ALBERT

Description: Landscape analyses of mutational patterns have shown that virtually all myelodysplastic syndromes (MDS) harbor somatic mutations in 〉80% of cases. These molecular alterations provide useful clonality markers with a potential for early diagnosis of MDS when only cytopenia without marked dysplasia is observed. These markers have been proposed as future prognostic tools to guide therapeutic strategies (Bejar et al., 2011; Itzykson et al, 2013; Mufti et al 2013). Mutational analysis is finally a good way to track disease complexity by deciphering oligoclonality in MDS and better understand clonal evolution. Alterations in the TP53 gene are the most common cause of tumor escape from apoptosis. The aim of this study was to identify TP53 mutations in consecutive samples of lower-risk MDS (IPSS ≤1) with del(5q)obtained at follow-up or progression after sequential classical treatments. Next-generation sequencing (NGS) was used to backtrack the mutant clone(s) identified in late samples. The study was performed both by conventional Sanger sequencing and NGS on a GS Junior Instrument (Roche Applied Science, Mannheim, Germany). For each sample, eight exons (4-11) were amplified from 320 ng of DNA with preconfigured primer plates provided within the IRON II study network. PCR reactions were performed using the FastStart High Fidelity PCR System kit (Roche Applied Science). After double purification with Agencourt AMPure XP beads (Beckman Coulter, Miami, FL), exon-specific amplicon pools were generated and quantified using the Quant-iT™ Broad-Range PicoGreen DNA Assay Kit (Invitrogen, Carlsbad, CA). Emulsion PCR was performed with GS Junior emPCR Reagents (Lib-A) (Roche Applied Science) using 5 x 106 beads at a copy per bead ratio of 0.6. Finally, a fraction of 5-7% enriched beads was loaded on GS Junior Titanium sequencing PicoTiterPlate kit (Roche Applied Science). Data were analyzed for sequence alignment and variant detection using the GS Junior Sequencer and GS Amplicon Variant Analyzer softwares, versions 2.7 and 2.9 (Roche Applied Science). The results were further processed using the Sequence Pilot software version 4.0.1 (JSI Medical Systems, Kippenheim, Germany). The sensitivity of variant detection was set to a lower limit of 〉1% for bidirectional reads. This threshold was chosen according to a recent study investigating the assay's lower limit of detection (Grossmann et al., 2013), thus underlining the strength of NGS to identify subclones at a low frequency, not detectable by conventional Sanger analysis. A total of 89 DNA samples were extracted from the cytogenetics pellets of a cohort of 40 MDS with del(5q). TP53 mutation analysis was performed on 40 initial and 49 follow-up or progression samples including serial samples for 23 subjects. The depth of coverage was at least 500X and up to 8,444X per amplicon. Of those samples obtained and analysed at time of last follow-up or progression, 14 (61%) had TP53 mutations, mostly in the DNA-binding domain. Performing backtracking on previously collected serial samples, TP53 mutations were retrieved by NGS in 43% of initial samples (n=6), which is different from what was previously described by Jädersten et al (2011). A complete scenario of clonal evolution was retrieved in 11 cases, evidenced by TP53 mutations and/or cytogenetics. These were always consecutive to treatment with lenalidomide, yet 6 of the 12 cases without clonal evolution were also consecutive lenalidomide. Figure 1 provides the example of a complete follow-up including nine time points. More correlation with treatment will be provided. Although lenalidomide remains the treatment of choice for MDS with del(5q), resistant subclones may survive and culminate even following therapy initiation. This theory was recently suggested by Landau et al. in CLL (2013) and our test results support this. Early detection of emerging subclones could lead to initiation of alternative treatment, and we thus propose that a monitoring of TP53 alleles is performed annually after the onset of therapy for MDS using NGS. Figure 1. Figure 1. Disclosures: Kohlmann: MLL Munich Leukemia Laboratory: Employment. Moreau:CELGENE: Honoraria, Speakers Bureau; JANSSEN: Honoraria, Speakers Bureau.

