ALBERT

Description: By reducing treatment intensity allogeneic hematopoietic stem cell transplantation (allo-HSCT) has become feasible for elderly patients. Different reduced-intensity conditioning (RIC) regimens are available, but there is little consensus about the optimal preparative regimen to use, in particular with regard to the outcomes counterbalancing the aim of feasibility and tolerability with higher rates of relapse. Here, we retrospectively evaluate the outcome of sequential therapy employing RIC with fludarabine 30 mg/m2, cytarabine 2g/m2 and amsacrine 100 mg/m2 for 4 days (FLAMSA; Schmid C et al. JCO 2005) followed by busulfan 10 x 0.8 mg/kg (FLAMSA-Bu) compared to RIC utilizing fludarabine 5 x 30 mg/m2, carmustine (BCNU) 2 x 150 mg/m2 and melphalan 110 mg/m2 (FBM; Marks R et al. Blood 2008) in elderly patients treated consecutively at our institution between July 2005 and October 2012. We analyzed the course of 114 patients (pts) with acute myeloid leukemia (AML; n=99) or myelodysplasia (MDS; n=15) aged ≥ 59 years with 59 pts aged ≥ 66 years who were treated with either FLAMSA-Bu (n=66; n=24 ≥ 66 years) or FBM (n=48; n=35 ≥ 66 years). All patients received sero-therapy with anti-thymocyteglobuline (ATG). Median patient age was 66 years for the entire cohort (68 years FBM; 64 years FLAMSA-Bu). 36 patients (75%) of the FBM and 42 patients (63 %) of the FLAMSA-Bu group suffered from high risk disease defined as relapsed or refractory AML or refractory anemia with excess of blasts in transformation (RAEB-T). The hematopoietic cell transplantation comorbidity index (HCT-CI) was higher for the patients of the FBM group than for the FLAMSA-Bu group with 26 (54 %) versus (vs) 24 patients (36 %) scoring ≥ 2 (p 0.085). Graft source after conditioning with FBM/FLAMSA-Bu was bone marrow (1/2), G-CSF mobilized peripheral blood stem cells (40/62) and double-umbilical cord-blood (7/1). In 23 pts (20 %) HLA-matched related and in 91 pts (80 %) HLA-matched unrelated donor transplantation was performed. Engraftment failure was observed in 1 patient after FLAMSA-Bu, while engraftment was achieved in all evaluable patients of the FBM group in a median of 23 days vs 18 days after FLAMSA-Bu (p 0.003), while 7 pts with double-umbilical cord-blood transplantation where included in the FBM group vs 1 pt in the FLAMSA-Bu group. Non-hematological treatment-related acute toxicity ≥ CTC III (gastrointestinal, hepatic, cardiovascular, renal, centralnervous system) occurred in 12/48 pts (25 %) after FBM and in 18/66 pts (27 %) after FLAMSA-Bu. Incidence of severe acute (III-IV) and chronic GvHD was 22.9 %/16.6 % for FBM vs 18.2 %/19.7 % for FLAMSA-Bu, respectively. After conditioning with FBM 2/48 pts vs 9/66 pts after FLAMSA-Bu were diagnosed with a secondary malignancy (p 0.08). Non-relapse mortality (NRM) after 12 months was 26.8 % for FBM versus 25.2 % for the FLAMSA-Bu group. Incidence of relapse after FBM vs FLAMSA-Bu conditioning was 22.9 % vs 15.2 % after 1 year and 31.3 % vs 16.7 % after 2 years. Occurrence of relapse was significantly related to an incomplete or mixed chimerism (donor cells ≤ 95 % in peripheral blood and/or bone marrow) at day +30 (p 0.001). After a median follow up of 31.4 months (range 4.4-97.5) estimated overall survival (OS) and relapse-free survival (RFS) after 2 years was 55.4 % and 51.4 % for the FBM vs 58 % and 56.7 % for the FLAMSA-Bu group, respectively. Analyzing different subgroups, FBM conditioning might be favorable for pts aged ≥ 66 years when suffering from high risk AML (n=26): Within this group 1-year OS after FBM vs FLAMSA-Bu was 71.4 % vs 66.7 % (p 0.58) and 1-year RFS was 71.4 % vs 58.3 % (p 0.59), respectively. Notably, for pts at highest risk (aged ≥ 66 years and suffering from secondary or therapy-related AML; n=24) the benefit of FBM conditioning becomes more pronounced: 1-year OS after FBM vs FLAMSA-Bu was 62.5 % vs 37.5 % (p 0.26) and 1-year RFS 54.2 % vs 37.5 % (p 0.17). Both conditioning regimens are feasible, and provide similar rates of acute toxicity, NRM and GvHD. There might be evidence for a benefit of conditioning with FBM for the subgroup of “the oldest patients at highest risk”. Taking into account that there is an increasing group of ‘medically fit’ elderly patients in the field of allogeneic transplantation, prospective clinical trials are needed to investigate different conditioning regimens considering their special requirements. Disclosures No relevant conflicts of interest to declare.

