients were enrolled in this phase 2 study combining radioimmunotherapy (RIT) using yttrium-90-ibritumomab-tiuxetan (15 MBq [0.4 mCi]/kg) with reduced-intensity conditioning (RIC) using fludarabine (90 mg/m2) and 2 Gy total body irradiation followed by allogeneic hematopoietic cell transplantation (HCT) from related (n = 13) or unrelated (n = 27) donors for the treatment of advanced non-Hodgkin lymphoma. Diagnoses were follicular lymphoma (n = 17), chronic lymphocytic leukemia (n = 13), mantle cell lymphoma (n = 8), marginal zone lymphoma (n = 1), and lymphoplasmacytic lymphoma (n = 1). Median age was 55 years (range, 34-68 years). All patients were high risk with refractory disease or relapse after preceding autologous HCT. No additional toxicities attributable to RIT were observed. Engraftment was rapid and sustained. Incidences of acute graft-versus-host disease 2-4 and chronic graft-versus-host disease were 43% and 53%, respectively. Kaplan-Meier–estimated nonrelapse mortality was 45% at 2 years. Twenty-two of 40 patients (55%) are alive, resulting in a Kaplan-Meier–estimated 2-year survival of 51% for all, 67% for follicular lymphoma, 49% for chronic lymphocytic leukemia, and 37% for mantle cell lymphoma patients. The combined use of RIT with RIC is feasible with acceptable toxicity, even in elderly and heavily pretreated patients. This study is registered at www.clinicaltrials.gov as #NCT00302757.

Dendritic cells (DC) are highly specialized antigen-presenting cells with a unique ability to activate resting T lymphocytes and initiate primary immune responses. Angiotensin II (AII) is involved in key events of the inflammatory response. Because our previous work implicated an effect of AII on differentiation and function of murine and human DC, we investigated the impact of AII type 1 receptor (AT(1)) deficiency on the phenotypical and functional properties of mouse DC in vitro and in vivo. Bone marrow (BM) cells isolated from mice lacking AII subtype 1a receptor (AT(1a)), AII subtype 1b receptor (AT(1b)), or both receptor isoforms and control littermates [wild type (WT)] were cultured for 7 days in the presence of recombinant mouse granulocyte/macrophage colony-stimulating factor to generate myeloid DC in vitro. Generation of CD11c(+) cells was less efficient in both AT(1a)- and AT(1b)-deficient BM cells than in WT BM cell cultures. Moreover, DC generated from AT(1)-deficient progenitors showed lower levels of expression of major histocompatibility complex II (MHC-II) and CD11c (p < 0.01) and a marked reduction in their allostimulatory activity (p < 0.01 or 0.001). Although AT(1)-deficient DC released comparable levels of interleukin (IL)-10 and IL-12p70 to WT DC, they produced significantly lower levels of tumor necrosis factor alpha (TNF-alpha) (p < 0.05). Remarkably, CD11c(+) cells isolated from the spleen of AT(1) knockout mice challenged with lipopolysaccharide in vivo up-regulated MHC-II, CD40, and CD80 as did WT, but released significantly lower levels of TNF-alpha (p < 0.01). These data provide clear evidence that AT(1) controls differentiation and functionality of DC and thus may have a crucial impact on inflammatory processes where local angiotensinergic systems are known to be activated.

From 2002 to 2007, 103 patients with primary myelofibrosis or postessential thrombocythemia and polycythemia vera myelofibrosis and a median age of 55 years (range, 32-68 years) were included in a prospective multicenter phase 2 trial to determine efficacy of a busulfan (10 mg/kg)/fludarabine (180 mg/m2)–based reduced-intensity conditioning regimen followed by allogeneic stem cell transplantation from related (n = 33) or unrelated donors (n = 70). All but 2 patients (2%) showed leukocyte and platelet engraftment after a median of 18 and 22 days, respectively. Acute graft-versus-host disease grade 2 to 4 occurred in 27% and chronic graft-versus-host disease in 43% of the patients. Cumulative incidence of nonrelapse mortality at 1 year was 16% (95% confidence interval, 9%-23%) and significantly lower for patients with a completely matched donor (12% vs 38%; P = .003). The cumulative incidence of relapse at 3 years was 22% (95% confidence interval, 13%-31%) and was influenced by Lille risk profile (low, 14%; intermediate, 22%; and high, 34%; P = .02). The estimated 5-year event-free and overall survival was 51% and 67%, respectively. In a multivariate analysis, age older than 55 years (hazard ratio = 2.70; P = .02) and human leukocyte antigen–mismatched donor (hazard ratio = 3.04; P = .006) remained significant factors for survival. The study was registered at www.clinicaltrials.gov as #NCT 00599547.

From 2002 to 2007, 103 patients with primary myelofibrosis or postessential thrombocythemia and polycythemia vera myelofibrosis and a median age of 55 years (range, 32-68 years) were included in a prospective multicenter phase 2 trial to determine efficacy of a busulfan (10 mg/kg)/fludarabine (180 mg/m2)–based reduced-intensity conditioning regimen followed by allogeneic stem cell transplantation from related (n = 33) or unrelated donors (n = 70). All but 2 patients (2%) showed leukocyte and platelet engraftment after a median of 18 and 22 days, respectively. Acute graft-versus-host disease grade 2 to 4 occurred in 27% and chronic graft-versus-host disease in 43% of the patients. Cumulative incidence of nonrelapse mortality at 1 year was 16% (95% confidence interval, 9%-23%) and significantly lower for patients with a completely matched donor (12% vs 38%; P= .003). The cumulative incidence of relapse at 3 years was 22% (95% confidence interval, 13%-31%) and was influenced by Lille risk profile (low, 14%; intermediate, 22%; and high, 34%; P= .02). The estimated 5-year event-free and overall survival was 51% and 67%, respectively. In a multivariate analysis, age older than 55 years (hazard ratio = 2.70; P= .02) and human leukocyte antigen–mismatched donor (hazard ratio = 3.04; P= .006) remained significant factors for survival. The study was registered at www.clinicaltrials.govas #NCT 00599547.