Description: Abstract 4981 Background. In the two International Prognostic Scoring Systems of Myelodysplastic Syndrom (MDS), the percentage of BLAST cells in the Bone Marrow is the most important parameter implicated in score: either directly in the IPSS (International Prognostic Scoring System) (1) or indirectly in the WPSS (Who classification-based prognostic scoring system) (2). In these two systems, as in the recommendations of the WHO 2001 (3), the morphological criteria defining blasts cells compared to promyelocyte cells are not specified. In 2005, the IWGM-MDS (International Working Group on Morphology of myelodysplastic syndrome) (4-6) has established morphological criteria defining Blasts cells. PURPOSE OF THE WORK. Our objective in this study is to evaluate the reproducibility, among five observers of our laboratory, of the counting of the marrow blasts of 73 myelodysplasia. Materials and Methods. This study was conducted in several stages. 1st step was to test the implementation of the MDS-Foundation (http://www.mds- foundation.org/virtualmicroscopy) by 5 observers. The 2nd step was to perform correlation test of the selected cells from RAEB (selected and stained in our laboratory conditions), for the 5 observers. Finally Step 3 was to assess the correlation of blast percentage of 73 bone marrow smears from MDS patients, selected due to the presence of an excess of blast cells in the first reading on the bone aspiration. Results. Our results show that the correlation on counting blasts is generally satisfactory (percentage agreement: Test 1 = 86% and test 2 = 94%), while the concordance on the counting blasts of bone marow smears of 73 MDS patients and concordance on the WHO classification seems less satisfactory (agreement 3/5 observers. = 95% but agreed to 4–5/5 observers = 64%). These results can be explained partly by the inter-observer variability and by the variability of some parameters specific to smear marrow (poor quality smears, the staining and/or poverty of smears). Conclusion. The evaluation of the blasts in the MDS must be achieved: (i) at least 500 cells counted (ii) by at least two different observers (iii) by a third observer in discordant case. Despite these recommendations, the assessment of the blasts in myelodysplasia is difficult to achieve in many cases. Disclosures: No relevant conflicts of interest to declare.

Description: Although great progress has been made in the treatment of acute lymphoblastic leukemia (ALL), relapse remains a major issue in the follow-up of these patients. Recent data about the emergence of subclones during haematological malignancies suggest that relapses could result from resistant cells initially in minority or from cells driven to resistance by previous treatments. Among the tools allowing for the characterization of leukemic cells, flow cytometry (FCM) is an essential approach. Increasingly used to evaluate minimal residual disease (MRD) based on the immunophenotypic features of the blasts at diagnosis, it can also allow to identify immunophenotypic shifts related to clonal evolution. Such an approach would be best studied by comparing follow-up samples from the same patients. In order to be thorough, this would however require that conditions as similar as possible are applied to both types of cells. This work was designed 1) to compare the immunophenotypic features of B-ALL blast cells with those of normal hematogones, 2)to assess potential immunophenotypic shifts at relapse 3)to determine the stability of markers not classically used at diagnosis during follow-up and their potential utility for MRD. FCM was performed simultaneously on thawed paired samples from 15 patients (9 children aged between 1 and 12 years old and 6 adults aged between 23 and 71 years old) with B-lineage ALL. With a three-tubes 8 colours panel comprising a backbone of CD45, CD34, CD22 and CD10, the expression of 8 markers was examined and compared to that observed on normal hematogones contained in 29 bone marrow samples from healthy donors. These 8 markers were CD7, CD19, CD20, CD24, CD38, CD58, CD81, CD123. Moreover, an additional four colours panel was used to examine the more recently described antigens CD44, CD200, CD304 and Her2Neu. The presence of leukemia associated immunophenotypes (LAIP) was defined as a difference in mean fluorescence intensity (MFI) between hematogones and blasts of at least 2 standard deviations. At diagnosis, the expression of each marker was at