Acute Myeloid Leukemia - Clinical 3 Leukemia (AML/ALL) Aged 16-64 Years at Diagnosis in Estonia (N=407) and in a Well-Defined Region in Western Sweden (N=517)

Stockholm, Sweden, June 13–16, 2013

de novo acute leukemia (AL) patients with acute myeloid or lymphoblastic Acute myeloid leukemia - Clinical 3 leukemia (AML/ALL) aged 16-64 years at diagnosis in Estonia (n=407) and in a well-defined region in western Sweden (n=517). Estonia regained independ - ence in 1991 after 5 decades of occupation by the Soviet Union and entered a P661 transition from planned to market economy. Aims: The current population-based study was designed and aimed to inves - ACUTE MYELOID LEUKEMIA WITH NORMAL KARYOTYPE: IS THE NUM - tigate how changes as regards political and socio-economic circumstances are BER OF CONSOLIDATION CYCLES IMPORTANT? 1,* 1 1 1 1 1 1 likely to affect hematologic health care in a society, especially incidence and D Pereira , S Chacim , L Leite , I C Ferreira , C Moreira , M Nunes , M Brito , survival. M Marques 1, L Viterbo 1, I Oliveira 1, Ne Domingues 1, I Moreira 1, A Santo 1, Methods: The present population-based study comprises all de novo AL A Martins 1, J Mariz 1 patients aged 16-64 years in Estonia and western Sweden diagnosed 1982- 1Servico de Onco-Hematologia, Instituto Portugues de Oncologia Francisco 2006. Estonia covers an area of 45,000 km 2 and the Western Swedish Health Gentil, Porto, Portugal Care Region covers an area of 27,000 km 2. During the years of the population- based retrospective work (1982-1996) the average population in Estonia and To this date, the standard treatment of acute myeloid leukemia Background: western Sweden were 1.5 and 1.6 million inhabitants, respectively. The aver - (AML) in young patients has been the use of increasingly myelotoxic induction age number of inhabitants in Estonia and the western Swedish region during followed by consolidation treatment, in which appropriate doses remains uncer - the two prospective population-based studies (1997-2006) were 1.4 and 1.7 mil - tain. Furthermore, treatment response is heterogeneous, being that the risk of lion, respectively. Identification of de novo AL. Since the aim of the present work relapse remains a constant threat. Defining the best consolidation treatment is was to compare the incidence and survival of de novo AL in the two countries, still a major concern in the treatment of young adults with AML, particularly in the delineation of the diagnosis was critical. Therefore, patients with a history patients with a normal or unfavorable karyotype without a HLA-compatible of pre-existing myelodysplasia, polcythemia vera, essential thrombocythemia, donor. A relevant issue is the ideal number of consolidation cycles with high idiopathic myelofibrosis, chronic myeloid leukemia or leukemia secondary to dose cytarabine (HDAC) that would have an impact on prognosis. chemo-/radiotherapy were excluded from the study. Statistical methods. Indi - Determination of the prognostic impact of the number of consolidation Aims: vidual data for the two cohorts of acute de novo leukemias from western Swe - cycles in younger patients with non-promyelocytic de novo AML and normal den and Estonia were put together in a database for the analyses. The inci - karyotype. dence in the population was compared by use of age standardized incidence In a total of 336 patients with AML, treated in our center from 1998 Methods: rates. Survival analyses were carried out by estimating relative survival. The rel - to 2010, a retrospective analysis was conducted in the subgroup of younger ative survival is the ratio between the observed survival of the patients and the patients (age ≤65 years old according to our protocol) with non-promyelocytic expected survival of a comparable group from the general population. Mortal - de novo AML, normal karyotype who were undergoing the induction regimen ity data of the general population in Sweden and Estonia were used to estimate “7+3” (Idarubicin 12mg/m 2 ev, days 1,2,3; Cytarabine 200mg/m 2 ev, days expected survival rates for the study populations. Internal age standardizing of 1,2,3,4,5,6,7) and didn’t have an HLA compatible donor. The patients’ charac - the relative survival rates was done by use of the age distribution of all individ - teristics were compared with a chi-square test for binary variables and a Mann- uals in the two cohorts. Whitney test for continuous variables. The survival curves were estimated With regard to age adjusted incidence rates for AL in our current work based on Kaplan-Meier curves and data for the various groups were compared Results: there is no statistically significant difference between Estonia (1.8/100000) and with a log-rank test. The multivariate analysis was carried out using a Cox mod - Western Sweden (1.8/100000)during the period 2002-2006. Corresponding fig - el, after the proportional hazard assumption was checked. A p value below ures for AML were 1.4/100000 and 1.2/100000 respectively and for ALL 0,05 was considered as being statistically significant. 0.4/100000 and 0.6 /100000, respectively. Similar incidence rates are noted The median follow-up was 28 months ([1-134]). Of the 47 patients, Results: since the start of our study 1982. Since the first five year cohort 1982-1986, five 48,9% (n=23) were male - with an average age of 52 years old ([20-65]). All year relative survival for all AL patients in Western Sweden has increased from patients (n=47) presented intermediate risk SWOG. In this analysis, of the 20% to 56%, as compared to 2002-2006 (Figure 1). Corresponding figures 78,7% (n=37) patients who reached complete response (CR) after one cycle from Estonia is an increase in survival from 3% to 22%. The pattern is similar of induction, 76,6% (n=36) underwent consolidation treatment with Cytarabine for AML, Estonia (n=280) 2% to 22% and Western Sweden (n=370) 19% to 3mg/m 2 (CALGB). The median number of consolidation cycles was 3 [(1-4)]. 58%.There is also a brisk improvement in survival for Swedish AML patients This difference was conditioned by several factors (intention to treat, toxicity, during 1997-2001 (45% vs. 58%) as compared to Estonia (15% to 22%). relapse during consolidation, death). 59,6% of the patients relapsed after 13 months (median) [0-92]. It was observed that the disease-free survival (DFS) was better in the group of patients who complete 3 or 4 consolidation cycles, in contrast with the group who complete 1 or 2 cycles (P>0,05). If we done a stratification according to the number of consolidation cycles they carried out, 66.7% (n=24) of them underwent more than 2 consolidation cycles. This group of patients presented higher DFS in comparison to patients who underwent ≤2 cycles (median of 41 vs. 9 months, P=0,046), with an impact on overall survival (OS: P=0,021). Given the Cox regression analysis, the completion of more than 2 cycles of consolidation is an independent predictive factor of OS (HR=2,72; 95% CI1,12-6,58; P<0,05), comparatively with the group that performs ≤2 cycles of consolidation, age, number of leukocytes (at diagnosis) and type of response to induction. There was no noted difference between the high/intermediate risk group (n=11) and the low/intermediate risk one (n=4). Summary and Conclusions: This study demonstrates that the number of cycles performed by a particular patient has prognostic value. For the hetero - geneous intermediate risk group with normal karyotype, more cycles of consol - idation with CALGB could be considered an important component to the AML treatment and thereby it could be an alternative to allogeneic transplantation in patients without an HLA-matched donor. This analysis needs to be enhanced with randomized studies in order to evaluate and validate this approach.

