ALBERT

Description: Introduction: CD22 is widely expressed on leukemic lymphoblasts in patients with B-cell acute lymphoblastic leukemia (ALL). Inotuzumab ozogamicin (InO), an anti-CD22 monoclonal antibody conjugated to calicheamicin, has shown significantly higher remission rates than standard of care (SC) chemotherapy in relapsed or refractory B-cell (R/R) ALL, independently of CD22 expression level. Here we report safety outcomes by CD22 expression in patients with R/R ALL receiving InO (vs SC) as salvage therapy in the INO-VATE trial (NCT01564784). Methods: Adults with CD22-positive ALL in 1st or 2nd salvage were randomized to InO (n=164; starting dose 1.8 mg/m2/cycle [0.8 mg/m2 on day 1; 0.5 mg/m2 on days 8 and 15 of a 21-28 day cycle for ≤6 cycles]) or SC (n=162; fludarabine/high-dose (HD) cytarabine (Ara-C)/granulocyte colony-stimulating factor, Ara-C plus mitoxantrone, or HD Ara-C).The InO dose was reduced to 1.5 mg/m2 per cycle in patients who achieved complete remission [CR] or CR with incomplete hematologic recover [CRi]. Last patient visit was January 4, 2017. Central flow cytometry was used to assess CD22 expression, which was quantified as % leukemic blasts CD22-positive and as Molecules of Equivalent Soluble Fluorochrome (a quantitative measure of receptor density on leukemic blasts [MESF]). Outcomes were reported in patients by baseline leukemic blast positivity (≥90% vs

Description: Background: Allogeneic hematopoietic stem cell transplantation (alloHSCT) is a standard of care for patients with acute myeloid leukemia (AML) being at high risk of relapse. However, the procedure using conventional myeloablative conditioning regimens based on either busulphan (Bu) or total body irradiation (TBI) in combination with cyclophosphamide is associated with significant risk of non-relapse mortality (NRM). The use of reduced-intensity or non-myeloablative conditioning regimens results in improved tolerance but increased incidence of disease recurrence. In order to reduce toxicity while maintaining the efficacy, alternative approaches have been proposed including the use of moderately reduced doses of Bu or TBI in combination with fludarabine (Flu). The goal of this retrospective study to compare BuFlu and TBI/Flu reduced-toxicity regimens. Patients and methods: Adult patients with AML treated in CR1 with alloHSCT from either HLA-matched sibling or unrelated donor between January 2006 and June 2018 were included in the analysis. The following conditioning regimens have been selected for the comparison: intravenous Bu at a total dose 9.6 mg/kg (3 days) + Flu (Bu3Flu, N=350) or TBI at a total dose of 8 Gy + Flu (TBI8Flu, N=168). In both groups the proportion of patients with intermediate risk karyotype was 74% while high risk - 24%. The proportion of patients with secondary AML was also the same (14%). Patients in the Bu3Flu group were significantly older, were treated more frequently with alloHSCT from unrelated donors and were treated in more recent period. Results: The engraftment rate was 99% for both regimens. In a univariate analysis the use of TBI8Flu was associated with reduced incidence of relapse (20% vs. 30% at 2 years, p=0.01) and tendency to increased leukemia-free survival (LFS; 66% vs. 60%, p=0.15) and overall survival (OS; 74% vs. 58%, p=0.051) as well as reduced incidence of grade III-IV acute graft-versus-host disease (GVHD, 4% vs. 9%, p=0.03). The effects on non-relapse mortality (NRM), grade II-IV acute GVHD and chronic GVHD were not significant. In a multivariate analysis a chance of LFS was reduced for patients with high risk karyotype and secondary AML, with no effect of donor type. Due to significant interaction between type of conditioning and age, further analyses were performed stratifying patients above or below 50 years. For patients ≤50 y.o. the use of TBI8Gy was associated with reduced incidence of relapse (21% vs. 35%; Cox model: HR=0.49, p=0.049), tendency to improved NRM (4% vs. 7%,; HR=0.17; p=0.1), significantly improved LFS (75% vs. 58%, p=0.02; HR=0.45, p=0.02), improved OS (84% vs. 60%; HR=0.29, p=0.002) improved survival free from both GVHD and relapse (GRFS; 56% vs. 46%; HR=0.53, p=0.02) and reduced incidence of grade II-IV acute GVHD (15% vs. 26%; HR=0.53, p=0.02). For patients 〉50 y.o. the effect of TBI8Flu on relapse was not statistically significant (19% vs. 29%; HR=0.64, p=0.21) while this regimen was associated with increased risk of NRM (26% vs. 11%; HR=3.98, p=0.0006) leading to a tendency to decreased OS (59% vs. 63%; HR=1.53, p=0.1). Conclusion: Reduced-toxicity regimens using either TBI8Flu or Bu3Flu for patients with AML in CR1 are associated with relatively low risk of relapse and NRM leading to enhancing survival rates after alloHSCT. The use of TBI8Flu appears more effective compared to Bu3Flu and may be advised in younger patients where reduced risk of relapse translates into improved survival. In older individuals it should be used with caution due to increased risk of NRM. Prospective studies are needed to verify our findings. Disclosures Labopin: Jazz Pharmaceuticals: Honoraria. Byrne:Ariad/Incyte: Honoraria, Speakers Bureau. Beelen:Medac GmbH Wedel Germany: Consultancy, Honoraria. Blaise:Jazz Pharmaceuticals: Honoraria; Molmed: Consultancy, Honoraria; Sanofi: Honoraria; Pierre Fabre medicaments: Honoraria. Mohty:Jazz Pharmaceuticals: Honoraria, Research Funding.