variance from that on hematogones yielding LAIP in all patients, with at least 4 aberrant markers (up to 11). Antigens with the most abnormal expression were CD10 (100%), CD24 (93.3%), CD81 (80%), CD38 (60%) and CD58 (53.3%). Antigens with the least aberrant expression were CD19 (20%), CD22 (20%), CD123 (20%), CD34 and CD20 (46,7%). CD44 which is at a low level on hematogones, was present for 80% of the patients at diagnosis and overexpressed in ALL with MLL rearrangement (3/15 cases). CD200 was overexpressed in 73% of the patients while CD304 was present in only 40% of the patients. A single patient was positive for Her2Neu, which remained present at relapse. All patients retained at relapse the same global immunophenotype without any change in the EGIL classification (3 B-I, 8 B-II, 4 B-III) and the difference with hematogones remained. The expression of most markers was similar at diagnosis and relapse. There was no change at all for the expression of CD38 which therefore appears as the most interesting marker for follow-up and MRD in ALL. Only one patient each showed a change in the expression of CD44, CD58 or CD123. As a whole, stable markers were CD58, CD44, CD200, CD81 and CD24 in contrast with CD19, CD22, CD123, CD304, CD24 and CD20 which changed in 27 to 67% of the patients. Four patients displayed no immunophenotypic change at relapse while 3 showed a modification of a single marker. For 5 patients, with respectively 6 and 7 LAIP, two markers were modified at relapse. Three markers changed for the patient with Her2Neu expression. Finally, only two patients (13%) showed major changes possibly associated with the emergence of a new clone. This study confirms that B-ALL blast cells differ immunophenotypically from hematogones, although the latter have been reported to possibly be their normal counterpart. These data moreover comfort the interest of using LAIP in the detection of MRD in multiparameter FCM. Finally, since molecular targets of therapeutic monoclonal antibodies do not shift sensibly, their use can also be considered at relapse. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction: Peripheral lymphocytosis encountered after myeloablative (MAC) or reduced-intensity conditioning (RIC) allogeneic stem cell transplantation (allo-SCT) is an ill-defined feature. Most reports in the literature deal with large granular lymphocytes (LGL) expansions and only seldom of B-cell increases (Bellucci, Blood, 2002). With an incidence of 3 to 18%, LGL proliferations occur generally late after allo-SCT with a median onset of 9 to 16 months. Such expansions can be polyclonal, oligoclonal or monoclonal, arising from either CD3+ T-cells or CD3- NK cells or both. LGL expansion has been frequently linked to CMV reactivation, indolent clinical course and a usually favorable outcome. Most available data were mainly described in the setting of allo-SCT using bone marrow (BM) or peripheral blood (PBSC) as stem cell source. Here, we report data regarding the incidence and features of lymphocyte expansions after unrelated cord blood (UCB) transplantation. Patients and Methods: Ninety-nine UCB allo-SCT performed in adults between October 2005 and October 2014 were considered for the purpose of this study. Most patients received double CB units (n=94) and a RIC regimen (n=89), for various hematological diseases. Whenever detected, we collected the date of onset and termination of peripheral blood lymphocyte expansions (4x109/L) among the 86 UCB-SCT patients alive at 3 months post-transplant. LGL expansion was defined as sustained LGL above 0.5x109/L and/or 〉40% of LGL in peripheral blood (Zambello, Haematologica, 1998). Concomitant immunophenotypic results, allowed to discriminate expansions of cytotoxic T-cells (CD3+CD8+CD56+), NK-cells (CD3-CD16+/CD56+) and B-cells (CD19+). LGL expansion data were also analyzed with respect to viral reactivation episodes, acute or chronic graft vs host disease, relapse and survival. Results: Lymphocytosis was observed in 21 cases (24%; 10 females and 11 males; median age: 58 y., range: 32-69). Most patients had a myeloid-lineage disease (67%) and were in complete remission at time of UCB-SCT (76%). The median onset of lymphocyte expansion after UCB-SCT was 12.6 months (range, 1.4-49). The median initial lymphocyte count was 4.76x109/L at time of expansion diagnosis. The median duration of expansion was 12 months (range: 1-52). Twenty patients could be further analyzed phenotypically, showing 8 CD8+ T, 1 NK and 1 T-NK LGL expansions. Interestingly, 7 cases of polyclonal B-lymphocytes expansions were also documented while 3 patients presented both T CD8+ and B expansions. Of note, B-cell expansions were CD5+. For 6 patients with T-cell expansion, concomitant DNA from CD3+ sorted cells is available to test clonality. Lymphocyte expansion were from donor origin for 12/14 tested patients. Acute and chronic GVHD developed respectively in 31% and 68% of lymphocytosis patients, and in 57 and 45% of the 65 patients without lymphocyte expansion (P=NS). Comparing these two groups for viral reactivations, the rates were 86% and 76% for HHV-6 (P=NS) and 23% and 39% for EBV (P=NS) respectively. CMV reactivation was significantly more frequent in the group of lymphocytosis patients (76% vs. 29%, P=0.0001). Interestingly, CMV reactivation was significantly higher in the 10 patients of the T or NK group compared to the 7 patients with B cell expansion (100% vs 57%, P=0.05). At time of analysis, 1 patient had relapsed and 4 had died, the causes of death being disease in 1 case and transplant-related mortality in 3. These events were significantly lower than in the group of patients without lymphocytosis (p=0.003 for relapses and p=0.04 for death). Two-year disease-free survival (Fig A) and overall survival (Fig B) were significantly different at respectively 85% vs. 55% (p=0.01) and 85% vs. 63%. (p=0.03). Conclusion: Lymphocyte expansion, at 24%, is not a rare event in adults receiving UCB allo-SCT. These expansions involve equally the T or B-lineages. The latter are often CD5+ suggesting a proliferation of innate B1 cells from the UCB. Lymphocyte expansions are significantly associated with previous reactivation of CMV, but not HHV-6 or EBV. Because these cells were of donor origin, it can be postulated that they represent primo-activation upon encounter with CMV. Finally, both types of lymphocyte expansions are associated with a significant favorable outcome, suggesting a possibly bystander anti-GVL effect. Figure 1. Figure 1. Disclosures Moreau: Celgene, Janssen, Takeda, Novartis, Amgen: Membership on an entity's Board of Directors or advisory committees.

Description: Abstract 4901 In the two International Prognostic Scoring Systems of Myelodysplastic Syndrom (MDS), the percentage of BLAST cells in the Bone Marrow is the most important parameter implicated in score: either directly in the IPSS (International Prognostic Scoring System) (1) or indirectly in the WPSS (Who classification-based prognostic scoring system) (2). In these two systems, as in the recommendations of the WHO 2001 (3), the morphological criteria defining blasts cells compared to promyelocyte cells are not specified. In 2005, the IWGM-MDS (International Working Group on Morphology of myelodysplastic syndrome) (4-6) has established morphological criteria defining Blasts cells. Our objective in this study is to evaluate the reproducibility, among five observers of our laboratory, of the counting of the marrow blasts of 73 myelodysplasia. This study was conducted in several stages. 1st step was to test the implementation of the MDS-Foundation (www.mds-foundation.org/virtualmicroscopy) by 5 observers. The 2nd step was to perform correlation test of the selected cells from RAEB (selected and stained in our laboratory conditions), for the 5 observers. Finally Step 3 was to assess the correlation of blast percentage of 73 bone marrow smears from MDS patients, selected due to the presence of an excess of blast cells in the first reading on the bone aspiration. Our results show that the correlation on counting blasts is generally satisfactory (percentage agreement: Test 1 = 86% and test 2 = 94%), while the concordance on the counting blasts of bone marow smears of 73 MDS patients and concordance on the WHO classification seems less satisfactory (agreement 3/5 observers. = 95% but agreed to 4–5/5 observers = 64%). These results can be explained partly by the inter-observer variability and by the variability of some parameters specific to smear marrow (poor quality smears, the staining and/or poverty of smears). In conclusion, the evaluation of the blasts in the MDS must be achieved: (i) at least 500 cells counted (ii) by at least two different observers (iii) by a third observer in discordant case. Despite these recommendations, the assessment of the blasts in myelodysplasia is difficult to achieve in many cases. Disclosures: No relevant conflicts of interest to declare.