P662 IMPROVEMENT, BUT STILL POOR PROGNOSIS FOR ESTONIAN PATIENTS COMPARED TO A WELL-DEFINED REGION OF WESTERN SWEDEN, A POPULATION BASED STUDY OF ACUTE DE NOVO LEUKEMIA PATIENTS AGED 16-64 YEARS E Hulegårdh 1,* , M Punab 2, L Wennström 1, K Palk 3, M Varik 3, E Holmberg 4, D Stockelberg 1 Figure 1. Relative survival* for AML de novo . 1Hematology and Coagulation Section, Department of Medicine, Sahlgrenska University Hospital, Gothenburg, Sweden, 22nd Clinic of Hematology and Summary and Conclusions: Over the years1982-1992 Estonia was still under Oncology, Tartu, 3Department of Hematology, North Estonia medical centre, the mentorship of the Soviet Union. Estonian hematologists did not have access Tallinn, Estonia, 4Department of Oncology, Institute of Clinical Sciences, to therapeutic measures readily available to Swedish hematologists, and the Sahlgrenska Academy, Gothenburg, Sweden results for survival for western Swedish patients with AL far exceeded those for their Estonian counterparts. Of course it is encouraging that survival appears Background: The present population-based survey was carried out over five to have increased in Estonia as well as in Sweden. However, despite progress, consecutive 5-year study periods (1982-2006) on the incidence and survival of the results in Estonia are still burdened by very poor survival in acute leukemia.

haematologica | 2013; 98(s1) | 279 18 th Congress of the European Hematology Association