Description: Introduction: Immunocompromised patients are increasingly threatened by multidrug resistant bacterial strains. Following allogeneic stem cell transplantation (SCT), patients are especially at risk for multidrug resistant Pseudomonas aeruginosa (MRPA) which as a pathogen can be the cause of life-threatening septicemias. A possible reservoir and source of nosocomial infections are biofilms in shower drains, sinks, and toilets. For this reason, substantial targeted renovation was performed in the bathrooms of our SCT unit. We subsequently performed extensive surveillance of patients status as carriers and took numerous environmental samples. Methods: Renovation/ conversion works of our transplant unit (2 wards with 20 HEPA filtered patients rooms, all equipped with separate shower baths) began in 2011 and included the installation of Pall-Aquasafe water filters (Pall Cooperation, Port Washington, NY) on all faucet and shower heads. Washbasins were equipped with Biorec Disinfection siphons (MoveoMed, Radebeul, Germany), which use UV and temperature together with frequent vibration and a bactericidal coating for disinfection. In addition, the toilets were replaced with rimless toilet basins and a system was developed to supply the flushing water with disinfectant at appropriate concentrations. The shower outlet drains were modified and covered by a heavy metal lid that cannot be easily removed and prevents aerosol formation. All renovation work was completed in April 2014. Cleaning and disinfection takes place once a day and the removable metal devices of the showers are sterilized once a week as well as every time a patient is discharged. In addition, extensive training of all personal in infection prevention was performed and the work of the cleaning personal was supervised closely. Environmental sampling was performed monthly in every room. Patients were screened upon admission and weekly during their treatment on the SCT unit by either anal swabs or stool cultures. Standard antibiotic prophylaxis consisted of ciprofloxacin and metronidazole for patients undergoing allogeneic SCT. Metronidazole prophylaxis was omitted in August 2013. Results: Peri-renovation environmental sampling detected MRPA in 22,6% of swabs from the toilets and 11,8% of swabs from the shower outlets in 2012. The rate decreased to 15,1% from the toilets and 4,4% from the shower outlets in 2013 and reached 10,1% from the toilets and 0,6% from the shower outlets in the first six months of 2014. In 2012, 15 out of 197 patients (7.6%), receiving an allogeneic SCT or treatment for severe post-transplant complications (mainly GvHD) were affected by MRPA. This number decreased to 13 out of 206 (6.3%) in 2013 and 2 out of 95 (2.1%) in the first half in 2014. Overall 18 patients (3.6%) were colonized only and 12 patients (2.4%) developed bacteremia from MRPA. Discussion: Environmental sources must be taken into consideration for the prevention of nosocomial infections. The implemented measures in our transplant unit can serve as an example for the reduction of biofilm, which can act as a reservoir for MRPA. Because of inconsistent local conformity of environmental and clinical isolates it still remains unclear whether the reason for colonisation and infection lays in environmental contamination and / or antibiotic selection in already precolonized patients. Further investigations will employ MRPA genotyping to determine genetic relationships between isolates from patients and environmental isolates. Disclosures No relevant conflicts of interest to declare.