This study investigated the immunogenicity of Wilms tumor gene product 1 (WT1)–peptide vaccination in WT1-expressing acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) patients without curative treatment option. Vaccination consisted of granulocyte-macrophage colony-stimulating factor subcutaneously days 1 to 4, and WT1.126-134 peptide and 1 mg keyhole limpet hemocyanin on day 3. The initial 9 patients received 4 vaccinations biweekly, then monthly, and the subsequent 10 patients received continual biweekly vaccination. Seventeen AML patients and 2 refractory anemia with excess blasts patients received a median of 11 vaccinations. Treatment was well tolerated. Objective responses in AML patients were 10 stable diseases (SDs) including 4 SDs with more than 50% blast reduction and 2 with hematologic improvement. An additional 4 patients had clinical benefit after initial progression, including 1 complete remission and 3 SDs. WT1 mRNA levels decreased at least 3-fold from baseline in 35% of patients. In 8 of 18 patients, WT1-tetramer+ T cells increased in blood and in 8 of 17 patients in bone marrow, with a median frequency in bone marrow of 0.18% at baseline and 0.41% in week 18. This WT1 vaccination study provides immunologic, molecular, and preliminary evidence of potential clinical efficacy in AML patients, warranting further investigations.

This study investigated the immunogenicity of Wilms tumor gene product 1 (WT1)–peptide vaccination in WT1-expressing acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) patients without curative treatment option. Vaccination consisted of granulocyte-macrophage colony-stimulating factor subcutaneously days 1 to 4, and WT1.126-134 peptide and 1 mg keyhole limpet hemocyanin on day 3. The initial 9 patients received 4 vaccinations biweekly, then monthly, and the subsequent 10 patients received continual biweekly vaccination. Seventeen AML patients and 2 refractory anemia with excess blasts patients received a median of 11 vaccinations. Treatment was well tolerated. Objective responses in AML patients were 10 stable diseases (SDs) including 4 SDs with more than 50% blast reduction and 2 with hematologic improvement. An additional 4 patients had clinical benefit after initial progression, including 1 complete remission and 3 SDs. WT1 mRNA levels decreased at least 3-fold from baseline in 35% of patients. In 8 of 18 patients, WT1-tetramer+T cells increased in blood and in 8 of 17 patients in bone marrow, with a median frequency in bone marrow of 0.18% at baseline and 0.41% in week 18. This WT1 vaccination study provides immunologic, molecular, and preliminary evidence of potential clinical efficacy in AML patients, warranting further investigations.

Allogeneic hematopoietic stem cell transplantation (SCT) regimens incorporating the lymphocytotoxic CD52 antibody alemtuzumab demonstrate efficient engraftment and reduced graft-versus-host disease (GVHD). However, these protocols substantially impair posttransplantation antiviral and antitumor immunity. To accelerate immune reconstitution after alemtuzumab-based reduced-intensity SCT, we administered prophylactic CD8-depleted donor lymphocyte infusions (DLIs) starting on days 60 and 120 after transplantation. DLIs were processed in an immunomagnetic good manufacturing practice depletion procedure resulting in a 2.5- to 6-log reduction in CD8 T cells. Of 23 high-risk patients with hematologic malignancies, 11 received a total of 21 CD8-depleted DLIs. Five patients developed transient grade I acute GVHD following transfer. Only 2 patients with HLA-C–mismatched donors showed grade II and III acute GVHD and subsequently progressed to limited chronic GVHD. Following DLIs, 4 patients with declining hematopoietic donor chimerism converted to full chimeras. A 2.1-fold median increase of circulating CD4 T cells was observed within 2 weeks after infusion. Non-DLI patients did not show a comparable rise in CD4 counts. Four patients demonstrated enhanced frequencies of cytomegalovirus-specific CD4 and CD8 T cells following transfer. Our results suggest that prophylactic CD8-depleted DLIs accelerate immune reconstitution after lymphodepleted HLA-matched SCT and carry a low risk of inducing severe GVHD.

Allogeneic hematopoietic stem cell transplantation (SCT) regimens incorporating the lymphocytotoxic CD52 antibody alemtuzumab demonstrate efficient engraftment and reduced graft-versus-host disease (GVHD). However, these protocols substantially impair posttransplantation antiviral and antitumor immunity. To accelerate immune reconstitution after alemtuzumab-based reduced-intensity SCT, we administered prophylactic CD8-depleted donor lymphocyte infusions (DLIs) starting on days 60 and 120 after transplantation. DLIs were processed in an immunomagnetic good manufacturing practice depletion procedure resulting in a 2.5- to 6-log reduction in CD8 T cells. Of 23 high-risk patients with hematologic malignancies, 11 received a total of 21 CD8-depleted DLIs. Five patients developed transient grade I acute GVHD following transfer. Only 2 patients with HLA-C–mismatched donors showed grade II and III acute GVHD and subsequently progressed to limited chronic GVHD. Following DLIs, 4 patients with declining hematopoietic donor chimerism converted to full chimeras. A 2.1-fold median increase of circulating CD4 T cells was observed within 2 weeks after infusion. Non-DLI patients did not show a comparable rise in CD4 counts. Four patients demonstrated enhanced frequencies of cytomegalovirus-specific CD4 and CD8 T cells following transfer. Our results suggest that prophylactic CD8-depleted DLIs accelerate immune reconstitution after lymphodepleted HLA-matched SCT and carry a low risk of inducing severe GVHD.

