ALBERT

Description: Introduction: In a recent French FRALLE versus LALA comparison, we have demonstrated a large benefit in outcome when adolescents and young adults were treated in a pediatric rather than an adult ALL protocol (Boissel JCO 2003). Similar provocative results have been observed in three other pediatric versus adult ALL studies (Stock ASH 2000, de Bont Leukemia 2004, Testi ASH 2004). One explanation may be the larger amounts of steroids, vincristine (VCR), and L-asparaginase (L-aspa) administered in pediatric patients. The GRAALL-2003 study was thus designed to offer a pediatric approach in adults with Ph-negative ALL until 60 years of age. Methods: Treatment included a 5-drug induction, high dose-intensity consolidation blocks, delayed intensification, and 2-year maintenance. The comparison with the former LALA-94 protocol showed a 8.6-fold, 3.7-fold, and 16-fold increase in cumulative doses of prednisone, VCR, and L-aspa, respectively. One difference with childhood ALL therapy remained indication of allogeneic SCT in first CR which was offered to all patients with high-risk factor and a donor. In addition, induction was reinforced with a hyper-cyclophosphamide sequence (HyperC) in case of poor early response (cortico- and/or chemoresistant ALL). In the present report, 212 GRAALL-2003 patients with Philadelphia-negative ALL aged 15–55 years with a median follow-up of 18 months were compared to 712 patients previously treated in the LALA-94 trial. Results: Cohorts were comparable in terms of prognostic factors. CR rate was significantly higher in GRAALL patients (93 vs 88%, P=0.02) due to a reduction in resistant disease (0.5 vs 8%, P

Description: A high proportion of Ph-positive Acute Lymphoblastic Leukemia (Ph+ ALL) patients still undergo relapses despite the use of Tyrosine Kinase Inhibitors (TKIs) in addition to chemotherapy as frontline therapy. In this leukemia subtype, the role of BCR-ABL1 kinase domain (KD) mutations as a driver of resistance to TKIs has already been documented by previous studies and such mutations have been reported in up to 80% of the patients at relapse. Next-generation sequencing (NGS) has been proposed to characterize these mutations with a higher sensitivity than Sanger. We report here a prospective study aiming at detecting potentially resistant cell populations by NGS in Ph+ ALL patients enrolled in the GRAAPH 2014-trial. Between March 2016 and February 2019, 156 patients aged 18 to 59 years with newly diagnosed Ph+ and/or BCR-ABL1 positive ALL have been included in the GRAAPH 2014 trial (NCT02619630). BCR-ABL1 isoforms were E1A2 69%, B2A2/B3A2 29%, atypical 2%. After a prephase of steroid, treatment consisted of 4 blocks of chemotherapy + nilotinib. 118 patients (76%) underwent allogeneic or autologous stem cell transplantation (SCT). 22 medullary relapses were recorded within a median time of 9 months (range, 2 - 35). Blood and marrow samples harvested at diagnosis, after each treatment block, before and 3 months after SCT, and at relapse, were sequenced if BCR-ABL1/ABL1 ratio were above 0.001. Mutated BCR-ABL1 transcripts were detected by sequencing the KD of BCR-ABL1 transcripts by NGS with a limit of detection (LOD) of 0.03. T315I allele specific oligonucleotide (ASO) droplet digital RT-PCR (ddRT-PCR) with a LOD of 0.0005 was also performed at diagnosis on a subset of 63 patients, including 5 who have subsequently developed a T315I clone. NGS. At diagnosis, no KD mutation was found by NGS in pretreatment samples