Description: Introduction Next generation sequencing (NGS) has allowed to improve knowledge about the genomic landscape of hematological malignancies. Somatic mutations (SM) are valuable new biomarkers but the utility of incorporating routine sequencing to guide diagnosis and therapeutic decisions remains challenging. We report here an observational multicentric study aimed at assessing the impact of SM testing by NGS in a real-life setting on the diagnosis and treatment of chronic myeloid malignancies (CMM). Patients and Method All patients who benefited from molecular assessment, between 10/2014 and 03/2019 in our University Hospital were included. All provided informed consent for data collection. All NGS requests were validated during a regional multidisciplinary concertation meeting. A custom targeted panel of 34 genes (145kbp i.e. ASXL1,BCOR, BCORL1, CBL, CSF3R, DNMT3A, ETV6, EZH2, GATA2, IDH1, IDH2, JAK2, KDM6A, KIT, KRAS, MPL, NPM1, NRAS, PIGA, PTEN, PTPN11, RAD21, RUNX1, SETBP1, SF3B1, SMC1A, SMC3, SRSF2, STAG2, TET2, TNFAIP3, TP53, U2AF1, ZRSR2) was applied on DNA extracted from peripheral blood or bone marrow samples. DNA libraries, built with the Haloplex® target enrichment protocol (Agilent Technologies, Santa Clara, CA), were paired-end sequenced (150bp reads) with a MiSeq® Instrument (Illumina, San Diego, CA). Data analysis used an in-house pipeline including three variant callings (GATK HaplotypeCaller, VarScan and SAMTools). In a first group (A), NGS indication was to search for clonal hematopoiesis (CH), defined by the presence of at least one SM, in order to confirm or rule out a diagnosis of Idiopathic Cytopenia of Undetermined Significance (ICUS), Clonal Cytopenia of Undetermined Significance (CCUS), myelodysplastic syndrome (MDS), mixed myelodysplastic/myeloproliferative neoplasm (MDS/MPN), aplastic anemia (AA)/hypoplastic myelodysplasia (hMDS) or myeloproliferative neoplasm (MPN), based on recommendations of the WHO classification. In a second group (B), the theranostic impact of SM was studied. Prognostic SMs according to Bejar (2011) were used for MDS and MDS/MPN excluding chronic myelomonocytic leukemia that were analyzed with Itzykson score (2013) and/or CPSS-Mol score (Elena 2016). Prognostic SMs according to Vannucchi (2013) were used for myelofibrosis. Results The median age of the cohort was 60 years old (range: 10-87) with a median follow up of 1.1 years from molecular assessment to last follow-up. Within group A (94 patients), the most frequent blood count anomalies were cytopenia (68%), thrombocytosis (16%), and monocytosis (13%). The karyotype was normal in 77% and failed in 5% of the cases. Non-specific abnormalities (i.e. loss of chr Y, del 20q), were found in 8% of the cases. Before molecular assessment, the diagnoses proposed were ICUS (n=37), suspicion of MDS/MPN (n=16), AA/hMDS (n=16), or MPN (n=25). CH was detected in 31 patients comforting the diagnosis of CMM for 33% of group A (8 CCUS, 3 MDS, 7 MDS/MPN, 6 medullary hypoplasia, 7 MPN) patients. Considering the patients for whom no CH was detected (n=63), the initial suspected diagnosis of CMM was ruled out in 47 patients (i.e. 50% of group A). For the 16 remaining (i.e. 17% of group A), no firm diagnosis could be retained. Within group B (95 patients), NGS identified prognosis SM in 33% of the patients, i.e. poor prognosis SM in 24, including 8/40 MDS, 10/29 MDS/MPN and 6/17 myelofibrosis and good prognosis SM(SF3B1) in 7 of them, respectively 6/40 MDS and 1/29 MDS/MPN. Prognostic SMs had a therapeutic impact in 18/95 pts (19%). Indeed 13 patients with poor prognosis SM had a therapeutic change including 12 allogeneic stem-cell transplantation and 1 hypomethylating