There is considerable interest in immunotherapeutic approaches for lymphoma. The expression of proteinase inhibitor 9 (PI-9), a molecule that inactivates granzyme B, is considered an immune escape mechanism in lymphoma. Further, lymphomas frequently overexpress the antiapoptotic molecule bcl-2, which is able to inhibit perforin-dependent cytotoxic pathways. In this study, the impact of PI-9 and bcl-2 expression on the sensitivity of lymphomas to T- and natural killer (NK) cell–mediated cytotoxicity was analyzed. We found PI-9 expression in 10 of 18 lymphoma cell lines and in 9 of 14 primary lymphomas. Overexpression of bcl-2 was found in 8 of 18 cell lines and in 12 of 14 primary lymphomas. All lymphoma cells were sensitive to cytolysis by specific T cells and cytokine-activated NK cells, and no difference in sensitivity was observed with respect to PI-9 or bcl-2 expression. Cytolysis was mediated predominantly through perforin-dependent pathways despite expression of PI-9 and bcl-2. Interestingly, the majority of lymphoma cells were resistant to cytolysis by resting allogeneic NK cells. This was due to the failure of lymphomas to induce degranulation of resting NK cells. These results show that resistance to perforin-dependent pathways is not a relevant immune escape mechanism in lymphoma and therefore is unlikely to impair clinical outcome of immunotherapeutic approaches.

There is considerable interest in immunotherapeutic approaches for lymphoma. The expression of proteinase inhibitor 9 (PI-9), a molecule that inactivates granzyme B, is considered an immune escape mechanism in lymphoma. Further, lymphomas frequently overexpress the antiapoptotic molecule bcl-2, which is able to inhibit perforin-dependent cytotoxic pathways. In this study, the impact of PI-9 and bcl-2 expression on the sensitivity of lymphomas to T- and natural killer (NK) cell–mediated cytotoxicity was analyzed. We found PI-9 expression in 10 of 18 lymphoma cell lines and in 9 of 14 primary lymphomas. Overexpression of bcl-2 was found in 8 of 18 cell lines and in 12 of 14 primary lymphomas. All lymphoma cells were sensitive to cytolysis by specific T cells and cytokine-activated NK cells, and no difference in sensitivity was observed with respect to PI-9 or bcl-2 expression. Cytolysis was mediated predominantly through perforin-dependent pathways despite expression of PI-9 and bcl-2. Interestingly, the majority of lymphoma cells were resistant to cytolysis by resting allogeneic NK cells. This was due to the failure of lymphomas to induce degranulation of resting NK cells. These results show that resistance to perforin-dependent pathways is not a relevant immune escape mechanism in lymphoma and therefore is unlikely to impair clinical outcome of immunotherapeutic approaches.

CD4+CD25+regulatory T cells (Tregs) control immune responses to self- and foreign antigens and play a pivotal role in autoimmune diseases, infectious and noninfectious inflammation, and graft rejection. Since recent experimental studies have indicated that Tregs were able to ameliorate graft-versus-host disease (GvHD), we analyzed the number of infiltrating Tregs in the intestinal mucosa as one site of GvH reactivity using immunoenzymatic labeling to enumerate FOXP3+T cells in 95 intestinal biopsies from 49 allografted patients in comparison with healthy controls and patients with infectious inflammation. While patients with cytomegalovirus (CMV)-colitis or diverticulitis showed a concomitant increase of CD8+effectors and Tregs, acute and chronic GvHD were characterized by the complete lack of a counter-regulation indicated by a FOXP3+/CD8+T-cell ratio identical to healthy controls. In contrast, specimens without histologic signs of GvHD demonstrated increased numbers of FOXP3+per CD8+T cells, indicating that the potential for Treg expansion is principally maintained in allografted patients. Our findings provide evidence that GvHD is associated with an insufficient up-regulation of Tregs in intestinal GvHD lesions. The determination of FOXP3+/CD8+ratio can be a helpful tool to discriminate GvHD from infectious inflammation after allogeneic stem cell transplantation.

CD4+CD25+ regulatory T cells (Tregs) control immune responses to self- and foreign antigens and play a pivotal role in autoimmune diseases, infectious and noninfectious inflammation, and graft rejection. Since recent experimental studies have indicated that Tregs were able to ameliorate graft-versus-host disease (GvHD), we analyzed the number of infiltrating Tregs in the intestinal mucosa as one site of GvH reactivity using immunoenzymatic labeling to enumerate FOXP3+ T cells in 95 intestinal biopsies from 49 allografted patients in comparison with healthy controls and patients with infectious inflammation. While patients with cytomegalovirus (CMV)-colitis or diverticulitis showed a concomitant increase of CD8+ effectors and Tregs, acute and chronic GvHD were characterized by the complete lack of a counter-regulation indicated by a FOXP3+/CD8+ T-cell ratio identical to healthy controls. In contrast, specimens without histologic signs of GvHD demonstrated increased numbers of FOXP3+ per CD8+ T cells, indicating that the potential for Treg expansion is principally maintained in allografted patients. Our findings provide evidence that GvHD is associated with an insufficient up-regulation of Tregs in intestinal GvHD lesions. The determination of FOXP3+/CD8+ ratio can be a helpful tool to discriminate GvHD from infectious inflammation after allogeneic stem cell transplantation.