of 137 patients. During follow-up (FU), only 12 mutations were found by NGS in 7 out of the 88 patients tested (81, 45, 30, 20, 19, 9 after block 1, 2, 3, 4, before and 3 months after SCT, respectively). Mutations were T315I (N=6), Y253H (N=1), E255K (N=2), E255V (N=1), Q252H (N=1), Y253F (N=1). At relapse, 16 mutations were identified by NGS in 12 patients out of the 17 tested (71%). Mutations were T315I (N=7), Y253H (N=n=3), F359V (N=2), E255K (N=1), E255V (N=1), Q252H (N=1), Y253F (N=1). More than 1 mutated clone were present in 2 patients (E255V+T315I+F359V and Y253H+F359V), and a compound mutation was found in 1 patient (Q252H/Y253F). Out of the 7 patients found mutated during FU, 5 have relapsed with a rapid expansion (1 to 3 months) of the mutated clone. One patient harboring a sub-clonal (10%) E255K at MRD1 has relapsed 9 months later without any detectable mutation. One patient identified with 3 mutated clones (E255K 10%, E255V 10%, T315I 80%) underwent SCT and has not relapsed so far. We failed to anticipate expansion of any mutated clone in the 7 remaining patients found mutated at relapse. T315I ASO ddRT-PCR on diagnostic samples. Low-level T315I mutated BCR-ABL1 transcripts (0.00051 to 0.0013) were detected in 14 out of 63 patients (24%) tested. Only one has expanded a T315I clone later on. In the context of the GRAAPH 2014 trial, 71% of the 17 relapses tested so far were associated with BCR-ABL1 KD mutations. Expansion of the mutated clone could have been characterized before the onset of hematological relapses in only 5 out of 12 patients (42%). Unfortunately in these cases, lags between first detection and relapse were very short (1 to 3 months). On the contrary, occurrences of relapses associated with expansion of KD-mutated clones could not have been anticipated in 58%. All mutations identified, including T315I, F359V, E255K/V and Y253F/H, Q252H/Y253F are known for conferring resistance to nilotinib. NGS is a valuable method for KD mutation detection in Ph+ ALL. It allows a quantitative characterization of KD mutations at relapse. However in our hands and in the context of an intensive therapy combining chemotherapy, nilotinib and SCT, its enhanced sensitivity as compared to Sanger (3% vs 20%) does not translate into the capacity of anticipating expansion of KD-mutated clones. Moreover, in this study, NGS did not detect any mutation in pre-therapeutic samples while T315I mutated BCR-ABL1 transcripts were found at low-level in 24% of these samples by ddRT-PCR. However it should be emphasized that when detected, low-level T315I mutated sub-clones present at diagnosis failed to expand in most instances. Disclosures Cayuela: Incyte: Honoraria, Speakers Bureau; Novartis: Consultancy, Honoraria, Speakers Bureau. Chalandon:Pfizer: Consultancy, Honoraria; Bristol-Myers Squibb: Consultancy, Honoraria; Incyte Biosciences: Consultancy, Honoraria; Novartis: Consultancy, Honoraria. Rousselot:Pfizer: Research Funding. Thomas:PFIZER: Honoraria; ABBVIE: Honoraria; INCYTE: Honoraria; DAICHI: Honoraria. Huguet:Amgen: Honoraria; Servier: Honoraria; Jazz Pharmaceuticals: Honoraria; BMS: Honoraria; Pfizer: Honoraria; Incyte Biosciences: Honoraria; Novartis: Honoraria. Chevallier:Incyte: Consultancy, Honoraria; Jazz Pharmaceuticals: Honoraria; Daiichi Sankyo: Honoraria. Boissel:NOVARTIS: Consultancy. Vey:Novartis: Consultancy, Honoraria; Janssen: Honoraria. Berthon:PFIZER: Other: DISCLOSURE BOARD; JAZZPHARMACEUTICAL: Other: DISCLOSURE BOARD; CELGEN: Other: DISCLOSURE BOARD.