It is highly unsatisfactory that the improvements in relative survival have not 1Divisions of Hematology, 2Tumor Epidemiology Unit, A.O. Città della Salute been faster. Differences in survival between the two countries are mainly due e della Scienza, Università di Torino, Torino, Italy to higher intensity chemotherapy regimens and a higher rate of hematopetic stem cell transplantation in Sweden. However, other still unknown factors may Background: Acute myeloid leukemia (AML) is the most frequent acute play a substantial role. Therefore, it is of particular interest and great importance leukemia in adults. Prognosis of patients with high risk AML improves with allo - that we are in the process of collecting specified data also regarding support - geneic stem cell transplantation. ive care for de novo ALs for the next prospective period from 2007-2011. Aims: To evaluate outcomes in newly diagnosed patients, younger than 66, who achieved complete remission (CR) after induction/consolidation therapy at the Divisions of Hematology at Città della Salute e della Scienza, Università P663 di Torino, Torino, Italy, between 2000-2011. Methods: Three-hundred and two AML patients (except FAB-M3) were con - PROGNOSTIC IMPACT OF NPM1, IDH1/2 AND DNAH11 GENE MUTA - secutively diagnosed and stratified by risk as follows: low risk included pres - TIONS ON NORMAL KARYOTYPE ACUTE MYELOID LEUKEMIA ence of t(8;21), inv(16)/t(16;16); high risk features included WBC>50.000/uL at PATIENTS NOT HARBORING FLT3/ITD MUTATION 1,* 2 3 4 4 5 6 6 diagnosis, secondary leukemia, presence of extramedullary AML, complex Y Kim , I Lee , D Kim , C Won Jung , J Jang , H Kim , J Moon , S Sohn , karyotype, chromosomal monosomy, no remission after induction, and FLT3/ J Won 7, S Kim 8, J Huh 9, J Lee 1, H Kim 1 1 2 MLL mutations (since 2004). Intermediate risk included patients who did not Hematology, Genome Research Center for Haematopoietic Diseases, Chon - meet either low or high risk criteria. Moreover, the standard risk group includ - nam National University Hwasun Hospital, Jeollanam-do, Korea, Republic of 3 ed low+intermediate risk patients. Patients were treated according to Center Korea, Medical Oncology and Hematology, Princess Margaret Hospital, Uni - guidelines or on clinical trials active at the time of diagnosis. All high risk patients versity of Toronto, Toronto, Canada, 4Hematology-Oncology, Samsung Med - 5 were considered for an allograft since diagnosis. ical Center, Sungkyunkwan University School of Medicine, Hematology, The Results: After induction/consolidation, 229/302 patients (76%) achieved com - Catholic University of Korea, Seoul St Mary’s Hospital, Seoul, 6Hematology- 7 plete remission: 16/229 (7%) were at low, 54/229 (24%) at intermediate, and Oncology, Kyungpook National University Hospital, Daegu, Hematology- 159/229 (69%) at high risk respectively. Eighty/159 (50%) high risk patients Oncology, Soon Chun Hyang University Hospital, Seoul, 8Hematology-Oncol - st 9 received an allograft as 1 line treatment; 56% from a HLA-matched sibling, ogy, Dong-A University College of Medicine, Busan, Laboratory Medicine, 42% from an unrelated donor and 2% received a haplo-identical transplant. Ewha Womans University School of Medicine, Seoul, Korea, Republic of Korea Seventy-nine/159 (50%) did not receive an allograft primarily because of fail - ure to find a suitable donor either sibling or unrelated. At median follow-up of Background: Acute myeloid leukemia with normal karyotype (AML-NK) is 53 months from induction therapy and 49 months from achieving CR, 5-year known to be heterogeneous in the molecular level. Accordingly, it has become overall survival (OS) and 5-year event free survival (EFS) of the entire patient more critical to dissect this group of patients according to their prognosis using cohort were 45% and 35% respectively. By risk category, standard risk patients a molecular genetic technology. showed a 5-year OS of 56% and high risk patients of 40% (P=0.008). Five-year Aims: We attempted to analyze the incidence and prognostic implication of EFS was 37 and 34% in standard and high risk patients respectively (P=0.194). genetic abnormalities on survival in 426 adult patients with AML-NK. High risk patients who underwent an allograft uP-front showed a 5-year OS of Methods: A total of 67 AML-NK patients achieved complete remission (CR), 53% and showed a statistically significant advantage as compared with those candidate mutations in 21 genes were identified by whole exome sequencing who did not receive a transplant (P=0.018). which has 41-89X coverage and by single-nucleotide polymorphism array analy - Summary and Conclusions: Allografting plays a pivotal role in OS and EFS sis using marrow mononuclear cells at diagnosis of AML-NK. Subsequently, for high risk acute myeloid leukemia. The lack of a donor is associated with bad mutation analysis of 11 genes ( i.e. FLT3/ITD, NPM1, DNMT3a, IDH1, IDH2, clinical outcomes. Prospective clinical trials designed to evaluate the use of TET2, NRAS, WT1, DNAH11, SF3B1, and PHF6 ) which are known to be involved more readily available donors such as haploidentical siblings or parents are in the pathogenesis of hematologic diseases, were performed using Sanger needed. sequencing in another subset of 359 AML-NK patients as a validation cohort. Results: Of 426 patients in total (median age: 51, ranges: 