ALBERT

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: To improve the results of allogeneic SCT for high-risk AML and MDS, the FLAMSA-RIC conditioning regimen for allogeneic SCT combines cytoreductive chemotherapy (fludarabine, HD AraC, amsacrine), followed three days later by reduced intensity conditioning (4Gy TBI/EDX). Since in particular patients with an unfavorable karyotype seemed to benefit from this approach (Schmid et al., JCO, 2005), we analysed the outcome of 172 patients with poor risk cytogenetics according to NCCN criteria, who had been allografted following FLAMSA-RIC conditioning in 11 European centres between 1999 and 2008. Median time from diagnosis to transplantation was 5 months. Donors were matched siblings, matched unrelated, or mismatched unrelated donors in 34%, 47%, and 19%. Patients suffered from progressive MDS (10%), de novo AML (47.5%), or secondary AML (43.5%). SCT was performed upfront, after primary induction failure, in CR1 and in relapsed disease in 17%, 33%, 22% and 28% of patients, respectively. Median patient age was 53 (18–71) years. 95 patients (56%) had a complex aberrant karyotype, 55 and 65 had abnormalities of chromosome 5 (−5/5q-) and 7 (−7/7q-), respectively. After a median follow up of 20 months, overall survival (OS) at 2 and 4 years was 46.4% and 40.5%, the respective leukemia-free survival was 37.7% and 32.0%. Causes of death were leukemia in 30%, and non-relapse mortality in 21%. Encouraging results were observed in patients with chromosome 7 aberrations or with a complex karyotype leukemia (4y OS=49.3% and 40.3%). In contrast, results were inferior in patients with chromosome 5 aberrations (4y OS=30%), mainly due to an increased rate of leukemia-associated death (p=.008). Patiens with MDS, who received allogeneic SCT as first line treatment, achieved a 4y OS of 80% despite unfavorable cytogenetics. Unlike, patients with secondary AML after MDS had an inferior outcome (4y OS=28%, p=.018). In a Cox regression model, a stage of remission at transplantation, a 8/8 or 10/10 matched family or unrelated donor, and lack of monosomy 5 or deletion 5q were associated with superior OS (p=.025, .05, and .05). In conclusion, allogeneic SCT following the FLAMSA-RIC regimen is a highly effective treatment for MDS and AML with unfavorable karyotype, comparing favourably with published data. In MDS, SCT should be performed before transformation into sAML. Long term remission is achieved in a substantial percentage of patients with complex karyotype disease and aberrations of chromosome 7. Aberrations of chromosome 5 may require alternative or additive strategies.

Description: No standard treatment is available for relapse of acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) after allogeneic stem cell transplantation (alloSCT). Efficacy of donor lymphocyte transfusion (DLT) was limited, at least in part, by the rapid pace of the disease, overwhelming any allogeneic immune response. However, induction chemotherapy before DLT was complicated by severe toxicity. Based on a pilot trial conducted in Munich, a prospective, multicenter phase II trial for relapsed AML and MDS after alloSCT was initiated in 2000 by the German transplant cooperative group. The study evaluated the sequence of low-dose(ld)AraC for leukemic control (intensive chemotherapy for progressive disease only), transfusion of donor PBSC without immunosuppression as adoptive immunotherapy, and systemic application of GM-CSF. GM-CSF was included, since in combination with other cytokines, it has been able to improve the antigene presentation capacity of myeloid blasts in vitro. Between 2000 and 2006, 41 patients with hematological relapse of AML or MDS 〉3 months after alloSCT were included. Median age was 47y, 50% had an unrelated donor. Median remission after SCT was 223d (93–1614), median percentage of BM blasts at relapse was 40%. Median follow up was 28 months. Control of leukemic proliferation by ldAraC was achieved in 61%, allowing outpatient care in the majority of these patients. 39% required intensive chemotherapy. Three patients died from infections before donor cell transfusion (DCT), one patient was not transfused due to progressive leukemia. Median time from relapse to DCT was 52 days. 25/34 evaluable patients were found to be free of blasts in BM at d35 after PBSC and were considered initial responders. In an intent-to-treat analysis, overall survival (OS) of the entire cohort at 1 and 2 years from relapse was 41% and 32%. Among initial responders, 1y- and 2y-OS was 68% and 49%. A remission 〉6months after alloSCT, and control of leukemia by ldAraC prior to DCT were associated with better outcome. At last follow up, 8 patiens were in continous CR, 2 were alive with second relapse, 18 had died from leukemia, and 13 had died in remission or aplasia. In conclusion, ldAraC seems to be effective for initial control of leukemic proliferation in relapsed AML and MDS after alloSCT. A longer duration of post-transplant remission and response to ldAraC may identify patients who will benefit from adoptive immunotherapy. On an intend-to-treat-basis, the overall results of our trial compare favorably to other published strategies; nevertheless, outcome is still unsatisfying and warrants further investigation.