Description: PAX5 is a transcription factor central to B-cell differentiation, frequently mutated in pediatric B-ALL (38.9% in 192 B-ALL, Nature,2007;446:758). PAX5 mutations consist in whole or partial gene deletions or amplifications, fusion gene or point mutations. Adult B-ALL differs in many regards from pediatric B-ALL as demonstrated by differences in prognosis and genetic abnormalities (ETV6-RUNX1 almost absent from adult; BCR-ABL1 rare in children). We screened the occurrence of PAX5 mutations in a series of 119 adult patients with B-lineage ALL prospectively treated in the GRAALL-2003 (pediatric-like protocol for BCR-ABL1 negative patients) or GRAAPH-2003 (imatinib-containing protocol for BCR-ABL1 positive patients), with a median follow-up of 667 days. PAX5 copy number was evaluated using quantitative PCR of each exon performed in hexaplicate (10 exons). PAX5 point mutations of the paired and transactivation domains (exons 2, 3, 7, 8 and 9) were evaluated by sequence analysis. PAX5 rearrangement was evaluated by caryotype and FISH analysis. Globally, PAX5 mutations were identified in 35 cases (30%). Isolated complete deletion of 1 allele was found in 13 cases (11%). Hypomorphic PAX5 alleles were detected in 22 cases (19%) consisting either of partial deletion (13 cases), partial amplification (1), point mutations (7 cases, frequently associated with complete deletion of the remaining allele in 5 cases) or fusion gene (1 PAX5-ELN). In addition, 2 cases of whole PAX5 amplification were found (2%). The remaining 82 cases had two normal PAX5 alleles (68%). BCR-ABL1 fusion gene was significantly associated with PAX5 mutations (Pearson Chi2, p=0.045) occurring in 40% of patients with PAX5 mutations and 18.3% without. PAX5 mutations were rare in the early stages of B-ALL as estimated by the immunological EGIL classification (only 2/35 PAX5 mutant cases [5.7%] vs. 16/82 [19.5%] in normal PAX5 B-ALL cases blocked at the B-I stage). Frequencies of CD10+ and CD20+ blasts were significantly higher in PAX5 mutant cases (Mann-Whitney, p=0.022 and p=0.038, respectively). As PAX5 is essential to the immunoglobulin heavy chain DHJH to VHDHJH transition, we analyzed IGH rearrangement. No difference was detected regarding the frequency of IGH rearrangement (78% in normal vs. 74% in mutants) or the rearranged VH. Survival was not significantly different between the two groups of patients (log rank, p=0.45). In conclusion, PAX5 mutations were a frequent event in adult B-ALL, associated with BCR-ABL1 fusion gene but were not associated with a significant clinical evolution.

Description: Background: In recent years, oncogenic understanding of T-ALL has led to the identification of multiple molecular markers. However, each molecular abnormality accounts for a small proportion of cases and risk stratification at diagnosis still relies on age, clinical presentation, and early response to therapy. NOTCH1 and/or FBXW7 mutations have been recently reported to be a recurrent abnormality in T-ALL, both leading to activation of the NOTCH pathway. Studies in pediatric series have suggested a favorable outcome for NOTCH1 mutated T-ALL, but this has not been evaluated in large series of adult cases. Furthermore, FBXW7 prognostic impact remains unknown in both populations. Methods: In order to evaluate the incidence of these mutations and their prognostic impact in adults, we performed a retrospective analysis of 141 patients (median age, 28 years) with T-ALL treated within the LALA-94 (N=87) or the more recent GRAALL-2003 (N=54) French trials. These patients were representative of the overall population since their outcome did not differ from the overall T-ALL cases treated in either LALA-94 (estimated 3-year OS, 41%) or GRAALL-2003 (estimated 3-year OS, 66%) trial. Furthermore, the patients from the LALA-94 and the GRAALL-2003 trials did not differ with respect to sex ratio, age, WBC, and initial complete remission rate (92% and 98%, respectively). Exons 26 (HD N-terminal), 27 (HD C-terminal), 28 (juxtamembrane domain) and 34 (transactivation domain TAD and the PEST domain) of NOTCH1 and exons 9, 10 and 12 of WD40 domain of FBXW7 were sequenced. Results: We identified 101 cases with NOTCH1 and/or FBXW7 mutations (72%) and 40 wild type (WT) samples (28%). NOTCH1 was mutated in 88 patients (59 HD only, 15 HD+PEST, 9 PEST only, 5 other). FBXW7 was mutated in 34 patients, alone in 13 cases or in association with NOTCH1 mutations in 21 cases. There was no significant correlation between NOTCH1 and/or FBXW7 