During the last 15 years, important progress has been made in the understanding of the biology and prognosis of myelodysplastic syndromes (MDS). It is a clonal disorder, characterised by ineffective haematopoiesis, which can lead to either fatal cytopenias or acute myelogenous leukaemias. Risk-adapted treatment strategies were established, due to the high median age (60 – 75 years) of the MDS patients and the individual history of the disease (i.e. number of cytopenias, cytogenetic changes, transfusion requirements). Allogeneic bone marrow transplantation currently offers the only potentially curative treatment, but this form of therapy is not available for the ‘typical’ MDS patient, who is > 60 years of age. Therapy with erythropoietin and granulocyte colony-stimulating factor has improved the quality of life of selected patients. The development of target-specific therapies, including antibodies and small molecules directed against specific molecular alterations in MDS, with minimal adverse effects, is the hope for the future. Furthermore, the innovative use of immunomodulatory agents and the optimising of cytotoxic treatment should continue to help in the treatment of MDS.

Exocytosis of perforin, subsequent binding of perforin to the target cell membrane, and formation of lytic pores form an important pathway involved in the induction of tumor cell death by cytotoxic effector cells. Here we describe a novel escape mechanism employed by tumor cells to protect themselves from granule-mediated cell death: We were able to demonstrate that the resistance of the human leukemia cell line ML-2 to natural killer (NK)-cell–mediated killing is not caused by impaired NK-cell activation but by resistance against effector molecules contained in the granules of cytotoxic cells. No resistance was observed against other pore-forming agents like complement and streptolysin O. By using the NK-susceptible leukemia cell line K562, we could show that the induction of cell death by cytotoxic granules can be blocked completely by anti-perforin antibodies, indicating that perforin is essentially involved in this process. Flow cytometric data revealed that an impaired binding of perforin on the tumor cell membrane is mainly responsible for target cell resistance, because perforin turned out to bind well on K562 cells but is not able to attach to the surface of ML-2 cells. After impaired binding of perforin was identified as a potential mechanism of tumor cell resistance, leukemia cells from 6 patients with acute myeloid leukemia (AML) were examined. As predicted, AML cells that failed to bind perforin on their surface demonstrated complete resistance toward NK-cell–mediated cytotoxicity. Thus, perforin resistance could represent an important tumor escape mechanism that should be considered when cytotoxic effector cells are used for cellular immunotherapy.

Peripheral blood progenitor cells (PBPCs) are increasingly being used to replace bone marrow cells (BMCs) as a source of hematopoietic stem cells also in the field of allogeneic transplantation. Whereas it is well known that PBPC grafts and BM differ significantly in progenitor cell content and lymphocyte dose, the clinical consequences of these differences with respect to engraftment, graft-versus-host disease (GVHD), and the graft-versus-leukemia (GVL) effect are more difficult to assess. We present a murine model that allows us to evaluate engraftment, GVHD, and GVL effect of allogeneic PBPC transplantation (PBPCT). Balb/c mice (H-2d) served as recipients. Donors were major histocompatibility complex-matched DBA/2 mice or syngeneic Balb/c mice, respectively. Experiments with increasing numbers of BMCs or Filgastrim-mobilized PBPCs showed that the number of progenitor cells in the graft was correlated with the probability to engraft, irrespective of the graft type. With identically high cell numbers transferred (1 × 109 nucleated cells/kg body weight [BW]), the mortality rates due to GVHD (25%) were about the same after allogeneic BM transplantation (BMT) and allogeneic PBPCT, although PBPC grafts contained four times more CD3+ T cells as compared with BM grafts (6.2 × 108v 1.4 × 108/kg BW). For investigation of GVL activity, Balb/c recipients were injected with syngeneic cells of the B-lymphocytic leukemia cell line A20 2 days before transplantation. After total body irradiation to a dose of 7.5 Gy, 1 × 109/kg BW Balb/c PBPCs, DBA BMCs, or DBA PBPCs were infused. The relapse rates observed were 80% after syngeneic PBPCT (n = 22), 60% after allogeneic BMT (n = 23), and 34% after allogeneic PBPCT (n = 26) (allogeneic BMT v PBPCT, P = .032). We conclude that transplantation of allogeneic PBPCs instead of BM may enhance the GVL effect without an increase of GVHD.

Ciceri: MolMed SpA: Consultancy. Bonini:TxCell: Membership on an entity's Board of Directors or advisory committees; Molmed SpA: Consultancy. Colombi:MolMed: Employment. Lambiase:MolMed: Employment. Bordignon:MolMed SpA: Employment.

Background:Haploidentical HSCT is a readily available platform for most patients lacking an HLA-matched donor, but the T-cell depleted (TCD) approach without any post-transplant donor cell therapy is impaired by high non-relapse mortality (NRM) due to slow IR, while the T-cell replete (TCR) approach followed by cyclophosphamide and immunosuppressive therapy is hampered by high chronic GvHD and relapse incidence (RI)

No relevant conflicts of interest to declare.

Off Label Use: 90Yttrium ibritumomab tiuxetan (Zevalin(R)), radioimmunotherapy of CD20+ lymphoma.

Background:Patients with aggressive B cell Non-Hodgkin-Lymphoma failing from immune-chemotherapy show disappointing results with standard salvage therapies and consolidation of high-dose chemotherapy with autologous stem cell transplantation. Especially in cases never in remission or with early relapse outcome is poor. Several groups started trials combining high-dose chemotherapy of carmustin, etoposide, cytarabine and melphalan (BEAM) with 90Yttrium ibritumomab tiuxetan to enhance the efficacy of the preparative regimen. In this trial we tested safety, feasibility and studied efficacy of reduced intervals of 90Yttrium ibritumomab tiuxetan to autologous stem cell transplantation after BEAM.

Al-Ali: Novartis: Consultancy, Honoraria, Research Funding; Celgene: Honoraria, Research Funding. Wolf:Bayer: Honoraria; Geo Pharma: Honoraria.