Description: Abstract 327 Introduction: Older patients (≥ 60 years) with acute myeloid leukemia (AML) display a dismal prognosis with only half of the patients reaching a complete remission after intensive induction therapy. The decision whether or not to use intensive chemotherapy is often difficult given the limited number of known markers that might predict the chance to achieve a complete remission. The aim of this study was to establish a risk score for therapy-failure based on clinical variables for older but medically fit AML patients. Patients and Methods: 1379 patients ≥ 60 years of age with AML evaluable for an induction result after treatment with an intensive induction regimen (randomized standard-dose cytarabine containing TAD or high-dose cytarabine (1g/sqm × 6) containing HAM, with a second induction course HAM in case of blast persistence after the first course) within the AMLCG1999 study of the German AML Co-operative Group were evaluated. The following parameters were evaluated for risk prediction in an exploratory analysis: Age, sex, de novo versus secondary leukemia, performance status, body mass index, body temperature, presence of extramedullary manifestations, spleen size, presence of lymph node enlargement, hemoglobin, peripheral blood leukocytes, platelets, blast percentage in peripheral blood and bone marrow, total protein in serum, alanine aminotransferase (ALT), alkaline phosphatase (AP), bilirubin, lactate dehydrogenase (LDH), FAB classification, prothrombin time (PT) and fibrinogen. Since cytogenetics and molecular markers are often not available at the time of decision making, these data were not included. Results: Within this cohort of patients, 744 (54 %) achieved a complete remission. Among the analyzed parameters, age, de novo versus secondary leukaemia, body mass index, body temperature, hemoglobin, peripheral blood leukocytes, platelets, bone marrow and peripheral blood blasts, ALT, AP, LDH, PT and fibrinogen were significantly associated with a CR (p 36°C – 38°C, 〉 38°C), hemoglobin (≤ 10.3 g/dl versus 〉 10.3 g/dl), platelets (≤ 28,000, 〉 28,000 - 53000, 〉 53,000 – 103,000, 〉 103,000 per μl), AP (≤ 89 U/l versus 〉 89 U/l), PT (≤ 75 %, 〉 75 – 87 %, 〉 87 – 98 %, 〉 98 %) and fibrinogen (≤ 150 mg/dl versus 〉 150 mg/dl). Based on these parameters, the predicted remission rates were: minimum, 18.3 %; 1st quartile, 43.9 %; median, 54.4 %;3rd quartile, 64.3 %; maximum: 88.1 %. The observed remission rates were: 1st quarter, 35.7 %; 2nd quarter, 50.3 %; 3rd quarter, 60.2 %; 4th quarter, 69.8 %. Conclusions: Taken together, this risk prediction score based on pre-treatment values predicted the remission probability in patients ≥ 60 years of age with AML receiving an intensive induction therapy. This score may be useful for the determination of the therapy strategy in elderly patients with AML. Disclosures: No relevant conflicts of interest to declare.

Description: Background: Patients with AML who are not eligible for intensive therapy or stem cell transplantation have a dismal prognosis. Autocrine and paracrine secretion of angiogenic and hematopoietic growth factors such as vascular endothelial growth factor in the bone marrow (BM) microenvironment may promote proliferation and survival of leukemic blasts. The oral multikinase inhibitor pazopanib was reported to exert growth inhibitory and proapoptotic effects in myeloid cells. Methods: This phase II study evaluated pazopanib (800 mg orally once daily) in patients with relapsed or refractory AML or at initial diagnosis when no intensive treatment is possible. All patients who received pazopanib for 14 days or longer were included into the analysis of safety, tolerability and efficacy. Response criteria are defined according to the Revised Recommendations of the International Working Group for Diagnosis, Standardization of Response Criteria, Treatment Outcomes, and Reporting Standards for Therapeutic Trials in Acute Myeloid Leukemia. Co-primary endpoints were cumulative response rate (CR, CRp, CRi, PR) within up to one year and reduction of BM microvessel density (MVD) on day 28. Overall survival (OS) and progression free survival (PFS, time from first dose until progression or death from any cause) were measured from the first day of treatment until death of any cause or progression of disease. Results: Between February 2012 and