ALBERT

Description: Introduction: Grey zone lymphoma (GZL), a B-cell lymphoma with features intermediate between large B-cell lymphoma (LBCL) and classical Hodgkin lymphoma (cHL), is a rare and poorly defined entity. To decipher its mutational landscape and discover new therapeutic targets, we performed exome sequencing of 31 GZL cases. Methods: GZL cases from the LYSA group (N=139) and BC Cancer (N=30) were centrally reviewed and classified as previously published (Sarkozy et al, Am J Surg Pathol 2019). Whole-exome sequencing was performed on 31 cases with available fresh frozen tissue, using laser micro-dissection (LMD, MMI technology) to enrich for tumor cells and obtain matching normal DNA from microenvironment cells. DNA was extracted (Agencourt® DNAdvance kit) and genomic libraries were constructed with the Ovation ultra-low kit (Nugen®). Exome capture was performed using Agilent SureSelectXT V6+UTR followed by paired-end sequencing (NextSeq®). Somatic nucleotide variants (SNVs) and indels were identified using VarScan, Strelka and Mutect. Parameters affecting the sensitivity and specificity of variant calling were optimized using 7 "gold standard" cases for which DNA from peripheral blood cells was additionally available. Possible oncogenic drivers were identified based on rate of recurrence, MutSigCV and literature review. Results: Among the 31 GZL cases, the median age was 41 y (14-83) with a sex ratio of 15M:16F; 21 cases had mediastinal involvement, including 15 within the thymic area; EBER in-situ hybridization (ISH) was positive in 8 cases. Seven (23%) cases were classified as group-0 (cHL morphology with 100% CD20 expression), 22 (71%) with an intermediate morphology as group-1 (N=9, cHL-like morphology) or group-2 (N=13, LBCL-like morphology) and 2 (6%) as group-3 (LBCL with 100% of CD30 expression). The mean coverage was 96X (42-203) for tumor samples. One case was excluded due to failure in the LMD process. Among the 30 cases, 6628 variants across 4826 genes were found, including 2903 coding mutations (325 indels and 2808 SNVs, mean of 104/sample, range: 15-678), 721 affecting the 5' UTR and 2774 the 3' UTR. A total of 152 genes were identified as being potential oncogenic drivers, with a mean of 11 mutated genes per case (range 2-36). The most recurrently mutated genes were SOCS1 (33%), B2M (23%), GNA13 (20%), LRRN3 (17%), and ZNF217, NCOR1, ITPKB, IRF2BP2, CSF2RB, and CSMD3 (13% each). The epigenetic SWI/SNF and transcription regulation pathway (including NCOR1/2, ARID1A, KMT2D, KMT2A) was affected in 73% of the cases, JAK/STAT in 70% and NF-kB in 19%. As assessed by CNVkit and GISTIC, the most recurrent gains/amplifications identified were in 9p24.1 (JAK2, CD274, PDCD2LG2; 69%) and 2p16.1 (REL, BCL11A; 62%), and losses in 11q14.3 (ATM; 48%) and 12q24.33 (NCOR2; 48%). Based on mutational signature analysis, individual base substitutions were linked to mutagenic processes, with the highest contributions associated with aging (29%) and defective DNA mismatch repair (27%); moreover, mutations attributable to AID/APOBEC activity (5%), were found to be significantly enriched in EBV- vs. EBV+ cases (p = 0.013). EBV+ cases had fewer total variants (mean 98 vs 258, p=0.08) and potential oncogenic variants (mean 7 vs 15, p=0.03) compared to EBV- cases. EBV+ cases also lacked mutations in the NF-kB pathway and MHC-class I components (B2M and HLA-B: 0% vs 43% in EBV-, p=0.06) but had mutations in STAT3, DHX58, ACTB and ATP13A4 (6/7 cases) not present in the 23 EBV- cases. LRRN3 and GNA13 mutations were significantly associated with thymic area involvement (40% vs 0%, p=0.01). Furthermore, fluorescence-ISH