agent. Conversely, 5 patients with a good prognosis SM or absence of poor prognosis SM had a de-escalation of treatment intensity. Conclusion The use of NGS in daily practice had a clinical impact in both diagnostic and therapeutic decisions provided that the prescription is made in a critically explored context and not as a systematic test. In this "real life" cohort, the presence or absence of SM was a useful complement for integrated diagnoses in 83% of the patients, allowing to confirm (33%), or exclude (50%) a suspected condition. Moreover, in this cohort 34% of the patients had a SM with a reported prognostic impact and the treatment was modified in 19% of the cases. Yet, it remains necessary to integrate these results with other diagnostic criteria. Disclosures Peterlin: AbbVie Inc: Consultancy; Jazz Pharma: Consultancy; Astellas: Consultancy; Daiichi-Sankyo: Consultancy. Moreau:Janssen: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; AbbVie: Consultancy, Honoraria; Celgene: Consultancy, Honoraria; Amgen: Consultancy, Honoraria. Le Gouill:Celgene: Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Travel support; Roche-Genentech: Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Travel support. Chevallier:Daiichi Sankyo: Honoraria; Incyte: Consultancy, Honoraria; Jazz Pharmaceuticals: Honoraria.

Description: Introduction: Prophylactic T cell depletion with antithymocyte globulin (ATG) remains a standard of care for GVHD prophylaxis during allotransplant (ASCT). Although the optimal ATG dosing strategy is still unknown, recent studies have reported that recipient absolute lymphocyte counts (ALC) at the time of ATG administration may predict survivals in ASCT with unrelated donors, suggesting that the dose (especially at the cut off of

Description: The cytokine Fms-like tyrosine kinase 3 ligand (FL) is a key regulator of hematopoiesis. In a previous Phase 1 study testing a radioimmunotherapy regimen for relapsed/refractory acute lymphoblastic leukemia (ALL), responders showed increased soluble FL serum concentration (sFLc) after salvage regimen (Chevallier, Lancet Haematol., 2015). This prospective monocentric study (ClinicalTrials.gov NCT02693899) aimed to assess the impact of sFLc in ALL and acute myeloid leukemia (AML) patients treated according to standard-of-care intensive first-line chemotherapy regimens. Serum samples were collected on days 1, 8, 15, 22 of induction, at days 1, 8, 15 of each intensive consolidation or day 1 of each non intensive consolidation when appropriate, frozen-stored then tested by ELISA (DY308, R&D Systems, Minneapolis, MN). The following outcomes were considered to assess the impact of sFLc: refractory status after induction (≥5% bone marrow blasts or persistent aplasia 〉45 days), morphologic, immunophenotypic, cytogenetic or molecular relapses, event-free (EFS) and overall survival (OS). All patients provided informed consent. Between May 2016 and January 2018, 80 patients were included. Data were ultimately available for 16 ALL and 62 AML patients. A total of 579 samples were assayed. Analysis of the results disclosed 3 sFLc kinetic profiles during induction i) sustained increase from days 1 to 22 (FLI group), ii) increase from days 1 to 15, then decrease at day 22 (FLD group) and iii) stagnation of low levels all along (