15-85), FLT3/ITD, NPM1, and DNMT3a mutations were associated with higher leukocytes counts P665 at presentation of AML-NK. In 284 patients who received standard remission induction (RI) chemotherapy (excluding 119 patients with conservative treat - PROGNOSTIC IMPACT OF A MONOSOMAL KARYOTYPE ON OUTCOME ment and 22 early death/1 follow-up loss after RI chemotherapy), those with IN ACUTE MYELOID LEUKEMIA FLT3/ITD mutation were significantly associated with a higher risk of relapse I Djunic 1,* , N Suvajdzic-Vukovic 1,2 , M Virijevic 1, A Novkovic 3, M Mitrovic 1, (P=0.02), a shorter leukemic-free survival duration (LFS)(P<0.01) or overall N Colovic 1,2 , A Vidovic 1,2 , D Tomin 1,2 survival (OS) (P=0.01). Accordingly, we divided the patients into FLT3/ITD + and 1Clinic for Hematology, Clinical Center of Serbia, 2Faculty of Medicine, Univer - FLT3/ITD - population, and analyzed their treatment outcomes according to the sity of Belgrade, 3Clinical Hospital Center Zemun, Belgrade, Serbia other mutations. In the FLT3/ITD - group (n=200), those with NPM1 mutation showed a higher CR rates after one or two cycles of RI chemotherapy (P<0.01) Background: Recently, a new cytogenetic category of acute myeloid and a longer OS duration (P<0.01), hazard ratio (HR) 0.43, 95% confidence leukemias (AML), monosomal karyotype AML (AML MK+), was reported to be interval (CI) 0.25-0.73, adjusted by other clinical variables including age, leuko - associated with a poor prognosis and to add prognostic information even in cyte counts at diagnosis, and transplantation. In the FLT3/ITD + patients (n=84), patients with a complex karyotype. NPM1 mutation was found to be a favorable prognostic factor showing a low - Aims: The aims of this study were to determine the incidence and character - er relapse rate (P=0.00), a longer LFS duration (P<0.01, HR 0.35, 95% CI istics of AML MK+, and to estimate the impact of MK+ on overall survival (OS), 0.18-0.70), and OS duration (P=0.04, HR 0.55, 95% CI 0.31-0.98) in NPM1 complete remission rate (CR) and disease-free survival (DFS) in patients with mutated patients. In addition, OS was significantly different in favor of those with AML. IDH2, especially R140Q IDH2 mutation, (P=0.04, HR 0.30, 95% CI 0.09-0.99), Methods: This single-center study involved 298 patients with nonpromyelocytic whereas DNAH11 mutation was associated with inferior OS (P<0.01, HR 5.78, AML. The following parameters were recorded: age, hemoglobin level (Hb), white 95% CI 1.65-20.25). Accordingly, we stratified the FLT3/ITD + patients into three blood cell count (WBC), platelet count (Plt) and performance status (PS) evaluat - subgroups according to the NPM1, IDH1/2 and DNAH11 mutation status, ed by the Eastern Cooperative Oncology Group (ECOG), range 0-4 (<1 vs ≥2). Group 1: NPM1 mutation and IDH1 or 2 mutations (n=16), Group 2: isolated The clinical characteristics at presentation and outcome of two subgroups of DNAH11 mutation (n=4) and Group 3: all mutations were negative (n=64). The patients: AML MK+ and AML without MK were compared. MK+ is defined by the group 1 showed significantly better OS than group 2 (P<0.01, HR 16.90, 95% presence of one single autosomal monosomy (AM; excluding isolated loss of X CI 3.48-82.15) orgroup 3 (P<0.01, HR 3.40, 95% CI 1.20-9.55).In a subgroup or Y) in association with at least one additional AM or one structural chromoso - analysis of younger patients less than 60 years of age, similar outcomes were mal abnormality (in the absence of core-binding factor AML). Patients were treat - also observed in favor of group 1 in terms of OS. ed according to the Medical Research Council (MRC) 12 protocol. Statistical analy - Summary and Conclusions: Our study confirmed that NPM1 mutation is an sis included: the Fisher exact test, chi-square test and Kaplan-Meier method. independent prognostic factor in adult patients with AML-NK not harboring Results: The mean age of the patients was 57 years (range 19-79 years). The FLT3/ITD mutation. In addition, several other genetic markers were identified incidence of AML MK+ was 32.8% (98 pts). AML MK+ patients were significant - as prognostic including IDH1/2 or DNAH11 mutations as well as NPM1 muta - ly older than those from the control group (AML without MK; P=0.048). The tion in a subgroup of AML-NK patients with FLT3/ITD mutation. presence of severe anemia at presentation with Hb <80 g/L was more frequent in AML MK+ patients than in those with AML without MK (P=0.046). Likewise, the mean WBC was lower in AML MK+ patients than in participants with AML P664 without MK (P=0.008). AML MK+ patients had a higher rate of poor ECOG PS (≥2) than AML patients without MK (P=0.013). AML MK+ patients had shorter ROLE OF ALLOGRAFTING IN HIGH RISK ACUTE MYELOID LEUKEMIA OS (6 vs 18 months; P=0.002) and DFS (5 vs 12 months; P=0.013) in compar - B Bruno 1,* , E Audisio 1, S D’Ardia 1, D Ferrero 1, E Maffini 1, L Giaccone 1, ison with AML patients without MK. Similarly, AML MK+ patients had a lower B Allione 1, A Busca 1, C Casa 1, F Testa 1, C Frairia 1, F Ferrando 1, M Festuc - CR rate compared to AML patients without MK (27.3% vs 59.5%; P<0.001). cia 1, L Brunello 1, A Evangelista 2, M Boccadoro 1, U Vitolo 1, G Ciccone 2, M Fal - Summary and Conclusions: This research identified MK+ in approximately da 1, F Marmont 1 one-third of AML patients. The characteristics of these patients are older age,