September 2015, 20 AML patients with a median (range) age of 76 (52 - 86) years were treated with pazopanib. The majority of patients (n = 15, 75%) had relapsed (n = 7) or refractory (n = 8) AML, five patients (25%) were enrolled with newly diagnosed AML. Median (range) ECOG performance status was 1 (1 - 3). According to ELN 2010 criteria, four patients (20%) had adverse risk, 15 (75%) had intermediate risk, and one patient (5%) had favorable cytogenetic/molecular risk. Overall, the safety profile of pazopanib was similar to that reported in previous studies. The most common AEs of any grade, related to pazopanib as assessed by the investigator, were gastrointestinal AEs, including nausea (n = 8), diarrhea (n = 6), inappetence (n = 5) and vomiting (n = 3). Two out of 20 treated patients (10%) had a partial remission (reduction of blast count 〉 50%) and 14 (70%) a stable disease (SD) while on pazopanib. Four patients (20%) experienced initial PD. Median PFS was 65 days (95% CI 29 - 105). After the end of study period three remarkable responses occurred on subsequent therapies such as demethylating agents resulting in one CRi and one CRp and one CR after secondary BM transplantation. All these patients had SD while on pazopanib and improved general condition allowing escalation of therapy. However, at the time of OS evaluation all patients had died due to PD and/or infections. Median OS of the treated study cohort was 191 days (95% CI 87 - 435), and 1-year survival altogether was 35%. There was no significant change in BM MVD between day 1 and day 28. Conclusion: Pazopanib was found to be safe in patients with AML not eligible for intensive therapy. The survival data are encouraging but clearly necessitate a controlled randomized clinical trial for confirmation. Clinical trial information: NCT01361334. Disclosures Stelljes: Amgen: Honoraria; JAZZ: Honoraria; MSD: Consultancy; Pfizer: Consultancy, Honoraria, Research Funding; Novartis: Honoraria. Lenz:Roche: Consultancy, Honoraria, Other: Travel, Accomodations, Expenses, Research Funding; Bayer: Consultancy, Honoraria, Research Funding, Speakers Bureau; Novartis: Research Funding; Gilead: Consultancy, Honoraria; Janssen: Consultancy, Honoraria, Other: Travel, Accomodations, Expenses, Research Funding, Speakers Bureau; Celgene Corp.: Consultancy, Honoraria, Other: Travel, Accomodations, Expenses, Research Funding, Speakers Bureau. Brümmendorf:Takeda: Consultancy; Novartis: Consultancy, Research Funding; Pfizer: Consultancy, Research Funding; Janssen: Consultancy; Merck: Consultancy.

Description: Background: Inotuzumab ozogamicin (InO), a humanized anti-CD22 antibody-calicheamicin conjugate, produced a superior response compared with standard of care (SOC) for relapsed/refractory (R/R) acute lymphoblastic leukemia (ALL) in an intent-to-treat (ITT) analysis of the first 218 of 326 patients (pts) randomized (ITT218) in the INO-VATE trial (complete remission [CR], including CR with incomplete hematologic recovery [CRi], 80.7% [95% CI, 72-88] vs 29.4% [21-39]; 1-sided P

Description: Abstract 2773 Introduction: For patients with high-risk myelodysplastic syndromes an epigenetic therapy with hypomethylating agents is considered standard of care. Intensive chemotherapy can be offered to a subset of patients; however, data about the long-term outcome of MDS patients receiving intensive chemotherapy are scarce. Methods: For this evaluation, 104 adult patients with IPSS intermediate-2 or high-risk MDS with at least 10% bone marrow blasts of all age groups treated within the AMLCG1999 trial were included. Patients were randomized upfront to receive 1. double induction therapy with either standard-dose containing TAD - versus high-dose containing HAM–HAM, 2. TAD consolidation therapy followed by either a monthly maintenance therapy for 3 years after achievement of CR or an autologous stem cell transplantation (patients aged ≥ 60 years were all assigned to maintenance therapy), and 3. blast priming with filgastrim starting on day -1 of chemotherapy in selected centers. Results: Fifty-four patients had IPSS Score intermediate-2 and 50 patients were IPSS high risk. Median bone marrow blast count at diagnosis was 15%. The median age was 63.5 years (range: 27–76 years), 39 patients (37.5 %) were female. Median lactate dehydrogenase (LDH) serum level was 296 U/l, median leukocyte count