Description: Myelodysplastic syndrome (MDS) and secondary acute myeloid leukemia (sAML) are diseases of the hematopoietic stem cell that are potentially curable by allogeneic stem cells transplantation (ASCT). MDS and sAML are most frequent in elderly patients with significant co-morbidity. Here we report our results with reduced intensity conditioning regimen and preemptive donor lymphocyte transfusions that have been successful in patients with advanced and refractory AML. Sixty seven patients were treated for MDS and 90 patients for sAML. The FLAMSA regimen consisted of a 4 day course of chemotherapy (Ara-C 2g/m2, Fludarabin 30 mg/m2, and amsacrin 100 mg/m2) followed by 3 days rest and reduced intensity conditioning with 4 Gy total body irradiation (TBI), cyclophosphamide (CY) +/− antithymocyte globulin (ATG). Overall survival at 10 years was 50% for MDS and 26% for sAML (p=0.014, log rank). In the MDS group most patients had preleukemic forms (IPSS 3 and 4). Survival at 5 years was 58% with FLAMSA, 44% with myeloablative conditioning with TBI and 67% with Busulfan (BUS). Non-relapse mortality at 5 years was 41% with FLAMSA, 20% for BUS and 50% for TBI. The relapse or progression rate was 0% for FLAMSA, 24 % for BUS and TBI respectively. In sAML the 5 year survival was 17% for FLAMSA, 57 % for BUS and 30% for TBI. Non-relapse mortality was 60% for FLAMSA, 36% for BUS and 66% for TBI, whereas the relapse rate was 62% for FLAMSA, 12.5 % for BUS and 37% for TBI (p=0.04). In multivariate analysis adjusted for disease, conditioning treatment and source of stem cells – marrow vs. mobilized blood – differences were not significant and the only significant prognostic factor was age less than 30 years (p=0.003). Therefore patients of the age of 60 years and older with AML were treated with FLAMSA containing BUS (8mg/kg in two days) instead of 4 Gy TBI. The overall survival at 2 years was 64%, 61% with FLAMSA-TBI and 70% with FLAMSA-BUS. The non-relapse mortality was 9% for FLAMSA-BUS and 37% for FLAMSA-TBI; the relapse rate was 25% at 2 years. At day 120 patients received donor lymphocytes (DLT) in 3 escalating doses from 1 × 106/kg, 5 × 106/kg on day 150 and 1 × 107/kg on day 180, if they had no GVHD, infection or relapse. Preemptive DLT appear to prevent relapse when compared to a historical control and contemporary patients not given DLT. We conclude from these data that MDS/sAML may have a different biology in young patients, and that in the elderly we can optimize the treatment with FLAMSA-BUS as a dose reduced regimen and preemptive immunotherapy for maintenance of remission.