Description: Abstract 347 Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment for many hematologic malignancies or hematopoietic dysfunction syndromes, but the application is still limited due to major complications, such as severe graft versus host disease (GvHD) and infectious complications. Diagnosis GvHD is based on clinical features and biopsies, a non invasive, unbiased laboratory test does not exist. To date the proteomic pattern specific for aGvHD was evaluated blindly on 961 samples collected from 345 patients undergoing allo-HSCT at MHH and 7 additional clinics, including the University of Michigan. The majority of the patients included were transplanted for hematological malignancies (n=329), 16 for hematopoietic failure syndromes, mainly severe aplastic anemia. Conditioning regimens included dose reduced conditioning regimens (FLAMSA and ClaraC for the majority of the patients of MHH), as well as standard conditioning regimens (TBI+Cy or Busulfan+Cy) for about 35% of the patients. GvHD-prophylaxis was cyclosporine A (CSA) and mycophenolate (MMF) or CSA metothrexate (MTX) as appropriate. In addition, about 80% percent of the patients received ATG (antithymocyte globulin) prior to HSCT. A peptide pattern of 31 peptides - either absent/decreased (15) or present /increased (16) - was previously published (Weissinger et al. 2007). Prospective and blinded evaluation of the patients included in this diagnostic analysis for early recognition of patients at risk for aGvHD development revealed the correct classification of patients developing aGvHD about 7 days prior to the development of clinical symptoms for aGVHD with a sensitivity 76% and specificity of about 85% (fig.1). Additional data obtained from patients transplanted until September 2009 will be reported. Based on these data a pre-emptive therapy multicenter trial, administering steroids upon positivity of the proteomic pattern has been initiated now in 10 German centers, testing the efficacy of the pre-emptive therapy on incidence and severity of aGvHD and a possible benefit on overall survival of the patients. Disclosures: Krons: mosaiques-diagnostics GmbH: Employment. Metzger:mosaiques-diagnostics GmbH: Employment.

Description: Abstract 1164 Poster Board I-186 Allogeneic hematopoietic stem cell transplantation (allo-HSCT) is a curative treatment for many hematologic malignancies or hematopoietic dysfunction syndromes, but the application is still limited due to major complications, such as severe graft versus host disease (GvHD) and infectious complications. Diagnosis chronic GvHD is based on clinical features and biopsies, a non invasive, unbiased laboratory test does not exist. We used the urine collected from 20 patients (10 with limited cGvHD, 10 with extensive cGvHD) to establish a proteomic pattern that allowed the diagnosis of cGvHD development and tested the resulting set of polypeptide markers (27 differentially excreted peptides) on more than 200 patients prospectively and blinded for the correct classification of cGvHD samples. The majority of the patients included were transplanted for hematological malignancies (n=209), 6 for hematopoietic failure syndromes. Conditioning regimens included dose reduced conditioning regimens (FLAMSA and ClaraC for the majority of the patients of MHH), as well as standard conditioning regimens (TBI+Cy or Busulfan+Cy) for about 35% of the patients, with GvHD-prophylaxis including cyclosporine A and mycophenolate (MMF) or metothrexate (MTX) as appropriate. Eighty percent of the patients received ATG (antithymocyte globulin) prior to HSCT. A peptide pattern of 27 peptides, differentiating chronic from acute GvHD was developed. Controls were patients at least 100 days post HSCT, with no GvHD in the history, no infections and without relapse at the time of sampling. Prospective and blinded evaluation of the patients revealed the correct classification of patients developing cGvHD with a sensitivity of 85% and specificity of 95%.Further evaluation of the cGvHD patterns specific for particular organ manifestations of cGvHD are currently ongoing. Interestingly, the cGvHD pattern seems to be predictive for GvHD developing post DLI, while the aGvHD-specific proteomic pattern only predicts GvHD of the intestine, which may be more similar to “late acute GvHD”. Disclosures Krons: mosaiques-diagnostics GmbH: Employment. Metzger:mosaiques-diagnostics GmbH: Employment.