mutations and immunophenotypic or oncogenetic features. There was a trend for a higher WBC and more frequent CNS involvement in patients demonstrating WT NOTCH1 and FBXW7. Similarly, high-risk MRC/ECOG criteria (age〉35y and/or WBC〉100G/L) were found in 56% of patients from the mutated subgroup versus 73% in the WT subgroup (P=0.06). The prognostic impact of NOTCH1 mutations alone did not reach statistical significance on multivariate analysis (P=0.09). On the other hand, multivariate analysis showed that the GRAALL-2003 trial and the presence of NOTCH1 and/or FBXW7 mutations were the only factors associated with a longer EFS (P=0.001 and 0.035, respectively). Median EFS was 36 months for patients with NOTCH1 and/or FBXW7 mutations versus 17 months for WT patients. Conclusions: These data demonstrate that NOTCH pathway activation by either NOTCH1 or FBXW7 mutation identifies a large group of T-ALL adult patients (72%) with a favorable outcome that could be used for treatment stratification.

Description: Ph+ ALL accounts for approximately one third of ALL cases in patients aged 55 years or older. The median survival of older Ph+ ALL patients is one year, with practically no long-term survivor (Blood, 98, Supp1 p319a, 2001). Imatinib has demonstrated remarkable, although transient, activity in relapsed and refractory Ph+ ALL, which prompted the GRAALL to implement a treatment protocol associating imatinib and chemotherapy in previously untreated elderly patients: ALL patients aged 55 years or older are treated with steroids during one week and Ph+ve cases are then offered a specific therapy including an induction treatment with steroids, cyclophosphamide, daunorubicin and vincristine, followed, irrespective of response to induction chemotherapy, by imatinib, 600 mg daily, combined with intermittent steroids during 2 months. Patients in complete response are then given 10 blocks of alternating chemotherapy, including 2 additional two-month blocks of imatinib, for a total treatment duration of 2 years. Therapy of occult central nervous system leukemia includes 5 intrathecal injections of methotrexate and cranial irradiation. The study is intended to include 30 patients and its main objective is to improve overall one-year survival to 70%. Results are compared with those obtained in 21 Ph+ ALL elderly patients treated according to our previous protocol. Since January 2003, 21 patients aged 58 to 78 years (median: 64.7 years) were included in the AFR09 protocol. Their median follow-up is 3 months. 15/19 patients are in complete response after induction chemotherapy vs 6/21 in the historical controls given similar induction regimen but with no steroids before chemotherapy (p=0.002). The projected overall survival is 95% at 9 months vs 62% in the control group (p=0.08, log-rank test). The 9-month projected event-free survival is 83% vs 10% (p

Description: Patients with T-cell acute lymphoblastic leukemias (T-ALLs) within the Leucémies Aiguës Lymphoblastiques de l'Adulte-94 (LALA-94) prospective trial were treated with a 4-drug per 4-week induction, with intermediate-dose cytarabine and mitoxantrone salvage treatment for patients not achieving complete remission (CR) in 1 course. Only the latter received allografts, if possible, thus providing an informative setting for assessing early response. Representative patients with T-ALL (91 patients) were classified into surface T-cell receptor (TCR)–expressing T-ALL patients (TCRαβ+ or TCRγδ+), pre-αβ T-ALL patients (cTCRβ+, TCR–), and immature (IM) cTCRβ–, TCR– T-ALL patients; 81 patients underwent genotyping for SIL-TAL1, CALM-AF10, HOX11, and HOX11L2. Overall, CR was obtained in 81 (89%) patients; relapse rate was 62% at 4 years and overall survival (OS) rate was 38%. CR rate was significantly lower in IM T-ALL patients after 1 course (45% vs 87%; P 〈 .001) and after salvage (74% vs 97%; P = .002), with the latter inducing a higher rate of CR (9 [64%] of 14) than initial induction. Once CR was obtained, cumulative relapse rates were similar for IM, pre-αβ, and TCR+ T-ALL patients (P = .51), but were higher in HOX11L2 (83%) and SIL-TAL1 (82%) T-ALL patients compared with other genetic subgroups (48%; P = .05). This was associated with an inferior OS for HOX11L2 T-ALLs (13% vs 47% in HOX11L2-T-ALLs; P = .009). The majority of patients with HOX11 T-ALL underwent allografting, predominantly in second CR, but were not associated with a superior OS. Both TCR and genotypic stratification can therefore contribute to risk-adapted management of adult T-ALLs.