Description: The presentation of leukemic antigens can be improved in AML and MDS by in vitro conversion of leukemic cells in leukemia-derived DC (DCleu), thereby forming a platform for the generation of leukemia-specific cytotoxic lymphocytes (CTL). In preliminary analyses with 140 AML and 60 MDS-cases we could already define optimal serum-free culture conditions to generate DC/DCleu.(Kufner 2005 I–III). Now we want to predict or correlate the clinical response to a DC/CTL-based immunotherapy by detailed analyses of the ex vivo generated/activated DC/DCleu and T-cells: 1)By a combination of 3 different DC-generating methods (‘MCM-mimic’, Lee 2003; ‘Ca-Ionophore’, Houtenbos 2003; ‘Picibanil’, Sato 2003) we can generate DC/DCleu in every case of AML/MDS, independently from FAB-type or karyotype. DC/DCleu are quantified according to their surface DC/blast-marker profiles. On average 42–45%/39–66% DC in AML/MDS could be generated with 48–54%/39–51% mature (CD83+) and 31–34%/23–31% migratory (CCR7+) DC. 45–65% of DC were ‘DCleu’; on average 47% of blasts are convertible to DCleu.. 2) In AML-patients who had presented with a relapse after SCT we could correlate a better ex vivo convertibility of blasts to DCleu with the patients’ in vivo response to a GM-CSF/Donor-lymphocyte Infusion (DLI)-therapy of their relapse after SCT (33% vs 7% to DCleu convertible blasts in ‘non-responders’). 3) A ‘Mixed lymphocyte culture’ (MLC) of autologous AML-patients’ or allogeneic donor-T-cells showed an on average higher proliferation and stimulation of DC-primed compared to MNC-primed T-cells: Upregulation of CD80/CD86-CD28;CD40-CD154;CD137L-CD137; moreover DC-priming yielded higher proportions of CD4+ cells, CD3+CD45RO+ memory cells CCR4+ T-cells (+59%, +52%, +91%) compared to MNC-primed T-cells (+35%, +13%, +44%) and a higher leukaemia-cytolytic activity (average 62%) compared to MNC-stimulated CTL (average 26%). 4) A detailed analysis of data showed great individual variations depending on the quality and composition of DC and T-cells: a) non-DC-primed autologous or allogeneic T-cells an lead to an increase of naive blasts after 3h incubation with blasts b) in cases with an ineffective DC-mediated ex vivo lysis of naïve blasts lower proportions of mature DC (29% vs 63%), DCleu (41% vs 68%) or a reduced T-cell proliferation or even loss of CD4/CD8/memory T-cells were seen. In summary our data show 1. that DC/DCleu can be generated in every single AML/MDS-case. 2. Grade of ex-vivo generability of DC/DCleu correlates with the in vivo response to a GM-CSF/DLI-relapse therapy. 3. Composition and quality of DC and autologous or donor T-cells after DC-priming provides informations about the activability and quality of CTLs in individual patients. We conclude, that ex vivo analysis of the DC/anti-leukemic T-cell-activability is necessary to develop and select promising anti-leukemia-directed T-cells for the immunotherapy of AML and MDS.

Description: Abstract 783FN2 Introduction: The advent of second generation tyrosine kinase inhibitors (TKI) in the front line treatment setting of chronic myeloid leukemia (CML) has tightened the evaluation of imatinib response. Early assessment of response markers might identify slow responders harboring a BCR-ABL positive clone with an inferior susceptibility to tyrosine kinase inhibition. This group of patients could benefit from an early dose escalation or a change of treatment to a second generation TKI thus avoiding the risk of disease progression. Therefore we sought to evaluate the impact of molecular and cytogenetic response levels after 3 months of imatinib treatment on the further course of disease. Patients and methods: A total of 1,340 patients (median age 52 years, range 16–88, 40% female) were included into the randomized German CML study IV and treated with an imatinib based therapy consisting of imatinib 400 mg/d (n=381), imatinib 800 mg/d (n=399) and combinations of standard dose imatinib with interferon alpha (n=402) and low-dose cytarabine (n=158). Median follow-up was 4.7 years (range 0–9). Molecular response after 3 months was assessed in 743 patients, cytogenetic response in 498 patients. The BCR-ABL expression was determined by quantitative RT-PCR and standardized according to the international scale (BCR-ABL IS). Only patients expressing typical BCR-ABL transcripts (b2a2, b3a2, b2a2 and b3a2) were considered. Cytogenetic response was determined by conventional metaphase analysis. Disease progression was defined by the incidence of accelerated phase, blastic phase or death from any reason. A landmark analysis was performed for progression free survival (PFS) and overall survival (OS). Results: Disease progression was observed in 149 patients (11.1%), 127 patients died (9.5%). After 3 months of treatment the median BCR-ABL IS was 2.6% (0-100), the median proportion of Philadelphia chromosome positive metaphases (Ph+) was 8% (0-100). The BCR-ABL landmarks of 1% and 10% after 3 months of imatinib both proved to discriminate significantly for PFS and OS: BCR-ABL IS 65% (n=84), p=0.004 for PFS and p=0.011 for OS. A 3 group stratification did not reach statistical significance. Conclusions: The achievement of molecular and cytogenetic response landmarks after 3 months of imatinib treatment is predictive for long term progression free and overall survival. At 3 months a BCR-ABL IS of 10% or more is associated with a 5-year overall survival of 87% suggesting an early change of treatment, whereas a BCR-ABL IS of 1% or less indicates a favorable 5-year overall survival of 97%. Disclosures: Schnittger: Münchner Leukämie Labor: Equity Ownership. Haferlach:Münchner Leukämie Labor: Equity Ownership. German CML Study Group:Deutsche Krebshilfe: Research Funding; Novartis: Research Funding; BMBF: Research Funding; EU: Research Funding; Roche: Research Funding; Essex: Research Funding.