Description: Introduction Myelodysplastic syndromes (MDS) are diagnosed at median age of 70 years. Allogeneic stem cell transplantation (HSCT) is the only curative treatment option, but with an increasing age, morbidity escalates. Treatment guidelines suggest HSCT for intermediate-II and high risk constellations up to the age of 65, and reduced intensity conditioning (RIC) regimens are commonly used up to 70 years of age. However, increasing life expectancy, availability of RIC regimens and good Karnofsky performance status (KPS) of MDS patients more than 70 years of age, has led to an increased use of HSCT. We performed a retrospective analysis to investigate results after HSCT for those patients and influence of KPS on outcome. Patients and methods We analyzed data of 345 patients in the EMBT database older than 70 years with MDS/sAML. The disease status at transplantation was available in 233 patients and most of the them were in more advanced stage of the disease: RA/RARS,RCMD (n=25) , RAEB (n=68) and RAEB-T/secondary acute leukemia (sAL, n=140). Donor were: related (n=88) and unrelated (n=257). Cytogenetic data were available only in 73 patients and classified as good (58), intermediate (6), poor (5) and very poor (4). Median follow up was 29.7 months. Median age at transplantation was 72 years (70-79 years) with 249 male and 96 female patients. KPS was defined in 300 cases, being 90-100% in 61% and 80% or less in 39%. Stem cell source was peripheral blood (94%) or bone marrow (6%). The intensity of the conditioning regimen was mainly reduced intensity (78%) rather than myeloablative (22%). Negative or positive CMV sero-status of the patient were seen in 35% and 65%, respectively. Results The number of HSCT for MDS patients of 70 years or more has increased over time. While 2000-2004 only 19 patients received transplantation, the following 3-year periods included 28 (2005-2007), 97 (2008-2010) and 200 (2011-2013) patients, respectively. The estimated 3-year OS was 33% (27-39%). A significant better 3 year OS in the univariate analysis was seen for Karnofsky (90-100%) vs 80% or less (41 vs 23%, p=0.008) and for CMV negative sero-status (46% vs 27%, p〉

Description: Background: Patients (pts) relapsing with CML after allogeneic hematopoietic stem cell transplantation (alloHSCT) may be treated with tyrosine kinase inhibitors (TKI) and/or donor lymphocyte infusions (DLI). As nowadays the majority of CML patients would have received at least imatinib prior to transplantation, we were interested in analizing a) the type of TKI used after alloHSCT, b) the indication for TKI treatment, c) the outcome of this treatment and d) the temporal relationship with DLI if given. Patients and methods: 435 pts received TKI after first allogeneic HSCT for CML for different reasons. Transplants had been performed in first chronic phase (CP1, n=194), accelerated phase (AP, n= 60) or for more advanced disease (blast crisis (BC)/〉 CP1, n=177) from HLA identical siblings (n=231) or unrelated donors (n=204) between 2000 and 2013. TKI given prior to transplant was imatinib (n=268), dasatinib (n=162), nilotinib (n=88), bosutinib (n=4) and ponatinib (n=7). Median age at transplant was 44 (18.5-68) years, 274 pts (63%) were male. TKI post alloHSCT were given between 2000 and 2015. 1st TKI given was either imatinib (n=223), dasatinib (n=131), nilotinib (n=67), bosutinib (n=2) or ponatinib (12). The indications for TKI therapy were the same as for transplantation (n=262), for relapse/progression/persistent disease (n=124), for prophylaxis/pre-emptive (n=32), planned (n=5), others (n=8) and missing (n=4). Results: Median follow-up from start of TKI was 55 (1-171) months. The median time interval from transplant to TKI was 6 (0.2-165) months. It was longer for TKI given for relapse/progression with 15 (1-89) months and shorter for TKI given for prophylaxis/pre-emptive with 1.6 (0.2-43) months. It was longer for imatinib with 11 (0.2-121) months vs 3.8 (0.2-165) months for other TKI. Imatinib as 1st TKI was mainly given for relapse/progression/persistent disease (48%) and the other TKI for the same reason as for transplantation (83%). 