indicated that 20% (1/5) of EBV+ cases had a rearrangement in the CIITA locus (16p13.13) vs 53% (9/17) in EBV- cases. Patients with an intermediate morphology had more oncogenic variants than those in group 0 and 3 (mean of 15 vs 6 variants/case, p=0.01 affecting 12 vs 5 genes, p=0.004). Finally, NCOR1 (N=4) and NCOR2 (N=2) mutations were exclusively found in cases with intermediate morphology (23% vs 0% for those with group 0 or 3 morphology). Conclusion: These data suggest that GZL is a highly heterogenous disease harboring somatic driver events shared with PMBCL and HL. We also discovered novel gene mutations pointing to the importance of previously unrecognized pathways in the pathogenesis of GZL. The distinct mutational pattern in EBV+ GZL suggests divergent evolutionary trajectories. Disclosures Sarkozy: Takeda: Research Funding. Salles:Merck: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees; Novartis, Servier, AbbVie, Karyopharm, Kite, MorphoSys: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Educational events; Autolus: Consultancy, Membership on an entity's Board of Directors or advisory committees; Takeda: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Educational events; Epizyme: Consultancy, Honoraria; BMS: Honoraria; Amgen: Honoraria, Other: Educational events; Roche, Janssen, Gilead, Celgene: Consultancy, Honoraria, Membership on an entity's Board of Directors or advisory committees, Other: Educational events. Savage:BMS, Merck, Novartis, Verastem, Abbvie, Servier, and Seattle Genetics: Consultancy, Honoraria; Seattle Genetics, Inc.: Consultancy, Honoraria, Research Funding. Scott:Celgene: Consultancy; Roche/Genentech: Research Funding; Janssen: Consultancy, Research Funding; NanoString: Patents & Royalties: Named inventor on a patent licensed to NanoSting [Institution], Research Funding. Steidl:Juno Therapeutics: Consultancy; Tioma: Research Funding; Roche: Consultancy; Bristol-Myers Squibb: Research Funding; Nanostring: Patents & Royalties: Filed patent on behalf of BC Cancer; Seattle Genetics: Consultancy; Bayer: Consultancy.

Description: Diffuse large B-cell lymphomas (DLBCLs) are heterogeneous in clinical presentation, histopathological and biological findings, and outcome. Today, risk-adapted treatment decisions are mainly based on the International Prognostic Index (IPI). Recently, molecular profiling has been shown to correlate with survival in DLBCL patients treated with conventional chemotherapy. Immunohistochemistry (IHC) has been proposed as a surrogate of molecular profiling (« phenotypic profiling»). Besides, we and others have shown that early response evaluation (after 2 cycles) with 18FDG-PET scanning is predictive of patient outcome in DLBCL, independently of IPI (Haioun & al, Blood2005; 106: 1376). We retrospectively investigated the phenotypic profile of 81 patients (pts) from our recently published series with a confirmed diagnosis of DLBCL and available material for extensive IHC analyses. Diagnosis was based on a nodal (n=45), extranodal (n=25) or mediastinal (n=11) specimen. IHC was performed with the markers bcl2, CD10, bcl6 and MUM1, and scored by two observers. Pts were classified as having a germinal center (GC) or non germinal center (nGC) profile using the following algorithm: GC pts were to be CD10+ or bcl6+ and MUM1−, and all others were nGC (Hans & al, Blood2004; 103: 275). Bcl2 was positive in 53%, CD10 in 36%, bcl6 in 58% and MUM1 in 45% of interpretable cases. Seventy-four pts (91%) had an interpretable profile, 39 (52%) were in the GC group, 35 (48%) in the nGC group. All pts received a doxorubicin-containing regimen as induction