Description: Purpose: The use of rituximab, a chimeric monoclonal antibody to CD20, has led to significant improvement in the treatment of B-cell non-Hodgkin's lymphoma and mature B-cell ALL. CD20 is expressed in 30 to 50% of adult BCP-ALL patients. Although some single arm studies suggested that adding rituximab to chemotherapy could improve the outcome of these patients, no randomized study has been reported so far. Methods: To evaluate the potential benefit of adding rituximab, we conducted a multicenter randomized trial comparing the pediatric-inspired GRAALL protocol to the same regimen plus rituximab, in patients aged 18-59 years old with newly diagnosed CD20-positive Ph-negative BCP-ALL enrolled in the GRAALL 2005 trial. CD20 positivity was defined as expression of CD20 in more than 20% of leukemia blasts. Rituximab (375 mg/m2) was given during induction (day 1 and 7), salvage reinduction when needed (day 1 and 7), consolidation blocks (6 infusions), late intensification (day 1 and 7) and first year of maintenance (6 infusions) for a total of 16 to 18 infusions. Allogeneic stem cell transplantation (SCT) was offered in first complete remission (CR) to patients with one or more conventional high-risk criteria and a donor. The primary study objective was event-free survival (EFS). A study sample size of 220 patients was estimated in order to detect a 20% gain in EFS at 2 years (two-sided test, power 85%, type 1 error 5%). A sensitivity analysis was performed after censoring patients allografted in first CR at transplant time. This trial was registered at http://www.clinicaltrials.gov as #NCT00327678. Results: From 2005 to 2014, 220 patients from 56 centers were randomized. Eleven patients had non-eligibility criteria (n=5 Ph+ ALL; n=3 CD20-negative ALL; n=1 HIV infection) or withdrew their consent (n=2) and were accordingly excluded from this modified ITT analysis that dealt with 209 patients (105 in the rituximab arm and 104 in the control arm). Median age was 40 years. Both randomization arms were well balanced for pretreatment characteristics including age, ECOG status, WBC, and central nervous system (CNS) involvement (6% of the whole cohort). After induction ± salvage reinduction, CR rate was 92% and 91% in rituximab and control arm, respectively. In patients who reached CR after first induction and were evaluated for Ig/TCR minimal residual disease level (MRD), the rates of patients with MRD

Description: Patients undergoing treatment for acute lymphoblastic leukemia (ALL) are at risk for thrombosis, in part due to the use of L-asparaginase (L-ASP). Antithrombin (AT) replacement has been suggested to prevent VTE and thus might increase exposure to ASP. We report herein the results of the prophylactic replacement strategy in the pediatric-inspired prospective GRAALL-2005 study. Between 2006 and 2014, 784 adult patients with newly diagnosed Philadelphia-negative ALL were included. The incidence rate of VTE was 16% with 69% of them occurring during induction therapy. Most patients received AT supplementation (87%). After excluding patients who did not receive L-ASP or developed thrombosis before L-ASP, AT supplementation did not have a significant impact on VTE (8% versus 14%, OR: 0.6, p=0.1). Fibrinogen concentrates administration was associated with an increased risk of VTE (17% versus 9%, OR 2.2, p=0.02) whereas transfusion of fresh-frozen plasma had no effect. Heparin prophylaxis was associated with an increased risk of VTE (13% versus 7%, OR 1.9, p=0.04). Prophylactic measures were not associated with an increased risk of grade 3-4 bleeding complications. The rate of VTE recurrence after L-ASP reintroduction was 3% (1/34). In ALL patients receiving L-ASP therapy, the use of fibrinogen concentrates may increase the risk of thrombosis and should be restricted to rare patients with hypofibrinogenemia-induced hemorrhage. Patients developed VTE despite extensive AT supplementation which advocates for additional prophylactic measures. While this large descriptive study was not powered to demonstrate the efficacy of these prophylactic measures, it provides important insight to guide future trial design. NCT00327678.