Description: Cereblon (CRBN) was recently identified as a protein binding immunomodulatory drugs (IMiDs) (Ito et al., 2010, Lopez-Girona et al., 2011) essential for the activity of thalidomide, lenalidomide and pomalidomide in patients suffering from multiple myeloma (MM) (Zhu et al., 2011 & 2012). It is known that some patients develop resistance to these drugs, raising the question of a possible role of CRBN alterations (mutations, deletions, loss of transcription...) in the development of such treatment escapes (Broyl et al., 2012, Heintel et al, 2013). Ito et al (2010) showed that loss of the C-ter fragment of CRBN or missense mutations (Y384A and W386A) in the thalidomide binding domain (TBD) indeed impair the efficacy of this drug. We investigated whether the loss of CRBN expression, particularly of full length isoforms including exons 10 and 11 encoding TBD, or missense mutations could be detected by polymerase chain reaction (PCR) followed by fragment-length analysis or Sanger bidirectional sequencing. This method was applied to a cohort of 19 patients issued from a consecutive cohort of 45 elderly patients treated with lenalidomide and dexamethasone for relapsed or refractory MM (Touzeau et al., 2012). RNA was extracted from plasma cells purified by CD138 immunomagnetic sorting (STEM CELLS® beads) at the time of diagnosis. Sanger sequencing performed on RT-PCR products revealed no missense mutations but disclosed a high frequency of alternative splicing of CRBN in samples from MM patients (Lodé et al 2013). We thus developed a new PCR strategy to detect and semi-quantify alternative spliced isoforms of CRBN involving or not a loss of CRBN specific domains, focusing on exons 10 and 11. Two PCR were performed with custom-made primers: PCR-1 was designed to encompass the whole coding sequence of CRBN and PCR-2 was specifically designed to cover TBD. Fragment-length analysis of fluorescent PCR-2 products was obtained by capillary electrophoresis on an Applied Biosystems 3130xl Genetic Analyser (Applied Biosystems, Foster City, CA) allowing for precise sizing and semi-quantification by assessing the height of peaks. This PCR strategy was highly efficient to detect both unspliced CRBN transcripts (PCR-1 and PCR-2 fragments detected at 1562 nucleotides (nt) and 568nt, respectively) and CRBN alternatively spliced variants. The sizing strategy chosen allowed to identify exon skipping. For instance, PCR-2 fragments without exon 10 were visualized as peaks of 234nt, 319nt and 436nt depending on the presence or not of exons 7, 8 and 9. Alternative CRBN splicing in TBD was seen in nearly all patients of this cohort. Examples of representative splicing patterns are shown in figure 1. Semi-quantification of CRBN spliced transcripts showed that 58% of patients had lost more than 50% of CRBN full-length isoform and that 37% had lost more than 50% of exon10. Of note, two patients with a very short duration of response to lenalidomide and dexamethasone (16 days and 58 days respectively) had lost more than 95% of full-length CRBN and more than 90% of exon 10 (see MM6179 in figure1) in initial samples in spite of proper detection of an actin fragment of 858nt.Figure1Figure1. The fragment-length PCR analysis presented here demonstrates that CRBN splicing profiles in MM patients are very heterogeneous. A correlation with response to treatment was not established in this rather small cohort of patients, but this could be due to the fact that diagnosis samples were collected too early in relation to the time of initiating lenalidomide treatment. Indeed, CRBN splicing could be therapy-induced. A prospective study is needed to determine whether the semi-quantification of CRBN alternatively spliced variants would indeed correlate with, or even predict, clinical response to IMiDs. Moreover, further mutation studies are needed to elucidate the heterogeneity of splicing profiles which could reflect the presence of subclonal mutations in genes encoding the splicing machinery. Disclosures: Moreau: CELGENE: Honoraria, Speakers Bureau; JANSSEN: Honoraria, Speakers Bureau.