treatment, with rituximab for 46% of them. Pretreatment characteristics and early response by 18FDG-PET according to phenotypic profile are shown in the table below. With a median follow-up of 33 months, estimated 2-y OS and EFS were 75% and 67%. Survival analysis confirmed the poor prognostic value of a positive early 18FDG-PET scan: 2-y EFS was 46% in the PET-positive group and 80% in the PET-negative group (p = 0.0003). Two-year EFS was 61% and 73% in the bcl2-positive and negative groups, respectively (p = 0.08). We could not observe any prognostic influence of the GC vs nGC profile: Two-year EFS was 72% in the GC and 64% in the nGC group (p=0.62). In this limited series, we confirm the high predictive value of early response evaluation with 18FDG-PET scanning, but did not observe any significative influence of phenotypic profile, contrarily to several published reports. The reasons for this discrepancy, be they related to treatment type (including rituximab), disease presentation (many pts with extranodal disease) or IHC evaluation (fixation, scoring), remain to be understood. Meanwhile, 18FDG-PET should be an early guide for first-line strategies in DLBCL. Characteristics according to phenotypic profile GC (n=39) nGC (n=35) p Age

Description: Abstract 2632 Hans algorithm using immunohistochemistry correlates well with gene expression data in Diffuse large B-cell lymphomas (DLBCL) (Meyer PN, 2011) and has demonstrated in some studies clear survival differences in favor of germinal-centre (GC) vs non-germinal centre (n-GC) B-cell among DLBCL treated with R-CHOP. We undertook an immunohistochemical study among patients aged 18 to 59 years with aaIPI 1 included in the GELA trial LNH 03-2B that compared R-ACVBP intensified immunochemotherapy to standard R-CHOP. This trial demonstrated an improvement of EFS, PFS and overall survival (OS) of patients treated with R-ACVBP (C Recher et al, in press). Our goal was to evaluate survival of patients with GC and n-GC DLBCL according to treatment regimens. We analyzed by immunohistochemistry the expression of CD10, BCL6 and MUM1 and classified patients as GC or n-GC according to the Hans algorithm. Among the 380 patients enrolled in this study, 229 patients were available for Hans algorithm classification. There was no differences considering clinical characteristics of these 229 patients (age, sex, B symptoms, PS, Stage, LDH, number of extranodal sites, bulky mass, bone marrow involvement) compared to the whole LNH03-2B population. 175 DLBCL cases were present on a tissue microarray (TMA) and 54 other cases were analyzed using unstained slides. 101 patients were classified as GC and 128 as n-GC. 107 patients were treated by R-ACVBP and 122 by R-CHOP. EFS, PFS and OS were not different between the GC and n-GC profile among the whole population (P=.82, P=.90, P=.68, respectively). There was no statistical difference in EFS, PFS and OS between R-ACVBP and R-CHOP in GC patients (P=.78; P=.84, P=.33, respectively). Interestingly, EFS, PFS and OS were significantly much longer among n-GC patients treated by R-ACVBP compared to R-CHOP (P=.02; P=.007, P=.007, respectively). Results were similar considering only TMA population (P=.02, P=.001, P=.001, respectively). This subgroup analysis suggests that the survival benefit related to R-ACVBP over R-CHOP in the LNH 03-2B is in large part linked to a survival improvement in the n-GC population. This algorithm, easy to apply on routine paraffin-embedded tissue, might be useful in the future to select patients that can primarily benefit from this intensive regimen. Disclosures: No relevant conflicts of interest to declare.