Description: Background IA is a major cause of morbidity and mortality for patients with hematological malignancies and has been mainly reported in patients with acute myeloblastic leukemia (AML) and stem cell transplantation. Much less is known about the role of IA during the treatment of lymphoproliferative diseases like ALL. Here, we retrospectively evaluated the characteristics of patients and the incidence of IA occurring during the induction course in adult ALL patients enrolled in the GRAALL-2005 trial. Methods We collected the data of 36 patients with IA during induction chemotherapy. All were included between May 2006 and October 2012 in the multicentric GRAALL-2005 phase III trial. According to ALL characteristics, these patients were treated in 3 different substudies: GRAALL, GRAALL-Rituximab (GRAALL-R) for CD20+ ALL, GRAAPH for Philadelphia (Ph) chromosome-positive ALL. IA was defined retrospectively using the EORTC modified criteria (De Pauw, CID 2008). Results Among the 969 patients enrolled, 36 (3.7%) developed IA during induction therapy. The median age was 30 years for all patients receiving induction and 47 years (range, 18 to 59) for patients with IA. We observed 18 IA (3%) by the GRAALL protocol (593 patients enrolled), 16 (8.3%) by the GRAALL-R protocol (191 patients enrolled), and 2 (0.7%) by the GRAAPH protocol (270 patients enrolled). In the GRAALL-R protocol, IA was diagnosed in 6 patients randomly assigned to receive rituximab and in 10 patients treated in the control group. The median time between first day of induction therapy and IA diagnosis was 20 days (range, -2 to 71). The median time between the first day of neutropenia and IA diagnosis was 18 days (range, 0 to 76). At the time of IA diagnosis, 13 patients were hospitalized in laminar airflow rooms, 10 patients in rooms with overpressure, 1 patient in a room with high-efficiency particulate filter, 5 patients in conventional rooms, 1 patient received home care and the type of hospitalization was unknown for the 6 remaining patients. Four patients with IA (11%) had received antifungal drug for invasive fungal infection (IFI) prophylaxis: one patient received fluconazole, one received caspofungin, and two received liposomal amphotericin B. All patients with IA presented pulmonary symptoms associated with a sinusal or cutaneous localization in 1 and 2 patients respectively. The diagnosis of IA was classified as possible in 7 episodes (19 %), probable in 22 episodes (61%), proven in 4 episodes (11%), and indeterminate in the remaining 3 episodes. Detection of Aspergillus antigen in serum by latex agglutination was positive in 24 cases (67%). Chest CT scans were taken for 33 patients (92%): nodules were found in 19 patients, halo sign in 20 patients, and “air crescent sign” in 1 patient. Bronchoalveolar lavage was performed in 11 patients (31%): culture was positive in 7 of them and Aspergillus antigen was positive in 2 of them. Biopsies where positive in 4 cases: in pulmonary biopsy in 3 patients and in cutaneous biopsy in 1 patient. Of all the Aspergillus isolates identified to the species level, Aspergillus fumigatus was isolated from 4 patients and Aspergillus flavus from 1 patient. Overall and IA-attributable 12-week mortality was 8 (0.8%) and 6 (16.7%) patients respectively. Conclusion IA is less frequently observed during induction therapy in adult ALL as compared to adult AML patients. However, the attributable mortality appears to be high in these patients. They are as at risk patients for IA and should be included in prophylactic and empirical antifungal clinical trials. Disclosures: No relevant conflicts of interest to declare.