103/223 (46%) of pts with imatinib, 99/131 (76%) with dasatinib, 55/67 (82%) with nilotinib and 11/14 (79%) with bosutinib/ponatinib post-transplantation had been treated with imatinib prior to transplantation. In total, 196 (45%) patients received DLI after alloHSCT, of which 63/435 (14.5%) had DLI prior to TKI post-alloHSCT, 19/435 (4.4%) had DLI at the same time of TKI and 114/435 (26%) had DLI post-TKI. Best response after TKI was complete molecular remission in 17.7%, cytogenetic remission in 4.4%, hematological remission in 20.2% and no response/progression/relapse in 57.7% of pts. 50% of pts treated with imatinib had a response (molecular/cytogenetic/hematological) vs 34% with nilotinib, 33% with dasatinib and 33% with bosutinib/ponatinib, p=0.014. OS was 60% (55-65%) at 5 years. It was 66% (60-73%) with imatinib vs 51% (42-60%) with other TKI, p=0.0024. 5 years RFS was 47% (42-53%). It was 53% (46-60%) with imatinib vs 40% (32-48%) with other TKI, p=0.0102. 5 years RI was 25% (21-30%). It was 21% (16-27%) with imatinib vs 31% (24-38%) with other TKI, p=0.0454. 5 years NRM was 27% (23-32%). It was 26% (20-31%) with imatinib vs 29% (22-36%) with other TKI, p=0.365. In multivariate analysis for OS, imatinib vs other TKI post-transplant did not show anymore an effect, HR 1.19 (0.85-1.67), p=0.317. Factors influencing OS were time from diagnosis to transplant, HR 1.01 (1.00-1.01), p=0.009, AP vs CP1, HR 1.80 (1.11-2.91), p=0.017 and BC/〉CP1 vs CP1, HR 2.3 (1.58-3.33), pCP1 vs CP1, HR 2.11 (1.55-2.88), p

Description: Albeit clinical experience is limited adjuvant transfusion of donor lymphocytes (aDLT) may reduce the risk of relapse after allogeneic stem cell transplantation. We report our data on aDLT in high-risk AML. Cells were given within a prospective protocol that contained a sequence of chemotherapy, reduced intensity conditioning for allogeneic transplantation, and aDLT. For aDLT, patients had to be in CR at least 120 days from transplantation, off immunosuppression and free of GvHD. 30% of patients (n=46) alive at day +120 fulfilled the criteria for aDLT. They had been transplanted (24 from matched unrelated donors) for refractory (n=11) or relapsed leukemia (n=24) or in CR1 because of unfavourable cytogenetics (n=7) or other unfavourable criteria (n=2) or in CR2 with unfavourable cytogenetics (n=2). Twenty-four patients had an unfavourable karyotype, 10 with complex aberrations. Thirty-one patients with similar disease characteristics and fulfilling similar selection criteria (being alive in CR at d +120, no cGVHD and no history of aGVHD〉 II°) transplanted during the same time period served as control. Of these 10 pts. were transplanted at a center not using aDLT and 21 at the center in Wiesbaden during 2000 and 2002 prior to the introduction of the protocol at this center. The median time from transplant to first aDLT was 160 days (range 71–303). Median follow up of the surviving transfused patients is 594 days (range 113–1920). Nine patients received 1, 18 patients received 2, and 19 patients received 3 transfusions in escalating doses, containing a median of 1×106, 5×106 and 1×107 CD3+ cells/kg at aDLT 1, 2 and 3, respectively. Reasons for giving less than 3 transfusions were GvHD, relapse or refusal of the patient. Induction of GvHD was the main complication; grade II/III acute GvHD developed in 4, and chronic GvHD in 8 patients. So far, 8 (17%) patients have relapsed despite aDLT, as compared to 42% in the control group (Chi-square: p=0.018). Two died of refractory leukemia, whereas 5 achieved a secondary CR following adoptive immunotherapy. At the time of analysis (April 2006), 39/46 patients were alive and all but one patient who is actually receiving adoptive immunotherapy are in CR at a median of 441 days post DLT. The actuarial overall survival two years after transplant is 85% as compared to 53% in the control group and at four years 78% vs. 40%, respectively (p=0.005). In conclusion, aDLT is safe, when given in escalating doses and to a selected group of patients. Results are encouraging, and improved long term survival can be achieved.