No relevant conflicts of interest to declare.

Niederwieser: Novartis, Gentium, Amgen: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau. Colombi:MolMed: Employment. Antonio:MolMed: Employment. Bordignon:MolMed: Employment.

The incidence of acute myeloid leukemia (AML) is age-dependent with the majority of patients (pts) being older than 60 years at diagnosis. Treatment of these pts needs to be well balanced between sufficient efficacy and tolerable toxicity. Here, we report long term follow-up of the OSHO AML97 study. Pts with AML older than 60 years were registered after informed consent and received age-adapted intensive chemotherapy treatment (curative arm; induction therapy with AraC 2 g/m2iv day 1, 3, 5, 7 and mitoxantrone 10 mg/m2iv day 1 to 3 to induce complete remission (CR), followed by 2 consolidation courses with AraC 240 mg/m2iv day 1 to 5 and mitoxantrone 10 mg/m2iv day 1 to 2), low dose chemotherapy (palliative arm; idarubicin 10 mg po day 1 and either thioguanine 40 mg po day 1-5, or AraC 80 mg sc day 1-5 or etoposide 100mg po day 1-5) or supportive therapy (best supportive care including transfusions). A total of 618 pts were enrolled (curative arm n=471, palliative treatment n=115 and supportive therapy n=32 pts). In the curative arm, CR was obtained in 66.8% of pts. Treatment related mortality (TRM) was 11.2% after induction and 4.5% after consolidation I, respectively. Median overall survival for all pts in the curative arm was 12 months, event free survival (EFS) at 12 years was 0.11±0.02%. In multivariate analysis, cytogenetics at diagnosis was the most important prognostic factor for CR (p=0.001). With a median follow up of 10 years (range 0.1 - 11.8) probability of overall survival (OS) at 5 years was 0.48±0.11; 0.13±0.03; 0.10±0.04 and 0.08±0.03 for pts with favorable, normal, other and unfavorable cytogenetics. Median survival for pts treated with palliative chemotherapy was 54 days.

BACKGROUND:Short telomere length has been described in CLL as independent adverse prognostic factor in multivariate analyses of several single centre series and clinical trial cohorts. However, the clinical impact of short telomeres has not been studied in the context of allogeneic stem cell transplantation (allo-SCT) in CLL.

Background: Relapse of acute leukemia (AL) after allogeneic hematopoietic stem cell transplantation (allo-HSCT) is frequently treated with 2ndallo-HSCT. Donor change has not yielded a significantly different outcome over choosing the original HLA-identical donor (Christopeit et al., JCO 2013). Using a haploidentical donor at second allo-HSCT might represent a feasible option (Tischer et al., BMT 2014)

Background: Haploidentical family donors represent the ideal solution to offer a potential cure for every high-risk leukemia patient undergoing HSCT. Widespread use of haploidentical HSCT has been historically limited by high rates of late non-relapse mortality (NRM) and relapse, which are associated with the delayed immune reconstitution (IR) due to either ex vivo T-cell depletion or in vivo post-HSCT cyclophosphamide given as graft-vs-host disease (GvHD) prevention. We have previously shown in a phase 2 trial (TK007, Lancet Oncol 2009;10:489) that TK cells (donor T cells genetically modified to express the HSV-TK suicide gene) can safely induce an early IR when given after T-cell depleted haploidentical HSCT

Acute graft-versus-host disease (GVHD) is a major complication after allogeneic stem cell transplantation (allo-SCT). With the current treatment approaches focussing on escalating immunosuppression, mortality rates of therapy-refractory disease courses are approx. 80-90%, underlining the medical need for novel therapeutic concepts. The pathophysiology of refractory GVHD is incompletely understood, however, recent evidence points towards a crucial role for endothelial damage. We have previously shown that steroid-refractory GVHD associates with increasing serum thrombomodulin (TM) levels 1, 2 and loss of endothelial TM expression 3, a constellation that indicates endothelial stress. However prior to allo-SCT, baseline TM levels did not predict steroid-refractory GVHD or survival. We hypothesized that endothelial vulnerability caused by single nucleotide polymorphisms within the TM gene (THBD) is a risk factor for refractory GVHD predicting outcome after allo-SCT.Seven SNPs within the THBD gene were studied (rs1962.TC, rs1042579.TC, rs1042580.AG, rs3176123.TG, rs3176124GA, rs3176126.GA and rs3176134.CT) in a training cohort of 465 allografted patients. The relevant genotypes were then re-assessed in an independent validation cohort (n=386).Allele frequencies of seven THBD SNPs allowed meaningful statistical correlation with outcome after allo-SCT. An increased risk of non-relapse mortality (NRM) was associated with three SNPs: rs1962.CC, rs1042579.TT (455V) and rs1042580.GG. When patients were divided into risk cohorts (one vs. no high-risk SNP), a strong correlation with NRM was observed in both cohorts (training cohort: p=0.002, HR 2.32 CI 1.36-3.95; validation cohort: p=0.007, HR 2.13, CI 1.23-3.70) (Figure 1). NRM was predicted by THBD SNPs in particular for patients who later develop GVHD (validation cohort: p=0.00056, HR 3.03 CI 5.68-1.61, training cohort: p=0.012, HR 2.38 CI 1.21-4.69) but not in patients without GVHD. In contrast, THBD SNPs did not predict incidence of acute GVHD. Multivariate analyses adjusting for clinical variables confirmed the independent effect of THBD SNPs on NRM in both cohorts.Our results demonstrate that THBD-SNPs predict mortality of GVHD patients without influencing GVHD incidence. This supports our hypothesis that endothelial vulnerability contributes to therapy refractory courses of the disease. Rather than focussing on the escalation of combinations of (endothelial cell toxic) immunosuppressive drugs, therapeutic approaches aiming at endothelial protection should be evaluated.1. Dietrich S, Falk CS, Benner A, et al. Endothelial vulnerability and endothelial damage are associated with risk of graft-versus-host disease and response to steroid treatment. Biol Blood Marrow Transplant 2013; 19(1): 22-7.2. Luft T, Dietrich S, Falk C, et al. Steroid-refractory GVHD: T-cell attack within a vulnerable endothelial system. Blood 2011; 118(6): 1685-92.3. Andrulis M, Dietrich S, Longerich T, et al. Loss of endothelial thrombomodulin predicts response to steroid therapy and survival in acute intestinal graft-versus-host disease. Haematologica 2012; 97(11): 1674-7.Dreger: Riemser Pharma : Consultancy, Honoraria, Research Funding.