Description: Background. During the follow-up of treated myeloma patients, the assessment of minimal residual disease (MRD) is gaining an increasing importance. The detection of remaining abnormal plasma-cells (PC) may rely on molecular techniques investigating immunoglobulin rearrangements of the malignant clone or on multiparameter flow cytometry (MFC). The latter allows to obtain a rapid response by dealing with fresh cells. It also focuses on cells still alive, since dead cells are discarded as debris. Numerous publications have reported that the most reliable markers of PC in MFC are CD38 and CD138 their co-expression being a good way to select the population of PC in a bone marrow (BM) or, more rarely tested, blood sample. Malignant PC often but not always differ from normal PC by the loss of CD19 expression and the acquisition of CD56. Other immunophenotypic alterations are related, among others, to the expression of CD20, CD27, CD28, CD33, CD45, CD81 or CD117. Malignant PC also display the monotypic usage of light chains by the myelomatous immunoglobulin, which can readily be assessed in MFC after permeabilization of the PC, although this induces an additional technical step that could induce some cell loss. Here we compared the two panels proposed by the Euroflow consortium (Flores Montero, 2018) which use the same backbone of antibodies with a "surface" strategy associating CD81 and CD117 or a "cytoplasmic" strategy investigating for the expression of kappa and lambda immunoglobulin light chains. Methods. From a cohort of patients for whom MRD had been assessed in our MFC platform, 100 samples were retrospectively selected as displaying detectable MRD in the cytoplasmic strategy. All BM samples had first been submitted to bulk lysis to increase the PC concentration. Between 5 and 10x106 nucleated cells were used for surface staining, premeabilization and intracytoplasmic staining. Another aliquot of the same suspension, with 3 to 5x106 nucleated cells, was used for the "surface" tube. Briefly, both samples were surface stained with antibodies to CD45 (Ozyme), CD19 (Beckman Coulter), CD38 (Cytognos), CD138 (BD Biosciences) and CD27 (Ozyme). The "surface tube" also contained antibodies to CD81 (Clinisciences) and CD117 (BD Biosciences). After this incubation, the "cytoplasmic tube" was submitted to permeabilization (Intrastain® Dako) and cells further incubated with antibodies to kappa and lambda chains (Dako and Clinisciences). All samples were acquired on the same day. Listmodes of the "cytoplasmic tubes" were analyzed and data provided to the clinician within 24 hours. For this study, the listmodes of the "surface tube" were analyzed blindly using the Kaluza® software. Data were then compared to those of the "cytoplasmic tube" Results. A good linear correlation was observed between the two results, with a R2 coefficient of 0.73. The global difference between both tubes was usually a lower MRD level detected with the "cytoplasmic tube", seen in 68% of the cases (median -0.0113; range -0.0001 to -1.4). Of note higher levels (0.0007 to 1.44) were observed in 32%, ruling out a systematic loss of cells that could have been responsible for this difference. The gating strategy adopted (Robillard 2013) delineated four populations on a CD19/CD56 bivariate histogram. Monotypy was then investigated in each of the four subsets thus identified. The same strategy was applied for the "surface tube" looking at the coexpression profile of CD81 and CD117 in each subset. Globally, 58% of the samples were CD56 positive among which 43 were CD19-. CD19 was also absent in 40 CD56- samples. All configurations of CD117 and CD81 coexpression were seen, making each patient a challenging case. In about 10% of the cases, two suspect subsets were seen in the "surface tube" while monotypy was seen in only one in the "cytoplasmic tube". Conclusion. Although this study shows a good correlation between the two panels, it was found that a greater confidence could be attributed to the "cytoplasmic tube", where data are comforted by identification of a monotypic population with the same light chain as the monoclonal peak. Moreover, although confirmation of the abnormal subset was required in numerous cases with the "surface tube", the reverse was never observed. Single use of the "cytoplasmic" combination can thus be recommended as a robust method of MRD assessment in multiple myeloma. Disclosures Moreau: AbbVie: Consultancy, Honoraria; Amgen: Consultancy, Honoraria; Janssen: Consultancy, Honoraria; Takeda: Consultancy, Honoraria; Celgene: Consultancy, Honoraria.