280 | haematologica | 2013; 98(s1) Stockholm, Sweden, June 13–16, 2013 poorer ECOG PS, lower Hb level and WBC at presentation of disease. The Results: Four studies involving 44 patients were included. Individual patient study also confirmed a poor prognosis for these patients, with shorter OS and data were obtained. The Kaplan-Meier estimate of median overall survival was DFS and lower CR rate. These data indicate that this subset of AML patients 13 months (95%CI 5-21). The median overall survival for patients undergoing should be recognized as an adverse risk group and according to this studied chemotherapy and HSCT was 28 months (95% CI, 7 to 48) while the median in the new context of molecular markers. overall survival for patients locally treated was 9 months (95% CI, 4 to 13). The difference in overall survival was not signiﬁcant but potentially represents an absolute beneﬁt of 19 months (95% CI 11 to 28). The difference was found in P666 OS between patients treated with HSCT and patients who did not underwent to HSCT in population below 40 years old (P=0,031). The patients undergoing RISK FACTORS FOR CENTRAL NERVOUS SYSTEM INVOLVEMENT AT HSCT had a significantly longer overall survival time compared to those who DIAGNOSIS IN PATIENTS WITH ACUTE MYELOID LEUKEMIA did not [33 months (95%CI 22-44) vs 9 months (95%CI 5-13)].There was no I Djunic 1,* , N Suvajdzic-Vukovic 1,2 , N Kraguljac-Kurtovic 1, M Virijevic 1, 3 1 1,2 1,2 1,2 consistent evidence of a difference in effect according to age, sex, stage, site, A Novkovic , M Mitrovic , N Colovic , A Vidovic , D Tomin grade, histology, extent of resection, tumour size or exposure to radiotherapy. 1Clinic for Hematology, Clinical Center of Serbia, 2Faculty of Medicine, Univer - 3 Summary and Conclusions: While more individual patient data are needed sity of Belgrade, Clinical Hospital Center Zemun, Belgrade, Serbia to collect, chemotherapy with HSCT could be considered as the optimal ther - apy for patients with isolated MS. Background: Central nervous system (CNS) involvement in acute myeloid leukemia (AML) is observed in 2-4% of patients at diagnosis. Hematological parameters associated with an increased risk of CNS leukemia include a high P668 presenting white blood count (WBC), elevated lactate dehydrogenase (LDH), French-American-British (FAB) subgroups M4 and M5 and age less than 50 EVALUATION OF BONE MARROW ON DAY 5 OF INDUCTION THERAPY years. Recently, it was shown that the immunophenotypic pattern of AML–the FOR ACUTE MYELOID LEUKEMIA FOR IDENTIFICATION OF CHEMO- expression of CD56 antigen (CD56+) on leukemic cells, was associated with a RESISTANT BLAST SUBPOPULATIONS higher rate of CNS involvement in AML at presentation. Y Ofran 1,2,* , S Yaari 1, R Hoffman 1, N Horowitz 1, M Hayun 1 Aims: The aims of this investigation were to estimate the incidence of CNS involve - 1Hematology and Bone marrow tarnsplantation, Rambam health care campus, ment at diagnosis of AML and to determine the risk factors for its occurrence. 2Faculty of medicine, Technion - Israel institute of technology, Haifa, Israel Methods: This single-center study involved 191 patients with nonpromyelocytic AML. The following parameters were estimated as risk factors for CNS involve - Background: In most patients with acute myeloid leukemia (AML), different ment at diagnosis of AML: age, WBC (<30 ¥10 9/L vs ≥30 ¥10 9/L), serum lactate leukemia-associated immunophenotypes (LAIP) co-exist. At diagnosis, rapid - dehydrogenase (LDH) concentration >1.5 ¥ the upper limit of normal, expression ly proliferating clones determine the dominant LAIP. However, at relapse, a dif - of CD56 antigen on leukemic blasts (<20% vs ≥20%) and cytogenetic risk group ferent LAIP may become dominant. This phenomenon may reflect either a new (assessed according to European LeukemiaNet recommendations). The pres - clone or differentiation change within the original clone evolved over time. Alter - ence of leukemic cells in cerebrospinal fluid (CSF) detected by cytomorphologi - natively, a small resistant sub-clone masked at diagnosis, may become evident cal and/or flow cytometric analysis of CSF is considered as CNS involvement. following eradication of other, chemo-sensitive clones. Patients were treated by the Medical Research Council (MRC) 12 scheme. Risk Aims: The present study investigated differences in response kinetics among factors were identified using univariate and multivariate analysis. immunophenotypically distinguished cell subpopulations at day 5 of induction Results: The mean age of the patients was 55 years (range 21-79). CNS involve - in addition to routine assessment of day 14. ment was recorded in 5.8% (11) of them. Univariate analysis detected the follow - Methods: Fifteen consecutive adult AML patients receiving induction therapy ing significant risk factors for CNS involvement: age <55 years (P=0.009), WBC with daunorubicin 60mg/m 2 for three days and cytarabine 100mg/m 2 for sev - ≥30 ¥10 9/L (P=0.001), elevated LDH (P=0.036), FAB M4/M5 (P=0.05) and CD56+ en days (3+7) were evaluated. Immunophenotypes of bone marrow samples (P=0.001). Multivariate analysis identified CD56+ as the most important risk fac - at diagnosis (day 1), after five and fourteen days of therapy were compared. tor for CNS involvement at diagnosis in AML patients (P=0.001; relative risk Samples of relapsed patients were also analyzed. LAIPs were quantitatively (RR)=18.241; 95% confidental interval (CI)=3.305-100.675. The cytogenetic risk evaluated by 6-color FACS analysis (Cytec; FlowJosoftware) and their relative group had no influence on CNS involvement in this study. fraction of total blast population was recorded at each time point. Summary and Conclusions: According to the result s obtained, AML patients Results: During therapy, the number of detectable LAIPs per patient was found with CD56+ on leukemic cells are at high risk for CNS involvement at presen - to decrease. The median number of detectable LAIPs reduced from 5 through tation. Cytomorphological and/or flow cytometric analysis of CSF should be 3 to 2 at diagnosis, day 5 and day14, respectively. However, in the majority of recommended as a routine procedure in these patients for more precise diag - patients, only scanty leukemic cells were observed in the day 14 marrow and nosis of CNS involvement. CD56 antigen is an isoform of the neural cell adhe - not a single LAIP was detected in 2/12 (16%) patients whose samples were sion molecule, constitutively expressed in normal monocytes and monocyte- evaluable. In most patients, the cell number in the blast window decreased by derived cells, which could be an explanation for the high risk of CNS involve - day 5. However, eradication rate varied among different leukemic subpopula - ment in CD56+ patients with AML, particularly FAB subtypes M4/M5. tions. Heterogeneity of leukemic cells decreased with therapy, especially in patients who achieved CR in response to treatment, reflecting different sensi - tivity of various blast subpopulations to chemotherapy. LAIPs were quantitative - P667 ly evaluated as percentage of cells in the blast window. Evaluation of each patient was particularly focused on leukemic subpopulations with the slowest SYSTEMIC VERSUS LOCALIZED THERAPEUTIC APPROACH FOR ISO - eradication rate. A specific LAIP, which doubled its size in the blast window, LATED MYELOID SARCOMA: META-ANALYSIS OF INDIVIDUAL PATIENT from diagnosis to day 5 was defined as a personalized LAIP of concern. Per - DATA 1,* 2 sonalized LAIPs of concern were identified in 7/15 (46%) patients and their D Antic , N Milic prevalence was as high as 71.5% (5/7) among the individuals who achieved 1Clinical Center Serbia, Clinic for hematology. Medical Faculty, University in 2 complete remission (CR) following a single induction course. Two of respond - Belgrade, Institute for medical statistic and informatic, Medical faculty, Univer - ing patients experienced early relapses. Interestingly, the dominant LAIP at sity in Belgrade, Belgrade, Serbia relapse differed from that recorded at diagnosis in both patients. Meticulous analysis of samples obtained at diagnosis revealed that the dominant LAIPs at Background: Myeloid sarcoma (MS), also known as granulocytic/monoblas - relapse were identified at day 5 samples in both patients as the LAIP of con - tic sarcoma, extramedullary myeloid tumors, or chloroma, is a tumor mass com - cern. prising blasts or immature cells of the myeloid series, occurring at an anatom - Summary and Conclusions: In clinical practice, there is an urgent need for ical site other than the bone marrow. MS can occur in patients with active acute an approach allowing recognition and characterization of leukemia relapse at myeloid leukemia (AML), in patients with chronic myeloproliferative disease, as the earliest possible stage. Herein, a novel approach to early, real-time identi - the first manifestation of relapse in previously treated patients and as well as fication of patient-specific resistant blast subpopulation is suggested. Detection isolated MS in patients without bone marrow infiltration. The optimal timing and of a personalized unique “LAIP of concern” as early as day 5 of induction would treatment of isolated MS are not clear, but delayed or inadequately treated iso - allow sorting these subpopulations out and characterizing their metabolic and lated MS will almost always progress to AML. Also, in patients treated with genetic factors which could be attributable to later appearing clinically signifi - chemotherapy postremission therapy has not been adequately studied; there - cant chemo-resistance. fore, one of the most important questions in the treatment of MS is the role of hematopoietic stem cell transplantation (HSCT). Aims: The objective of this review was to assess the effects of systemic ther - P669 apy approach in adults with isolated MS rather than local treatment (surgery or radiotherapy). LOW MDR1 AND BAALC EXPRESSION IDENTIFIES A NEW SUBGROUP Methods: We searched the Cochrane Controlled Trials Register, UKCCCR OF ADULT CYTOGENETICALLY NORMAL ACUTE MYELOID LEUKEMIA Register of Cancer Trials, Physicians Data Query, EMBASE, MEDLINE and WITH A FAVORABLE OUTCOME CancerLit and after that studies in adults with isolated MS treated with X Guo 1,* , F Chen 1, P Shi 1, J Zha 1, B Liu 1, H Zheng 1, B Xu 1 chemotherapy and/or HSCT compared with local treatment (surgery or radio - 1Department of Hematology, Nanfang Hospital, Southern Medical University, therapy) were included. GuangZhou, China