Two prospective studies and small retrospective series have demonstrated feasibility and clinical efficacy of Azacitidine (Aza) as salvage therapy for relapse of AML or MDS after allo-HSCT. As a consequence, Aza has become a clinically relevant treatment alternative in this setting. Still, due to the heterogeneity and limited number of patients reported so far predictive factors for response and long-term survival are unknown.

Acute graft-versus-host disease (GVHD) is a major complication after allogeneic stem cell transplantation (allo-SCT). With the current treatment approaches focussing on escalating immunosuppression, mortality rates of therapy-refractory disease courses are approx. 80-90%, underlining the medical need for novel therapeutic concepts. The pathophysiology of refractory GVHD is incompletely understood, however, recent evidence points towards a crucial role for endothelial damage. We have previously shown that steroid-refractory GVHD associates with increasing serum thrombomodulin (TM) levels 1, 2and loss of endothelial TM expression 3, a constellation that indicates endothelial stress. However prior to allo-SCT, baseline TM levels did not predict steroid-refractory GVHD or survival. We hypothesized that endothelial vulnerability caused by single nucleotide polymorphisms within the TM gene (THBD) is a risk factor for refractory GVHD predicting outcome after allo-SCT.

Two prospective studies and small retrospective series have demonstrated feasibility and clinical efficacy of Azacitidine (Aza) as salvage therapy for relapse of AML or MDS after allo-HSCT. As a consequence, Aza has become a clinically relevant treatment alternative in this setting. Still, due to the heterogeneity and limited number of patients reported so far predictive factors for response and long-term survival are unknown.We analyzed data of 115 pts (median age 50 years, range 21-72 years) with AML (n=90), MDS (n=23) or myeloproliferative syndrome (MPS n=2), who had experienced relapse (hematological n=101; molecular n=14) after a median of 181 days (range: 19-3349 days) after allo-HSCT. Patients were treated with Aza ± DLI at 11 German transplant centers and in 109 of them (95%) Aza was the first treatment of relapse. Patients received a median of 3 courses Aza (range 1-14) and 78 patients (68%) received DLI (median number of DLI = 2, range: 1-6). Thirty patients were treated within a prospective phase-II trial (NCT-00795548). Their results have been published previously but were updated for this analysis.Following this treatment, 34 pts achieved CR (29%) and 8 patients achieved PR (7%) resulting in an overall response rate of 36%. Median time to CR was 84 days (range: 26-430 days) corresponding to 3 Aza cycles (range: 1-8 cycles). Of 6 patients with an extramedullary relapse, 3 patients achieved CR. The 2-year OS rate was 34%. With a median follow-up of 17 months (range 1-82) for survivors, 22 pts (65%) are in ongoing CR for a median of 16 months (range 2-48) without further treatment, while 12 pts relapsed again after a median of 11 months (range 2-48). Bone marrow blast count (< 20% vs. ≥20%, CR rate 47% vs. 12% p=0.0001) and type of relapse (molecular vs. hematological, CR rate 71% vs. 24%, p=0.0007) were identified as predictive factors for the achievement of CR. No correlation was found for any karyotype abnormalities, time interval from allo-HSCT to relapse, the presence of blasts in the peripheral blood and the diagnosis of MDS vs de-novo AML. Incidence and severity of acute GvHD (overall: 25%, grade I: 10%, grad II: 6%, grade III: 7%, grade IV: 2%) and chronic GvHD (overall 23%, limited 19%, extensive 4%) were low and mild.This analysis shows that the combination Aza and DLI is a valuable treatment option in the challenging situation of relapse after allo-HSCT. The association of disease burden with the likelihood to respond emphasizes the need for stringent disease monitoring and early intervention.Schroeder: Celgene: Financial travel support Other, Honoraria. Bug:Celgene: Honoraria, Travel grants Other. Platzbecker:Celgene: Honoraria. Kobbe:Celgene: Research Funding.