Description: Abstract 324 Introduction and classification: This is the largest adult T-ALL cohort treated according to immunologic subtypes. All patients were immunophenotyped in one central lab (Berlin). T-ALL (cyCD3+, CD7+) were subclassified into early T-ALL (sCD3-, CD1a-), thymic T-ALL (sCD3-/+, CD1a+) and mature T-ALL (sCD3+, CD1a-). T-ALL constitutes in 3 consecutive GMALL-studies 24% of ALL patients. Patients and methods: A total of 744 T-ALL pts (15 to 55 yrs) were accrued in 102 hospitals in the GMALL studies 05/93, 06/99 and 07/2003. In GMALL 05/93 239 adult T-ALL patients, were treated according to a multi-agent chemoprotocol. Stem cell transplantation (SCT) was not recommended in CR1. In GMALL studies 06/99 and 07/03 505 T-ALL pts received intensified chemotherapy; particularly with introduction of PEG-asparaginase in induction as well as HDMTX/PEG-Asp consolidation cycles. Based on study 05/93 results, SCT from sibling (Sib) as well as matched unrelated (MUD) donor in CR1 was recommended for all patients with early T-ALL, mature T-ALL and for high-risk (HR) pts with thymic T-ALL (defined as late CR, complex karyotype or MRD positivity (MRD+)). Results: T-ALL subtype distribution in the total cohort of 744 adult T-ALL was early-T 23% (N=170), thymic-T 56% (N=420), mature-T 21% (N=154), without any differences between the studies. GMALL Study 05/93: The overall CR rate was 86% (early-T 72%, thymic-T 93%, mature-T 84%. The lower CR rate in early T-ALL was mainly due to early death (19%). The overall CCR rate was 47% (early-T 45%, thymic-T 54%, mature-T 30%). The overall survival rate at 10 yrs for all pts was 47% (early-T 47%, thymic-T 55%, mature-T 25%). GMALL Study 06/99 and 07/03: Of the 505 patients, 87% achieved CR (early-T 84%, thymic-T 92%, mature-T 77%). PR/Failure was higher in early-T (13%) and mature-T (17%) compared to thymic-T (5%). Early death was 4% and equally distributed. 267 pts (64%) received chemotherapy only and the majority were 229 pts (86%) with thymic T-ALL, not considered for SCT in CR1. The CCR rate was 61%. The few early (n = 15) and mature (n = 23) T-ALL pts, which could not have a transplant in CR1, are a negative selection (e.g. early relapse, comorbidity, no donor) and their CCR rate was 33% and 22% respectively. This was due to a high relapse rate in early T-ALL (60%) and mature-T (74%) compared to 33% in thymic-T. Overall survival rate at 8 yrs for thymic T-ALL with chemotherapy was 68%, for the 77 adolescent pts (15 to 25 yrs) even 76%. Stem cell transplantation: 153 T-ALL pts in studies 06/99 and 07/03 received a SCT in first remission. SCT realisation rate in early T-ALL was 84%, in mature-T 68%. Overall CCR rate was 58% (early-T 47%, HR thymic-T 79%, mature-T 61%). Relapse rate after SCT was in early-T 33% and in mature-T 22%. The overall TRM rate was 18% despite more than half MUD SCT, without any TRM difference between the immunological subtypes. Overall survival rate after SCT in CR1 at 8 yrs was 53%, early-T 44%, thymic-T 67%, mature-T 59%. SCT modalit: 49% received alloSib, 55% alloMUD and 20% auto-SCT. Overall CCR rate after alloSib for the total cohort was 65% (early-T 60%, thymic-T 73% and mature-T 69%); for alloMUD total 55% (early-T 45%, thymic-T 77%, mature-T 61%) and for the small cohort of 20 pts with auto-SCT CCR was 35%. Conclusion: The strategy in three consecutive GMALL studies to stratify and treat adult T-ALL pts according to the immunologic T-subtypes was successful. Overall survival at 5 yrs could be improved to 56% from 44%. There was a particular improvement for mature T-ALL (49% vs. 30%) and early-T (40% vs. 33%). This was mainly due to a high realisation rate of SCT in early T-ALL and mature T-ALL and the substantial better results of SCT. Results of alloMUD SCT were comparable to alloSib SCT. The small cohort of HR thymic T-ALL pts also had a benefit from SCT. The excellent outcome of SR thymic T-ALL (∼ 50% of all T-ALL) with the OS of 68% and 76% in adolescents due to intensified chemo, partic. PEG-Asp, does not suggest SCT in CR1. Several molecular markers, such as ERG, BAALC, WT1, had in a retrospective analysis some prognostic relevance in this pt cohort. The new GMALL study generation will however focus in thymic T-ALL on early evaluation of MRD to decide for SCT (MRD+) or not (MRD-) whereas early/mature T-ALL remain allocated to high risk groups with SCT in CR1. Supported by Deutsche Krebshilfe 702657Ho2 and BMBF 01GI9971/8 Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 862 Rapid relapse after discontinuation of imatinib, the need for indefinite therapy and residual disease in most patients are the major challenges in management of CML. Combinations of imatinib with IFN simultaneously, or consecutively preceding imatinib, or with araC may improve treatment outcome. The German CML Study Group therefore designed a randomized trial to compare standard imatinib vs. imatinib + interferon alpha (IFN) vs. imatinib + low dose araC vs. imatinib after IFN failure (for low- and intermediate-risk patients, high risk patients received imatinib 800 mg instead). The current evaluation represents the prefinal results of the pilot phase of the trial. Inclusion criteria were newly diagnosed BCR/ABL positive CML in chronic phase (CP). Primary aims are: prolongation of survival (overall, OS, and progression free, PFS), determination of rates of hematologic, cytogenetic and molecular remissions, adverse events (AE) and role of allografting. By the end of 2005, 670 patients were randomized, 13 had to be excluded (no CML (n=3), pregnancy, no CP (n=1 each), imatinib 800 mg (n=8)). Analysis was according to intention to treat. 657 patients were evaluable (174 with imatinib 400 mg, 196 with imatinib+IFN, 158 with imatinib+araC and 129 with imatinib after IFN-failure). 656 patients were evaluable for hematologic, 611 for cytogenetic, and 618 for molecular responses. Patient characteristics of treatment arms were similar for age (median 53 years), sex (40% female), median values for Hb (12.6 g/dl), WBC (66.2/μl), platelets (383/μl) and for Euro risk score (low 35%, intermediate 54%, high 10%). The median dose of imatinib was 400mg/die in all arms, of araC 10 mg per treatment day and of IFN 4.2 Mio I.U./die in the imatinib after IFN arm and 1.8 Mio I.U./die in the imatinib+IFN arm. Median observation time was 57.3 months. 55 patients died, 73 patients were transplanted in 1st CP, 81 patients progressed, 59 patients were switched to second generation TKIs. After 3 years 126 patients (72%) of the imatinib 400mg arm still received the initial therapy as well as 60 patients (30%) of the imatinib+IFN arm and 53 patients (34%) of the imatinib+araC arm. 9 patients (7%) of the imatinib after IFN arm are still on IFN. 5-year OS of all patients is 91%. 5-year PFS of all patients (no death, patient still in first chronic phase) is 87%. 5-year-OS and PFS according to treatment arm are shown in the Table. At 5 years, the cumulative incidences of achieving complete cytogenetic remission or major molecular remission (MMR) as determined by competing risks (death, progression) are not different (Table). Type and severity of adverse events (AE) over a 5-years period did not differ from those reported previously (Table). Hematologic AEs grade III/IV were similar in all therapy arms except leukopenia grade III/IV, which was more frequently observed in the imatinib after IFN arm (14%). Non hematologic AEs were mainly fluid retention, neurological and gastrointestinal symptoms and fatigue. Neurologic symptoms and fatigue were more often reported for the therapy arms with IFN. Imatinib 400mgImatinib+IFNImatinib+AraCImatinib after IFN5-Year Survival and Response RatesOS87%93%92%92% PFS84%91%88%84% CCR92%92 %89%83% MMR83%78%80%70% Adverse Events, WHO Grade III/IVAnemia7%1%3%3% Leukopenia4%5%2%14% Thrombocytopenia5%6%6%6% WHO Grade I-IVEdema15%13%5%0% Neurological5%15%5%22% Gastrointestinal17%27%21%15% Fatigue8%13%9%23% This analysis shows excellent survival and durable response rates in all arms. Currently, survival in all treatment arms is equal to, or better than in IRIS. To verify possible differences in survival, e.g. imatinib 400 mg vs. imatinib + IFN, longer observation is planned. Although cytogenetic and molecular responses in the imatinib after IFN failure arm at 5 years are inferior to that in the other treatment arms, the question of whether the consecutive therapy with IFN first and imatinib after IFN-failure provides a survival advantage requires long term follow-up. Imatinib in combination with, or after IFN, or with low dose araC are feasible and safe treatment modalities. We expect that the study will optimize and improve therapy outcome in CML. Disclosures: German CML Study Group: Deutsche Krebshilfe: Research Funding; Novartis: Research Funding; German Competence Net : Research Funding; European LeukemiaNet: Research Funding; Roche: Research Funding; Essex: Research Funding.