haematologica | 2013; 98(s1) | 281 18 th Congress of the European Hematology Association

Background: Cytogenetically normal acute myeloid leukemia (AML) is a het - P671 erogeneous disease, in terms of genetic/molecular abnormalities resulting into marked differences in outcome. IMPACT OF DAY 14 BONE MARROW EXAMINATION ON RE-INDUCTION Aims: We demonstrated that multidrug resistance 1( MDR1 ), brain and acute DECISION AND PREDICTION OF COMPLETE RESPONSE IN ACUTE leukemia, cytoplasmic ( BAALC ) gene expression was of prognostic signifi - MYELOGENOUS LEUKEMIA (AML) A Aleem 1,* , F Anjum 2, K Alsaleh 2, F Algahtani 2, A Momen 2 cance and high MDR1 expression correlated with a high BAALC expression 1 2 (r=0.487,P< 0.001) in cytogenetically normal AML in the prophase study then Department of Medicine, Division of Hematology/Oncology, Medicine, Hema - we hypothesized that MDR1 and BAALC expression together would better tology/Oncology, King Khalid University Hospital, King Saud University, Riyadh, identify the patient’s risk profile. Saudi Arabia Methods: Pretreatment bone marrow samples from 92 cytogenetically normal AML patients were analyzed for MDR1 and BAALC mRNA expression by real- Background: The decision to re-induce patients with acute myeloid leukemia time reverse transcriptase polymerase chain reaction. Patients were divided (AML) based on results of the day 14 bone marrow (BM) examination is vari - into low MDR1 and BAALC expression group and combined group(high MDR1 able and has limited evidence based data. In adult AML a variety of clinical and and/or high BAALC expression) according to MDR1 and BAALC levels and biological parameters have been examined for their potential value in predict - were compared for clinical outcome. ing treatment response. Early response to induction therapy could be an impor - Results: 73 cases of 92 cytogenetically normal AML patients got CR after the tant prognostic factor in this disease. first block with a CR rate being 79.3%. However, 29 cases of the CR patients Aims: The aim of this study was to evaluate the accuracy of a day 14 BM relapsed with the relapsed rate being 39.7%. In contrast, Patients with low examination in determining the need for re-induction chemotherapy and pre - expression of both MDR1 and BAALC had a higher CR rate ( 93.3%vs72.6% , diction of complete response. P= 0.021), lower relapse rate( 7.1% vs. 42.5%, P= 0.000), longer OS ( 50.3% vs Methods: Eighty-four patients with newly diagnosed AML treated with standard 17.8%,P=0.00 1) than high MDR1 and/or high BAALC expression (combined induction chemotherapy were retrospectively reviewed for the purpose of eval - group) in cytogenecally normal AML. Results showed no statistical difference uating treatment decisions based on their day 14 BM biopsy from 2000-2012. in CR rate ( 93.3%vs85.7% , P= 0.341),relapse rate ( 7.1% vs. 8.8%, P= 0.000) Response to therapy in this analysis was based on morphology alone. and OS ( 50.3% vs 63.1%,P=0.43 1) for cytogenetically normal patients with Results: Of the 84 patients undergoing standard induction, 65 patients (77%) had both MDR1 and BAALC low expression comparing to those with low-risk cyto - hypocellular bone marrow with less than 5% blast on their day 14 examination. Thir - genetically abnormal. teen patients (16%) had definitive residual disease (RD), and 6 patient’s (7%) were Summary and Conclusions: The combined assessment of BAALC and MDR1 classified as indeterminate response (IR). Sixty three out of 65 patients with hypocel - expression can improve treatment stratification in adult cytogenetically normal lular bone marrow on day14 had complete response on day 28 bone marrow exam - AML. Low expression of both MDR1 and BAALC identifies cytogenetically nor - ination and 2 patients had residual disease on day 28 bone marrow and these 2 mal AML patients with a favorable long-term outcome. patients received re-induction on day 14. Patients with suboptimal response (SOR) (SOR=IR+RD), 2/13 patients with RD underwent re-induction chemotherapy on day 14. Seventeen patients with SOR (6 with IR and 11 with RD) were observed until P670 count recovery without any re-induction therapy. All 6 patients with IR and 8 out of 11 patients with residual disease who were observed without re-induction eventu - THE SURFACE MOLECULAR SIGNATURE OF LEUKEMIC CELLS WAS ally attained a morphologic complete remission (CR) on day 28. ASSOCIATED WITH NPM1 AND FLT3-ITD MUTATIONS IN ACUTE Summary and Conclusions: Day 14 BM examination may have suboptimal MYELOID LEUKEMIA sensitivity for the detection of residual leukemia in AML patients. Some patients S Gao 1, L Su 1,* , W Li 1, Y Tan 1, J Cui 1 with an indeterminate response and residual disease on day 14 may not require 1Cancer center, The first hospital, Jilin University, Changchun, China re-induction chemotherapy, but instead, may benefit from careful observation until count recovery to avoid the morbidity and mortality associated with pro - Background: Acute myeloid leukemia (AML) is a genetically heterogeneous longed bone marrow suppression related to early re-induction chemotherapy. clonal disorder characterized by the accumulation of somatically acquired genetic alterations in hematopoietic progenitor cells. Certain molecular muta - tions are associated with signs of cell morphology and differentiation in AML. P672 However, only limited data is available to analyze such correlations in detail. Here we report the analysis results of 259 consecutive Chinese AML patients ADVERSE PROGNOSTIC SIGNIFICANCE OF FLT3 MUTATIONS IN ACUTE by using J-Express 2009 analysis suite. PROMYELOCYTIC LEUKEMIA 1,* 1,2 3 1,2 1 1 Aims: To observe the associations between surface molecular signature of M Mitrovic , N Suvajdzic , N Tosic , A Bogdanovic , I Djunic , A Sretenovic , leukemic cells and NPM1 and FLT3-ITD mutations in patients with de novo A Vidovic 1,2 , M Virijevic 1, N Colovic 1,2 , T Kostic 3, T Karan-Djurasovic 3, acute myeloid leukemia. I Glumac 3, V Spasovski 3, S Pavlovic 3, I Elezovic 1,2 , D Tomin 1,2 1 2 Methods: In present study 259 patients with de novo AML were enrolled. The Clinic of Hematology, Clinical Center of Serbia, Faculty of Medicine, Univer - morphological differentiation status of the leukemia cells was evaluated by sity of Belgrade, 3Institute of Molecular Genetics and Genetic Engineering, FAB classiﬁcation. The mutational status of molecular markers NPM1 and Belgrade, Serbia FLT3-ITD were analyzed by polymerase chain reactions (PCR). Surface differ - entiation markers (CD11c, CD13, CD14, CD15, CD33, CD34 and HLA-DR) of Background: Internal tandem duplication (ITD) of Fms-like tyrosine kinase-3 the leukemic cells were analyzed by flow cytometry. Patients gave informed (FLT3 ) gene and point mutation in its activation loop ( FLT3 -D835) are frequent consent prior to enrolment in the study. The obtained data were analyzed using aberrations in acute promyelocytic leukemia (APL), but their prognostic value J-Express 2009 analysis suite and SPSS 17.0. is not well established. Results: The frequencies of NPM1 and FLT3-ITD mutations were 15.06% Aims: To evaluate prognostic value of FLT3 gene mutations in de novo APL (39/259) and 14.51% (37/255) in this study, respectively. Based on the surface patients. differentiation markers of leukemic cells, the patients were classified into 2 dis - Methods: FLT3- ITD and FLT3- D835 mutations were assessed by PCR and tinct subsets: Cluster I (CD34 low ) and Cluster II (CD34 high ). Patients in Cluster I PCR-RFLP method respectively in 45/61 newly diagnosed APL patients, on could be further subdivided into Cluster Ia (CD34 low HLA-DR low ) and Cluster Ib DNA isolated fromdiagnostic bone marrow aspirates or smears. The presence (CD34 low HLA-DR high ). Similarly, Cluster II patients could be classified into two of t(15;17)(q22;q12) was assested using cytogenetics or RT-PCR method. All subsets: Cluster IIa (CD34 high CD11c high ) and Cluster IIb (CD34 high CD11c low ). of the patients (median age 44 years, range 19-78; female/male 32/29; medi - The frequencies of NPM1 and FLT3-ITD mutations were significantly higher in an follow-up 32 months) were managed at the Clinic of Hematology from 2004 Cluster I patients [33.33% (35/105) and 31.43% (33/105)] than those in Cluster to 2012 with all-trans retinoic acid and anthracycline-based chemotherapy. II patients [2.01% (3/149) and 2.01% (3/149)] ( χ2=47.491, P< 0.001 and Results: Material for FLT3 gene mutations analysis was available in 45/61 (75%) χ2=45.642, P< 0.001, respectively). Neither mutation was detected in the Clus - patients. FLT3 -ITD was detected in 11/45 (24.4%) and FLT3 - D835 in 4/45 (8.9%) ter IIb patients. No associations were observed among NPM1 (P= 0.531) and patients. FLT3 -ITD was associated with significantly higher WBC count (median FLT3-ITD (P= 0.563) mutations and monocytic morphology in patients without 28.4 ¥10 9/L vs. 1.4 ¥10 9/L, P=0.0022), higer percentage of peripheral blood blasts CD34 expression. (median 19% vs. 4%, P=0.042), lower platelet count (median 21 ¥10 9/L vs. ¥ 9 Summary and Conclusions: NPM1 and FLT3-ITD mutations are associated 32 10 /L, P=0.029) higher relapse-risk score (low risk 1/11 (9%), intermediate 3/11 with signs of cell morphology and differentiation in AML. The absence of NPM1 (27%), high 7/11 (64%) vs. 12/34 (35%), 10/32 (30%), 12/34 (35%) respectively, mutations in the immature subset Cluster IIb (CD34 +CD11c -) patients in this P=0.034), more frequent microgranular subtype (3/11 (27%) vs. 2/34 (5.9%) study supports the hypothesis that the cell origin of the NPM1 mutations is a P=0.0001), CD2 expression (5/9 (55.5%) vs. 5/30 (16.7%) P=0.0007), CD56 common myeloid progenitor. Further studies are required to identify similar expression (3/9 (33%) vs. 1/12 (8%), P=0.047) and thrombotic events (4/11, (36%) possible correlations for FLT3-ITD mutations. vs. 5/34, (15%), P=0.025). No correlation was found with age, fibrinogen, prothrom - bin time, activated prothrombin time, D-dimer, International Society of Thrombo - sis and Haemostasis Scoring System for diseminated intravascular coagulation (ISTH DIC score), expression of the other surface markers, additional cytogenet - ic abnormalities and incidence of differentiation syndrome. In contrast, FLT3 -D835 was not significantly associated with any hematologic characteristic. Early death