at diagnosis was 5,950 per μl. The cytogenetic risk groups were as follows: favorable 3, intermediate 57, unfavourable 37, missing 7. Among 38 patients with normal karyotype, NPM1/FLT3 mutational status was available for 22 with 5 patients having the combination NPM1 mutated/FLT3 wildtype. Comparison with 2051 patients with de novo AML within the same trial revealed the following significant differences: patients with MDS were older, had a higher male to female ratio, a lower LDH serum level at diagnosis, a lower leukocyte count at diagnosis and were more likely to have adverse cytogenetic risk. Compared to 636 patients with secondary AML after MDS, cytotoxic therapy or irradiation, the cohort of patients with MDS did not display any significant differences except the sex distribution. Patients with MDS displayed a CR rate of 48% (50/104 patients), which was significantly lower than de novo AML patients (67%) and not different to secondary AML patients (47%). Median overall survival in MDS patients was 320 (95% CI: 236 to 505) days with a 2-year and 5-year survival of 33.4% (95% CI: 23.6% to 43.2%) and 22.7% (95% CI: 13.5% to 31.9%), respective, which was significantly (p=0.03) lower than in patients with de novo AML (median 484, 95% CI 435 to 541 days) and comparable to patients with secondary AML (median 282, 95% CI 224 to 311 days, p=0.13). Median relapse-free survival in responding MDS patients was 536 (95% CI: 264 to 1299) days with no significant differences of RFS compared to de novo or secondary AML patients. In multivariate analyses, the diagnosis of MDS remained an independent prognostic factor for CR probability but had no independent influence on survival compared with de novo AML patients. Nine patients proceeded to allogeneic stem cell transplantation in first complete remission of whom six remain in first complete remission between 1354 and 1911 days after achievement of CR. In addition, 16 patients remained in CR for more than one year without allogeneic transplantation. Discussion: Taken together, outcome of patients with intermediate-2 or high-risk MDS after intensive chemotherapy is comparable to the outcome of patients with secondary AML. Adjustment for known risk factors such as age, cytogenetic risk and LDH revealed that inferior outcome of MDS patients compared to patients with de novo AML is attributable to the higher incidence of adverse risk factors. CR-rates appear to be higher compared to hypomethylating therapy and a fraction of MDS patients experiences long-term survival by intensive chemotherapy. Allogeneic transplantation can improve long-term survival for patients achieving remission. Disclosures: Krug: MedA Pharma: Honoraria; Novartis: Honoraria; Alexion: Honoraria; Boehringer Ingelheim: Research Funding; Sunesis: Honoraria. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Schnittger:MLL Munich Leukemia Laboratory: Employment, Equity Ownership. Haferlach:MLL Munich Leukemia Laboratory: Employment, Equity Ownership.

Description: Abstract 1509 Introduction: Both, patients' age and genetic groups are important predictors of outcome in AML while their influence remains poorly quantified and compared. Methods: In the AMLCG (Acute Myeloid Leukemia Cooperative Group) 1999 trial 1470 patients (pts) were 16–59y and 1747 pts were 60–85y of age. 95% of pts could be classified according to genetic risk groups as standardized on behalf of the ELN (Blood 2010;115:453-74). Their treatment was randomized between TAD-HAM versus HAM-HAM induction (TAD, standard dose thioguanine, cytarabine, daunorubicin 60mg/m2 × 3; HAM, high-dose cytarabine 3g/m2 × 6, mitoxantrone 10mg/m2 × 3), TAD consolidation and monthly maintenance vs TAD and autologous SCT, any chemotherapy + vs – G-CSF priming. All assignment was done upfront. Pts of