of this analysis was to provide long-term follow-up of the GCLLSG CLL3X trial which aimed at evaluating reduced-intensity HSCT in patients with poor-risk CLL, and to compare the impact of the newly identified risk factors SF3B1 and NOTCH1 mutations on the outcome of HSCT with that of established prognostic factors, such as TP53 abnormalities.CLL3X included 100 patients with poor-risk CLL (median age 53 (27-65) years), of whom 90 patients were allografted with blood stem cells from related (40%) or unrelated donors (60%). 24% had refractory CLL at HSCT, and 13% received in-vivo T cell depletion (TCD) with alemtuzumab during conditioning. All genetic studies were performed in the Ulm central reference laboratory of the GCLLSG. Genomic aberrations and IGHV mutation status were analyzed by FlSH and direct sequencing as previously described. NOTCH1 was studied by direct sequencing of a PCR fragment from the PEST domain (exon 34, chr9:139,390,619-139,391,290). SF3B1 (exons 13–16) and TP53 (exons 4–10) were analyzed by DHPLC (WAVE® 3500HT, Transgenomic Inc) with subsequent sequencing of aberrant fragments, if needed after subcloning.With a median follow-up of 6.0 (0.6-10.8) years, 6-year overall survival (OS) of all 100 patients enrolled was 53%. The 90 allografted patients had a 6-year OS and event-free survival (EFS) of 58% and 38%, respectively. Of 28 patients who were relapse-free and MRD-negative 12 months after HSCT, only 3 subsequently experienced clinical disease recurrence at 2.8, 5.0, and 9.0 years post transplant, and 24 (86%) remained MRD-negative throughout the whole follow-up. SF3B1, NOTCH1, and TP53 mutational status could be obtained each in 80/100 patients, and 17p FISH results were available in 82/100 patients. SF3B1 abnormalities were found in 26%, NOTCH1 abnormalities in 14%, and TP53 mutations in 30% of patients; whereas 17p- was detected in 20% of patients with diagnostic material available. Whilst SF3B1 and NOTCH1 mutations were mutually exclusive, 29% of patients with TP53 mutations also had an SF3B1 or NOTCH1 abnormality, and 46% of TP53-mutated patients had a concomitant 17p deletion. By univariate comparisons, we did not find significant survival differences between SF3B1-mutated, NOTCH1-mutated and SF3B1+NOTCH1 wild-type patients (Figure). Multivariate analysis using Cox regression modeling did not show a significant impact of SF3B1, NOTCH1, and TP53 (TP53mut and/or 17p-) abnormalities on OS and EFS (Hazard ratios (HRs) of SF3B1mut 0.82, 95%CI 0.35–1.97 and 0.64, 95%CI 0.32–1.31; HRs of NOTCH1mut 0.69, 95%CI 0.22–2.12 and 0.73, 95%CI 0.29–1.81; HRs of TP53abn 0.99, 95%CI 0.42–2.32 and 0.61, 95%CI 0.31–1.21). In contrast, TCD and refractory disease at HSCT retained their adverse impact on OS and EFS as already observed in an earlier analysis at a follow-up time of 3.8 years (Blood Oct 7, 2010).HSCT can provide long-term EFS in approximately 40% of patients with poor-risk CLL. Patients who are in MRD-negative remission one year after HSCT have a >80% probability of durably remaining in this status. Long-term disease control does not appear to be affected by the presence of SF3B1, NOTCH1, and TP53 abnormalities, suggesting that HSCT can overcome the treatment resistance found to be associated with this abnormality.Hallek: Roche: Consultancy, Honoraria, Research Funding. Stilgenbauer:Roche: Consultancy, Honoraria, Research Funding.

Abstract 1964

Treatment with azacitidine (Aza) and donor lymphocyte infusions (DLI) can induce sustained remissions in some patients (pts) with AML or MDS relapsing after allogeneic stem cell transplantation (allo-SCT). Meanwhile incidence and severity of GvHD seems to be relatively low when compared to historical data using DLI alone. As a potential mechanism murine models have suggested that Aza upregulates the transcription factor FoxP3 thereby expanding CD4+ regulatory T cells (Tregs). This has also been recently shown in 17 AML pts receiving Aza maintenance therapy following allo-SCT (Goodyear et al., 2012).To confirm and expand this knowledge we monitored CD4+CD25+FoxP3+ Tregs and lymphocyte subsets (CD3+; CD3+/CD4+; CD3+/CD8+; CD3−/CD56+; CD20+) by flow cytometry in 46 pts during salvage therapy with Aza (up to 8 cycles either 100 mg/m2/day d1-5 or 75 mg/m2/d d1-7) and DLI (envisaged on day 34/90/146) for relapse following allo-SCT. PB samples were obtained prior treatment (d0), after the 1st (d6), 2nd (d34), 4th (d90) and 6th cycle (d146). To assure a serial measurement only pts who had received at least 4 Aza cycles were eligible. Thereby 13 pts could be included, while 33 pts were excluded as a consequence of early drop-of resulting from progression or death (n=21), or due to missing samples (n=12).Relapse of AML (n=8) or MDS (n=5) occurred in median 446 d (range:19–1688) following allo-SCT in these 13 pts. They received a median of 6 Aza cycles (range: 4–8). DLI were administered in all patients with a median number of 2 DLI per patient (range:1–4) resulting in a median total T cell dose of 5.0×106CD3+ cells/kg per patient (range:1–119). A CR rate of 62% (n=8) was observed in these 13 pts being overestimated in comparison to a CR rate of 33% (n=15) in the whole group due to positive selection of pts.Prior to relapse 6 pts (46%) had suffered from aGvHD (Io 1 pt, IIo 1 pt, IIIo 3 pts, IVo 1 pt) and 2 pts (15%) from cGvHD (limited 1 pt, extensive 1 pt). At the beginning of Aza treatment 3 pts were still on immunosuppresion which could be tapered in all cases without GvHD flare. Following treatment with Aza aGvHD was observed in 5 pts (overall 38%, Io 3 pts, IIIo2 pts) in median 129 d (range: 20 – 253) following the 1st DLI, while cGvHD developed in 6 pts (overall 46%, limited 5 pts, extensive 1 pt).In concordance with this rather mild presentation of GvHD, a 1.5-fold increase of Tregs was observed after 4 Aza cycles (d0: 8.23/μl vs. d90: 13.26/μl, p=0.0479). By grouping the pts on the basis of the median time to relapse (day 446), we observed a 3.2-fold increase of the absolute number (d0: 4.7/μl vs. d90: 14.8/μl, p=0.031) as well as an 1.9-fold increase of the frequency of Tregs (d0: 6.7% vs. d90: 12.9% of CD3+CD4+ cells, p=0.06) during treatment with Aza in the group of patients who relapsed early. On the other hand, in those patients who relapsed late the absolute number of Tregs (d0: 12.2/μl vs. d90: 11.9/μl, n. s.) was already higher and remained together with the Treg frequency (d0: 4.7% vs. d90 3.9%, n. s.) unchanged during treatment. Of interest, in those patients with early relapse only 1 pt developed aGvHD (Io), in contrast to 4 pts with aGvHD in those with late relapse. No significant changes were observed with regard to other lymphocyte subpopulations.We here demonstrate an intra-individual Treg expansion, which might be induced by Aza and may explain the low rate and mild presentation of GvHD observed following the combination of Aza and DLI. In line with the data of Goodyear et al. Aza-induced expansion of Tregs seems to be restricted to patients relapsing early after allo-SCT where the Treg repertoirs is still immature.Schroeder: Celgene: Travel support Other. Platzbecker:Celgene: Honoraria, Research Funding, Speakers Bureau. Bug:Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Speakers Bureau, Travel support Other. Germing:Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Research Funding. Kröger:Celgene: Research Funding. Kobbe:Celgene: Research Funding.