Description: Background : Bacterial infections (BI) are a major cause of morbidity and mortality in patients treated for hematological malignancies, especially those with acute myeloblastic leukemia or receiving allogeneic hematopoietic stem cell transplantation. Despite severe neutropenia and prolonged treatment with corticosteroids, there are little published data on BI during induction chemotherapy in adults with acute lymphoblastic leukemia (ALL). Between 2006 and 2014, 787 adult patients were included in the GRAALL-2005 study, a prospective, randomized and multicenter phase III trial for patients newly diagnosed patients with Philadelphia chromosome-negative B-cell precursor (BCP) or T-cell ALL. We retrospectively reviewed the occurrence of BI during induction treatment in these patients. Patients and Methods: The GRAALL-2005 study evaluated the value of hyperfractionated cyclophosphamide in the whole study population and of rituximab in patients with CD20+ BCP-ALL. All patients received a 5-drug induction therapy with corticosteroids (prednisone) for 21 days, associated with vincristine, daunorubicin, cyclophosphamide and L-asparaginase. A broad-spectrum antibiotic treatment effective on Gram-negative and positive germs was recommended when the neutrophil count was less than 0.5 G/L. Pneumocystis prophylaxis was made by trimethoprim/sulfamethoxazole or pentamidine. Results: During induction chemotherapy, 270 of the 787 patients (34.3%) experienced a total of 376 BI episodes (1.4 BI episodes per patient). The BI incidence rate was higher in the subgroup of patients combining hyperfractionated cyclophosphamide and rituximab as compared to those who received standard-dose cyclophosphamide and no rituximab (40.7% versus 29.5%; p=0.098). The median time from the first day of induction therapy to BI diagnosis was 10 days (range, 7-14). The infection was considered as serious in 58 patients, with a diagnosis of septic shocks in 57. Forty-one patients were transferred in intensive care unit. At 50 days after induction initiation, 22 patients had died from BI: 8.1% of patients with BI and 2.8% of all patients. Bloodstream was the most common site (82.7%), followed by gastrointestinal tract (6.5%) and lungs (6.5%). In less than 2% of cases, skin and soft tissues, central venous catheter, or urinary tract was concerned. Infections with Gram-positive cocci predominated as the etiology of microbiologically documented infections (46.9%), more specifically coagulase-negative Staphylococci. E. coli and Pseudomonas species were the most common Gram-negative organisms (40.5%). The patients received a median number of 3 antibiotics. The first-line was a monotherapy in 57% of cases, with the predominant use of betalactam. In one-third of the cases, it was betalactam in combination with aminoglycoside or glycopeptide. More than 2 antibiotics were prescribed in 12% of cases. Conclusion: Induction chemotherapy in adults with ALL is associated with a high incidence of bacterial infections and a significant related mortality. To our knowledge, this report is the only large adult ALL study dealing with bacterial infectious complications during induction chemotherapy. Despite an intensely myelosuppressive chemotherapy regimen, the infection-related mortality seems to be lower than that reported during induction in acute myeloid leukemia. Predictive risk factors for bacterial infections have to be analyzed, as well as prophylactic/empirical antibiotic strategies in order to improve care for this subset of patients. Disclosures No relevant conflicts of interest to declare.