282 | haematologica | 2013; 98(s1) Stockholm, Sweden, June 13–16, 2013

(ED) rate in our series was 8/45 (17.78%). ED rate was significantly higher Results: Out of the 45 patients, 39 (86%) showed WT1 overexpression at diag - (P=0.002) in patients with FLT3 -ITD (4/11, 36%) than in patients without mutation nosis. No correlation was observed between WT1 overexpression and clinical (4/34, 11.7%). Furthermore, multivariate analysis identified WBC count >20×10 9/L characteristics such as age, gender, FAB classification and cytogenetic risk at (P= 0.002337), ISTH score ≥6 (P=0.008) and FLT3-ITD (P= 0.02) as independent diagnosis. In 1 patient no LAIP could be identified and 8 more patients showed prognostic factors for ED. Seven relapses (7/37, 18.9%, 5 overt and 2 molecular) suboptimal LAIPs for follow-up. The post-induction and post-consolidation occured, three (3/7, 42%) in FLT3 -ITD positive cohort and four (4/30, 13%) in analysis showed MRD by both WT1 and MFC in 3 and 4 patients, respective - FLT3 -ITD negative cohort. Median time from the achievement of complete remis - ly (Table 1b). No patients were positive by WT1 and negative by MFC either sion to relaps was 30 months (range: 4–44). FLT3 -ITD positive patients had sig - post-induction or post-consolidation. Finally, 9 patients post-induction and 4 nificantly lower 4-year RFS than FLT3 -ITD negative ones (56% vs. 77%; P=0.002). post-consolidation showed positive MRD by MFC without WT1 overexpres - However, multivariate analysis identified only WBC count >20×10 9/L as an inde - sion. Among these patients 2 showed suboptimal LAIPs for follow-up. The lim - pendent predictive factor for relapse (P=0.05). Besides, 4-year overall survival ited number of patients in the present series as well as the heterogeneity of (OS) was significantly lower in FLT3-ITD positive patients (62% vs. 87%, P=0.016) post-consolidation subsequent treatments does not allow testing the clinical too. In contrast FLT3- D835 mutation had no impact on both 4-year RFS and OS. utility of MRD results. Summary and Conclusions: Our study speaks in favor of negative prognos - Summary and Conclusions: Overexpression of WT1 in the present series of tic value of FLT3 -ITD mutation in APL. Further prospective trials should 1) AML patients is similar to that reported in the literature. In most cases, MRD investigate independence of FLT3-ITD predictive value and 2) evaluate whether positivity by WT1 overexpression correlates with that of MFC. Although clinical this parameter might be included in risk stratification. value of these results could not be tested in the present study, WT1 expression would be especially useful in patients without other molecular markers and those with suboptimal or without LAIP. Further analysis with a larger cohort of P673 patients is needed to determine the true prognostic value of this MRD approach. MINIMAL RESIDUAL DISEASE BY WT1 GENE EXPRESSION IN PATIENTS WITH ACUTE MYELOID LEUKEMIA AFTER INDUCTION AND CONSOLIDA - P674 TION CHEMOTHERAPY: CORRELATION WITH MULTIPARAMETER FLOW CYTOMETRY SERUM FERRITIN AS A PROGNOSTIC MARKER IN YOUNG ADULT A Franco 1,* , M Kwon 1, C Martinez-Laperche 1, P Balsalobre 1, P Font 1, ACUTE MYELOID LEUKEMIA 1,* 1 1 1 1 1 C Encinas 1, A Escudero 1, J Gayoso 1, J Anguita 1, J Diez-Martin 1, I Buño 1 M Delia , D Pastore , P Carluccio , F Albano , A Ricco , T Perrone , 1 1 1 1 1 1HEMATOLOGIA Y HEMOTERAPIA, HOSPITAL UNIVERSITARIO GREGO - P Curci , F Gaudio , A Giordano , A Russo Rossi , G Specchia 1 RIO MARAÑON, Madrid, Spain Hematology - University of Bari, Bari, Italy

Background: Evaluation of minimal residual disease (MRD) in early stages of Background: Different studies have demonstrated an iron overload contribu - treatment can refine the risk stratification of patients with acute myeloid tion to post-transplantation liver toxicity, infectious events and poor survival in leukemia (AML). patients undergoing hematopoietic stem cell transplantation for haematologi - Aims: The objective of this study was to analyze the MRD detection by of the cal malignancies. So far, in the clinical setting of adult acute myeloid leukaemia Wilms tumor (WT1) gene expression after induction and consolidation (AML) there is no evidence of the possible role of iron in response and survival chemotherapy in patients with AML and to correlate the results with those rates. obtained by multiparametric flow cytometry (MFC). Aims: Our aim was to study the role as a prognostic factor of pre-treatment Methods: Between 2007 and 2012, 65 patients were diagnosis of AML in our insti - serum ferritin in young adult AML. tution. Twenty patients were excluded due to earlydeath, lack of intensive treatment, Methods: The study sample included 77 consecutive adult de novo AML promyelocyticleukemiadiagnosis or absence of available data. Therefore, 45 AML patients (38 males and 39 females, median age 55 years, range 16 to 60 yrs). patients were included in the analysis (Table 1a). WT1 expression was quantified The serum ferritin level was determined at onset of the disease in each case. in bone marrow (BM) samples at diagnosis, after induction and after consolidation According to the FAB criteria the subtypes were: 7 M0, 43 M2, 16 M4, 6 M5, 5 therapy, using real-time quantitative PCR (relative quantification with K562 cell line M6. M3 subtypes were excluded from the analysis. NPM, FLT3 and cytogenet - as calibrator and GUS as reference gene). The cutoff value for overexpression ic evaluation was performed for all cases. The NPM mutation was present in was set at 0.55% in BM. The presence of subpopulations associated with leukemia 35 patients (45%) and 34 (44%) harboured the FLT3 alteration. Twenty-five aberrant immunophenotype (LAIP) was analyzed by MFC with 4 colors (Beckman (32%) patients were in the adverse, ten (13%) in the favorable and forty-two Coulter FC500). Residual disease was reported as positive above 0.1%. (55%) in the intermediate molecular-cytogenetic risk group. The patients were subdivided into two groups according to serum ferritin values (less than 800 ver - sus greater than 800 ng/mL). Student t-test or the Mann-Whitney test was per - formed for comparisons of means. Two-tailed Fisher exact test exact test was Table 1. used to compare categories. Overall survival (OS) was measured from the time of diagnosis to death or last follow-up visit and was calculated using the Kaplan- Meier method; the log-rank test was used to compare survival curves. Logistic regression was performed for multivariate analysis. Only p values less than 0.05 were considered statistically significant. Results: Thirty (39%) patients showed a ferritin serum value greater than 800 ng/mL. Compared with the less than 800 ng/mL group, patients with serum fer - ritin grater than 800 ng/mL were more frequently non responders to chemother - apy (29 vs 58%, P=0.03) and they had a shorter median OS (250 vs 665 days, P=0.04). Moreover, patients with serum ferritin greater than 800 ng/mL showed a higher frequency of documented infections during induction treatment (48% vs 16%, P=0.004). At multivariate analysis, NK-FLT3ITD+,NK-NPM, Cytogenet - ic and Ferritin value (less than or greater than 800 ng/mL) all showed a statis - tical correlation with the response rate (P=0.04; P=0.05; P=0.04; P=0.05, respectively). Summary and Conclusions: The results of our study suggest a link between serum ferritin and AML prognosis. Further studies are needed to confirm the utility of serum ferritin as a prognostic marker in the adult acute myeloid leukaemia setting.