Abstract 966

Relapse after allo-SCT is a major cause of treatment failure in patients (pts) with myeloid malignancies and is associated with a poor prognosis. As therapeutic options are limited, treatment of these pts is challenging. Indeed, there is a need for novel strategies which ideally target the leukemic clone and direct the immune system towards an enhanced GvL effect without aggravating GvHD. The hypomethylating agent Aza might provide these properties, and results from retrospective studies investigating Aza+/−DLI in pts with AML/MDS relapsing after allo-SCT were encouraging.We here report the final results from a prospective single-arm EBMT multicenter phase II trial which aimed to investigate the efficacy and safety of a combination of Aza and DLI as 1st salvage therapy in pts with AML or MDS with hematological relapse after allo-SCT. Treatment schedule contained up to 8 cycles Aza (100 mg/m2/d d1-5, every 28 d) followed by DLI with increasing dosages (1-5×106–1-5×108cells/kg) after every 2ndAza cycle.A total of 30 pts (19 f/11 m, median age 56 years, range 29–71) from 6 German transplant centres were included between January 2009 and May 2010. The majority of pts (n=28, 93%) suffered from AML, while 2 pts (8%) had MDS or MDS/MPN, respectively. At transplant, 16 pts (53%) had active disease (6 induction failure, 7 relapse I/II, 3 untreated) and 14 pts (47%) were in remission (12 CR1, 2 CR2). Following standard-dose (n=4, 13%) or dose-reduced conditioning (n=26, 87%), 10 pts (33%) received a graft from MSD and 20 pts (67%) from MUD. PBSC were used in 29 pts (97%), while 1 pt (3%) received BM. Acute GvHD occurred in 13 pts (46%) and 4 pts (13%) had chronic GvHD prior inclusion. None of the pts had active GvHD at the time of relapse, but 6 pts were still on immunosuppressive therapy. All pts had hematological relapse (median BM blasts: 34%, median chimerism: 67%) at a median time of 160 days (range 19–1699) following allo-SCT. A median of 3 courses Aza (range 1–8) were administered, and 22 of 30 pts (73%) received DLI (median: 1, range: 1–4, median CD3 dose 5×106/kg/DLI, range: 1–100×106). Overall response rate was 47%. Seven pts (23%) achieved CR or CRi, 2 pts (7%) PR, and 5 pts (17%) had stable disease (SD). Median time and median number of Aza cycles to best response were 79 days (range 28–299) and 3 cycles (range 1–8) respectively. Of the 7 pts who achieved a CR, 5 pts continue in CR for a median of 605 days (range 307–763) without any additional treatment, while 1 pt relapsed after 490 days and 1 pt died from GvHD. By July 2011 median follow-up of surviving pts is 645 days (range 564–857) and 5 of 30 pts (17%) are currently alive. Twelve pts have died due to progressive disease (PD), while 7 pts died during (n=3, 2 infection, 1 bleeding) or after the end of therapy (n= 4, 1 GvHD, 2 infection, 1 bleeding). All 5 pts who underwent 2nd allo-SCT died. Median overall survival (OS) of all pts is 117 days (95% CI 66–168 days). Patients with CR/CRi had a significant longer OS than pts not reaching CR/CRi (not reached vs. 83 days, p<.001). Eleven pts (37%) developed aGvHD, while cGvHD was observed in 5 pts (17%). Any type of GvHD was seen in 78% of pts with CR/PR as opposed to 33% in pts with SD/PD (p=.03). Cytopenias grade III/IV occurred in all pts, but were considered to be drug-related in only 11 pts (37%). The most common drug-related grade III/IV non-hematological toxicities were infection and bleeding.Aza and DLI as first salvage therapy is a safe and active approach for pts with AML or MDS who relapse after allo-SCT, and induces durable remissions in a subgroup of pts. Further research to better define target patient groups and combination partners for AZA+DLI is needed.Schroeder: Celgene: Financial travel support. Bug:Celgene: Lecture fees, Membership on an entity's Board of Directors or advisory committees. Luft:Celgene: Research Funding. Kobbe:Celgene: Financial travel support, Research Funding.

Abstract 1019

Kobbe: Celgene: Consultancy, Research Funding; Ortho Biotec: Consultancy. de Witte:Novartis: Consultancy, Honoraria, Speakers Bureau.

Abstract 656