Description: Introduction: Antithymocyte globulin (ATG) is increasingly incorporated into conditioning in unrelated stem cell transplantation (UD-HCT) in an attempt to modulate alloreactivity. Several brands of ATG with different antibody spectrum and potencies are available. Little is known about their optimal use, their clinical action and their potentially different effects in this setting. We reported on the immune reconstitution of 108 patients (pts) after UD-HCT between 1998 and 2003 at our center, after conditioning with either rabbit ATG Genzyme (Thymoglobulin®, ATG-G) or ATG Fresenius (ATG-F) (BMT2004, 33, S3). We here analyze the clinical results of these pts. Patients: 66 pts (cohort 1) received ATG-G, the following 42 pts (cohort 2) ATG-F. Median (md) age of the cohort (co) was 40 y, underlying diseases were AML 40%, MPS 24%, ALL 14%, MDS 7%, lymphoma 15%. There were no significant differences (sigdif)between groups regarding age, Karnofsky index, underlying disease, disease or risk status, CMV risk status and HLA match. 69 % of donor/recipient pairs were 8/8 matched, 31 % had at least one class I Ag MM or one class II allel MM. There was a trend towards more female donors for male recipients (15 vs 7%) in co1 whereas co2 had more pAML (36 vs 17%), more use of reduced conditioning (47 vs 26%), PBSC (60 vs 42%) and MMF instead of MTX (26 vs 5%). Three quarters of ATG-G pts received the product from day (d)-5 to -2, the later ones from d-4 to -1, whereas ATG-F was given from d-3 to -1. The vast majority of co1 had a daily ATG-G dose of 2.5 mg/kg BW compared to 15–20 mg/kg BW ATG-F in co2. Results: At a md follow up of 938 d (1051 for co1 and 908 for co2) there was no sigdif in engraftment, chimerism, disease status or overall survival (OS). Probability of OS for co1 is 63% and 69 % for co2. ATG-G pts were more likely to be readmitted after discharge (73 vs 57%), time to readmission was significantly shorter. Infection (46 vs 29 %) and CMV reactivation (23 vs 14%) were the most frequent reasons for hospitalization with a md number of infectious episodes of 3 for ATG-G and 2 for ATG-F pts from d30 to 365. Infections were categorized as equally severe in both groups. 64% of ATG-G pts developed aGvHD °II–IV compared to 43% of ATG-F pts - however aGvHD responded well to treatment (complete resolution in 63 vs 57%). 76% of co1 and 67% of co2 pts developed cGvHD with a maximum severity of extended cGvHD in 46 vs 26%. ATG-F pts had more skin (57 vs 44%) and gut cGvHD (21 vs 9%) whereas liver (33 vs 14%), lung manifestations (12 vs 5%) and wasting (14 vs 5%) were more frequent after ATG-G. cGvHD in co1 was more often judged to be severe (11 vs 5%), treated by more than one modality in 85 vs 55%, lasted longer (md of 296 vs 69 d) and was more often the primary cause of death (9 vs 2%). Conclusion: Retrospective comparisons of sequential cohorts are subject to biases and have to be interpreted with great caution. Better donor selection, new immunosuppressive and anti-infectious drugs and less toxic conditioning regimens may work in favour of co2. Both treatment regimens gave good results in matched and mismatched UD-HCT without sigdif in OS and disease free survival. However differences in the occurrence and type of cGvHD seem to emerge which need further evaluation. Our results suggest that at least at the given schedule the probably lower equivalence dose of ATG-F was able to prevent at least as much severe cGvHD as ATG-G. A prospective randomized trial with ATG given from d-3 to -1 is warranted.