P675 PERIPHERAL BLOOD BLASTS ON DAY 7 OF INITIAL INDUCTION THER - APY PREDICT RESPONSE TO CHEMOTHERAPY AND SURVIVAL IN PATIENTS WITH ACUTE MYELOID LEUKEMIA S Gao 1, L Su 1,* , W Li 1, Y Tan 1, X Liu 1, P Yu 1 1Cancer center, The first hospital, Jilin University, Changchun, China

Background: Cytogenetics and molecular markers are powerful prognostic indi - cators for outcome of patients with acute myeloid leukemia (AML). However, pre - dictors of response after initiation of chemotherapy are not available. A rapid decrease in blood blasts is often considered a favorable indicator of response in patients with AML who receive induction chemotherapy. Time of peripheral blood

haematologica | 2013; 98(s1) | 283 18 th Congress of the European Hematology Association blast (PBB) clearance has been reported to be a predictor for complete remis - sion (CR) and survival of AML patients, but litter attention was paid to the corre - lation between response, outcome of AML and PBB percentage. Aims: To explore the associations between chemotherapy response, outcome of AML and PBB percentage on day 7 of the initial induction therapy. Methods: In this study, 46 consecutive de novo AML (excluding acute promye - locytic leukemia) patients treated with standard ‘3+7’ induction course were evaluable for bone marrow (BM) response on day 14 and day 28. Flow cytom - etry was carried out to identify the percentage of cells with leukemia-associat - ed aberrant immuno-phenotype (LAIP) in each patient from the initial BM aspi - rate. LAIP-positive blast percentage was determined on peripheral blood on day 7 of the induction therapy. Patients gave informed consent prior to enrolment in the study. Statistical analysis was performed with the software SPSS 17.0. Results: PBB percentage on day 7 (PBBPD7) was 0.35% (0.0%, 80.74%) in all patients, and in those achieved CR, the percentage was 0.03% (0.0%, 0.45%), which was significantly lower than those did not achieve CR (10.85% [1.13%, 19.38%]) ( u= - 3.92, P< 0.001). PBBPD7 was significantly correlated with CR rate of the first course ( r= 0.584, P< 0.001). The cutoff value was 0.945% (area under curve [AUC]=0.84, sensitivity [Se]=78.9%, specificity [Sp]=81.5%, P< 0.001). The CR rate of patients with PBBPD7 ≤0.945% (84.62%, 22/26) was significantly higher than those with PBBPD7 >0.945% (25.0%, 5/20) (χ2= 16.571, P< 0.001). PBBPD7 was significantly correlated with overall survival (OS) ( r= - 0.437, P= 0.003) and relapse free survival (RFS) ( r= - 0.388, P= 0.007). The cutoff value was 0.43% for both OS (AUC=0.723, Se=73.4%, SP=71.4%, P=0.009) and RFS (AUC=0.723, Se=66.7%, SP=69.2%, P=0.009). The OS and RFS were significantly higher in patients with PBBPD7 ≤0.43% compared with those with PBBPD7 >0.43% ( P=0.002 and P=0.005, respectively). PBBPD7 was proved to be an independent prognostic parameter in multivariate analy - ses for both OS ( P=0.036) and RFS ( P=0.035). Summary and Conclusions: Peripheral blood blast percentage on day 7 of the induction therapy may be an important predictor of early response to initial chemotherapy and long-term survival for AML.

P676 PROGNOSTIC VALUE OF TH17 CELLS IN ACUTE LEUKEMIA M Salah-Eldin 1,* , N Abousamra 2, R Helal 3 1Medical Oncology, Oncology Center, Mansoura University, 2Department of Clinical Pathology, Hematology Unit, 3Public Health and Community Medicine, Faculty of Medicine, Mansoura, Egypt

Background: The net outcome of the battle between the anti-tumor and the tumor promoting immune cells and their associated cytokines within the tumor environment will determine the ultimate fate of the affected tumors. Th17 cells and their effector cytokines have emerged as important mediators in inﬂamma - tory and autoimmune diseases and serve as an ambitious field in current immunology research. Recent studies suggest a potential impact of Th17 cells on solid tumors but relatively little is known about their contribution in hemato - logical malignancies. Aims: The current study was designed to investigate the possible involvement and clinical significance of circulating Th17 cells in acute leukemia Methods: Flow cytometry was used to analyze percentages of Th17 cells in peripheral blood from 93 acute leukemia patients and 40 controls Results: Compared with healthy controls; acute leukemia patients had a high - er proportion of circulating Th17 cells. Furthermore, increased serum concen - trations of IL-17 and IL-21 accompanied the increased Th17 cell levels in these patients. Circulating Th17 cells were reduced when patients achieved complete remission (CR) after chemotherapy, suggesting that circulating Th17 cells can be used to evaluate response to therapy in acute leukemia patients. Notably, the increased prevalence ofinitial Th17 cells was significantly associated with probability of achieving CR as well as with longer overall survival of acute leukemia. Summary and Conclusions: These results strongly suggest that Th17 cells may be a powerful new prognostic determinant which could serve as a poten - tial therapeutic target to modulate anti-tumor response in acute leukemia. Mechanisms by which Th17 cells influence acute leukmia, and whether these mechanisms mediate direct or indirect effects need to be determined.

284 | haematologica | 2013; 98(s1)