Description: Acute graft-versus-host disease (aGvHD) contributes significantly to morbidity and mortality after allogeneic hematopoietic stem cell transplantation (allo-HSCT). Diagnosis of GvHD is mainly based on clinical features and tissue biopsies. A noninvasive, unbiased laboratory test for GvHD diagnosis does not exist. Here we describe the application of capillary electrophoresis coupled online with mass spectrometry (CE-MS) to 13 samples from 10 patients with aGvHD of grade II or more and 50 control samples from 23 patients without GvHD. About 170 GvHD-specific polypeptides were detected and a tentatively aGvHD-specific model consisting of 31 polypeptides was chosen, allowing correct classification of 13 of 13 (sensitivity 100.0% [95% confidence interval {CI} 75.1 to 100.0]) aGvHD samples and 49 of 50 (specificity 98.0% [95% CI 89.3 to 99.7]) control samples of the training set. The subsequent blinded evaluation of 599 samples enabled diagnosis of aGvHD greater than grade II, even prior to clinical diagnosis, with a sensitivity of 83.1% (95% CI 73.1 to 87.9) and a specificity of 75.6% (95% CI 71.6 to 79.4). Thus, high-resolution proteome analysis represents an unbiased laboratory-based screening method, enabling diagnosis, and possibly enabling preemptive therapy.

Description: Non-HLA polymorphisms (NHP) influence risk of GVHD and outcome of allogeneic hematopoietic stem cell transplants (HSCT) however their influence on GvHD vs GvL remains to be defined. A cohort of 291 CML HLA matched sibling transplants with known clinical risk factors; eg stage of disease, gender mismatch (female donor/male recipient), patient age and time from diagnosis to transplant as defined by the EBMT risk score, were typed via SNPs or microsatellites for cytokines (IL-1Ra, IL-4, IL-6, IL-10, IFNγ, TNFα, TNFR 11), steroid hormone receptors (VDR and ERα) and NOD2/CARD15 mutations. TNFRII-196 allele R; IL-10 ATC/ACC; IL-1 Ra (allele 2) and IL-4T were significantly associated with survival using univariate analysis. Two clinical Cox proportional hazards models were generated for the statistical analysis and used as a basis for further development: (i) using the EBMT risk score as a single variable on an ordinal scale or (ii) using the individual clinical factors of the EBMT risk score as categorical variables. After step-wise variable selection using the significant genetic factors as candidates, the resulting multivariate models indicated that absence of TNFRII-196 R, i.e. down regulation of TNF in the recipient, absence of IL-10 ATC/ACC, i.e. intermediate IL-10 production in the donor and presence of IL-1Ra (allele 2) i.e. down regulation of IL-1 in the donor were associated with poor outcome. The addition of the genetic variables significantly improved the preferred model containing the EBMT risk score as a single variable. The Goodness of Fit of the models was assessed by Kaplan-Meier curves showing clinically relevant differences between good, intermediate and poor prognostic groups. The worst prognostic scores included the absence of ATA/ACC in the donor, evidenced by a steep change in survival probability. Relapse was associated with clinical factors; absence of female to male transplants; absence of bone marrow transplants and presence of T cell depletion but no significant association was found with genetic factors. This study suggests that distinct high risk patterns of NHP of patients and donors can be defined, which influences survival due to factors associated with an increased risk of GvHD without the potential benefit of increased GvL response. Data add to the clinical factors (eg age, sex, multiparity of the donor) where an unrelated donor might be the preferred choice compared to a high risk sibling donor.