Description: Introduction: MMM is a clonal myeloproliferative disorder of later life with a wide range of life expectancy from months to many years. Allogeneic hematopoietic stem cell transplantation (HSCT) is the only curative treatment option, however associated with considerable morbidity and mortality. We analyzed the results of HSCT in 15 consecutive patients (pts) with MMM referred to our institution during the last four years in order to determine the best time point for transplantation. Patients and Methods: Ten males and five females with primary (n = 13) and secondary MMM following myeloproliferative syndrome (n = 2) were grafted with unmanipulated bone marrow (n = 10) or peripheral stem cells (n = 5) from a related (n = 6) or unrelated (n = 9) donor. Median (md) age was 49 (36 – 64) years, md time from diagnosis to transplant 27 (6 – 132) months. Five pts had low, eight intermediate and two high risk disease (Dupriez score 0, 1, 2). All intermediate risk pts were categorized Dupriez 1 because of hemoglobin levels (Hb) 〈 10 g/dl. In low risk pts indication for transplant were severe constitutional symptoms. Mean risk score was 0.4 (0–1) in related and 1.1 (0–2) in unrelated transplants. 8 pts were transfusion dependent and two had portal hypertension. Eleven pts received a conditioning of intermediate intensity containing TBI(8Gy)/Flud/Cy (n = 10) or Bu(12)/Flud (n = 1). Four pts had standard TBI/Cy or Bu/Cy (n = 2 each). GvHD prophylaxis consisted of CsA and short course MTX. In all unrelated and two related transplants rabbit ATG (Genzyme n = 6, Fresenius n = 5) was added before grafting. Results: All pts engrafted at a md of 23 (11 – 28) days. Four pts developed extended cGvHD. Four pts died at a md of 266 (177–407) days: One from liver cirrhosis, one from cerebral aspergillosis and two from infections associated with cGvHD. At a md follow up of 21 (4 – 53) months probability of disease free survival (DFS) is 66% for the whole cohort and 100% vs 45% for pts with a related versus an unrelated donor. Results of MUD transplants are confounded with the higher risk score of pts grafted from an unrelated donor, as DFS according to disease risk was 100% for score 0 and 52% for score 1 and 2. Need for red blood cell transfusions before transplant was an even better predictor of DFS with 100% in untransfused and 36% in transfusion dependent pts. A log rank test was performed for the following variables: Conditioning intensity, donor type (MUD vs MRD), stem cell source, Dupriez score and need for blood transfusions: Only transfusion dependence proved to be significant (p = 0.03). Conclusions: Although the number of pts is small our results support the use of related and unrelated allogeneic SCT as a curative treatment option in MMM. Hb below 10 g/dl has been reported to be a predictor of adverse outcome after HSCT (Guardiola et al, Blood93, 1999: 2831; Deeg et al, Blood102, 2003: 3912). In our hands however, red cell transfusion dependence is the relevant cut off point for survival as pts with risk score 1 without need for transfusions had the same DFS as pts with risk score 0. Transfusion dependence may be a marker of disease progression and of increased toxicity at a time. AlloHSCT therefore should be performed earlier during the natural course of the disease, ie before transfusion dependence occurs, when DFS is excellent.