Description: In the German Multicenter ALL studies (GMALL) patients aged 〉55 years with high risk (B-lineage ALL with WBC at diagnosis 〉30000, late CR, t (4; 11), complex aberrant karyotype or prae-T or mature T-ALL) or very high risk (Ph+/BCR-ABL+) ALL are increasingly candidates for allogeneic stem cell transplantation (allogeneic SCT with a HLA identical sibling donor, MRD or a matched unrelated donor, MUD) or autologous SCT. Here, we report on 31 elderly patients transplanted within the GMALL studies 06/99 and 07/03. Median age of the patients was 61 years (56–65). 22 patients belonged to the very high risk group (VHR), 8 patients to the high risk group (HR) and 1 patient from the standard risk group (SR) was transplanted because of detection of minimal residual disease.17/31 patients were transplanted from a matched unrelated donor, 9/31 patients from a HLA identical sibling donor and 5/31 patients underwent autologous SCT. Conditioning regimens for MRD SCT were myeloablative (MAC) in 6 patients (TBI 12 Gy and chemotherapy n=2, radioimmunotherapy + chemotherapy n=2, chemotherapy only n=2) and 3 patients received reduced intensity conditioning (RIC).Conditioning regimens for MUD SCT changed over time with an increasing number of RIC in the study 07/03. In total, 7/17 patients received MAC (TBI 12 Gy and chemotherapy n=5, chemotherapy only n=2) and 10/17 patients received RIC. Conditioning regimens in autologous SCT were myeloablative (MAC) in 5/5 patients. Results: After allogeneic MRD SCT 4/9 patients (44%) are alive in CCR (d+ 24, d+ 611, d+ 1721, d+ 2321), 3/9 patients died due to leukemia, 2/9 due to transplant related mortality (TRM). After allogeneic MUD SCT 8/17 patients (46%) are alive in CCR (from d+ 165 to d+ 2176). 1 further patient is alive after re- SCT for treatment of relapse. 7/17 patients died due to TRM and 1 patient died due to relapse. After autologous SCT 2/5 patients are alive in CCR (d+ 1703, d+ 1731), 3/5 died due to relapse. Risk factors for TRM: In allo SCT and MAC 8/13 patients died due to TRM in contrast to 1/13 patients with RIC. In auto SCT none of the patients died due to TRM. Risk factors for relapse: In allogeneic MRD SCT 3/9 patients died due to relapse and 2/17 patients relapsed after MUD SCT. Due to the small number of patients, no difference between MAC and RIC could be found. In autologous SCT 3/5 patients died due to relapse. In conclusion: The study shows that allo MRD but also MUD SCT is very effective in a selected population of elderly ALL patients. Since the survival of elderly patients with chemotherapy only is about 25%, more patients should be encouraged to have a MRD or MUD SCT.

Description: In 1999 the German Multicenter Study Group for Adult ALL (GMALL) activated a pilot study (GMALL 06/99). One major aim was to develop a new, shortened and intensified induction regimen based on the following new principles compared to previous GMALL trials: 1) Dexamethasone (DEXA) instead of prednisone to improve antileukemic activity and prophylaxis of CNS relapse 2) prephase with cyclophosphamide (CYCLO) 3) G-CSF parallel to chemo 4) intensified daunorubicin with two 2day cycles (DNR) vs 4 wkly applications 5)1 dose PEG-L-Asparaginase (ASP) instead of 14 d conventional ASP Induction I was followed by GMALL induction phase II as previously reported and a uniform consolidation I. Remission control took place on d24 and d44. Thereafter treatment was risk adapted. Induction I consisted of DEXA, CYCLO and G-CSF. In addition pts received PEG-ASP 1000 U/m2 (d13), vincristin 2 mg (d4,11,18) and DNR 45 mg/m2 (d4+5,11+12). The regimen was modified by 3 amendments which separated the study to 4 pilot phases. The major modifications referred to reduction of DEXA/CYCLO and earlier application of G-CSF. Table 1: Major modifications of induction phase I Drug Pilot 1 Pilot 2 Pilot 3 DEXA 40 mg/m2 (d1–3) 10 mg/m2 (d4–17) 10 mg/m2 (d 1–6,11–16) 10 mg/m2 (d 1–5,11–14) CYCLO 200 mg/m2 (d1–3) none none G-CSF from d13 from d4 from d4 Overall 843 pts were included between 4/99 and 10/03. The median age was 36 (15–65) yrs. Subtypes distribution was c-/pre B 65%, pro B 8%, early T 8%, thymic 14%, mature T 6%. 23% had Ph/BCR-ABL+ ALL. The overall CR rate was 83%, with 12% failure/PR and 7% early death (ED). Significant differences were detected for the pilot phases (p=.0008). The high mortality in pilot I was mainly due to infections. With lower doses of DEXA the rate of ED (p=.0002) and severe infections decreased significantly whereas the failure rate increased slightly. The earlier application of G-CSF contributed to a significant decrease of grade III/IV granulocytopenias and probably also mucositis. Table 2: Results and major toxicities (grade III/IV) of induction therapy Pilot 1 Pilot 2 Pilot3 P Evaluable 103 100 605 CR 76% 83% 82% .0008 PR/Failure 9% 9% 14% ED 16% 8% 5% Survival (3y) 45% 47% 47% 〉.05 Granulopenia 84% 72% 69% .008 Median duration 17d 15d 12d