Description: Diffuse large B-cell lymphomas (DLBCLs) are markedly heterogeneous, and there is increasing evidence that the biological features of these tumors vary according to the primary site of disease (lymph node or various extranodal organs). The Waldeyer’s ring (WR), a circular band of mucosa-associated lymphoid tissue composed of the palatine, lingual and pharyngeal tonsils, is the second most common site of extranodal invovlement by DLBCL. The presence of focal follicular features in tonsillar DLBCLs has been previously reported. In this study we included 209 adult patients with de novo DLBCL presenting in the WR consecutively included in the GELA protocols during a 10-year period (1993–2004). All patients (M/F ratio: 1,77/1, mean age 59 years) comprising 81% with stages I–II and 19% with stages III–IV disease, received intent-to-cure anthracyclin-based polychemotherapy. Morphology, pattern of growth, immunophenotype and differentiation profile were analyzed and correlated to the clinical features. Survival and outcome were analyzed in comparison to a matched cohort of patients with primary nodal DLBCL. By morphology, 55% of tonsillar cases were classified as centroblastic, 39% as centroblastic-polymorphous, 3% as immunoblastic, and 3% were unclassifiable. Among large biopsy specimens (n=79), 18% had a prominent nodular pattern consistent with transformed follicular lymphomas, 35% had a minor nodular component (

Description: Introduction Diffuse Large B Cell Lymphoma (DLBCL) is the most common lymphoid malignancy, accounting for 30-40% of all Non Hodgkin Lymphomas. Gene expression profiling (GEP) has identified three main subtypes of DLBCL: Germinal Center B-cell like (GCB), Activated B-Cell like (ABC) and Primary Mediastinal B-cell Lymphoma (PMBL). Recently, Next Generation Sequencing (NGS) has enabled a more detailed characterization of DLBCL mutational profiles. Conventional techniques such as immunohistochemistry (IHC) and FISH are also widely used to describe DLBCL. However, no study has yet performed an integrative analysis of the mutational, gene expression, IHC and FISH profiles of DLBCL, in order to provide a comprehensive view of this disease. Methods 215 patients with de novo DLBCL in the prospective, multicenter and randomized LNH-03B clinical trials led by the LYmphoma Study Association (LYSA) were included in this study. Microarray-based GEP identified 81 ABC, 83 GCB, 18 PMBL and 33 other. Mutational profiles of patients' tumor DNA were established using Lymphopanel NGS, designed to identify mutations in 34 genes important for lymphomagenesis. For each recurrently mutated gene, we applied ROMER (Ritchie, Nucleic Acids Res, 2015) to perform gene set enrichment analysis on differential expression profiles of mutant and wild-type patients, using a multifactorial model accounting for subtype. The gene sets were obtained from the MSIGDB Hallmarks (Subramanian A, PNAS, 2005) and Signaturedb (Schaffer, Immunol Rev, 2006) collections. When possible, IHC was performed for IgM (n=150), MYC (n=140), BCL2 (n=148), BCL6 (n=146), CD10 (n=152), FOXP1 (n=147) and MUM1 (n=152); FISH was performed for MYC (n=131), BCL2 (n=133) and BCL6 (n=131). Results As expected, EZH2 mutations were significantly associated with upregulation of GCB gene expression (p

Description: The molecular heterogeneity of diffuse large B-cell lymphoma (DLBCL) has been highlighted by gene expression profiling (GEP), dividing DLBCL into the following three main molecular subtypes with different clinical outcomes and responses to immunochemotherapy: the germinal center B-cell-like (GCB) subtype, the activated B-cell-like (ABC) subtype and the primary mediastinal B-cell lymphoma (PMBL) subtype. Despite frequent translocations involving the IGH (heavy chain) locus, B-cell receptor (BCR) expression is retained in almost all DLBCL, indicating that BCR signaling is crucial for the survival of malignant cells. Furthermore, DLBCL are characterized by recurrent somatic mutations that are supposed to be oncogenic drivers. Here, with the aim to more accurately define the DLBCL pathogenic subgroups, we report a detailed immunogenetic analysis of IGH rearrangements in a well-defined GEP DLBCL cohort and correlate these features with the somatic mutation status of recurrently mutated genes involved in lymphomagenesis. Methods: 204 DLBCL enrolled in the prospective LYSA LNH03 trial program were classified into molecular subtypes by GEP using Affymetrix arrays [82 ABC (40.2%), 77 GCB (37.7%), 29 (14.2%) unclassified and 16 PMBL (7.8%)]. The amplification of complete VDJ rearrangements was realized using BIOMED-2 protocols. Somatic hypermutation (SHM) characteristics were studied (mutation rate, glycosylation N-X-S/T sites, RGYW hotspots, composition of CDR3) and correlated with the molecular subtypes. To obtain a more comprehensive molecular portrait of the DLBCL cases, GEP and IGH VDJ analyses were correlated with the mutational status of a panel of 34 recurrently mutated genes in DLBCL such as MYD88, EZH2, CD79B, TNFAIP3, CARD11 and GNA13 (Dubois et al. Clinical Cancer Research 2016). Results: A total of 153/204 (75%) clonal VDJ rearrangements were successfully determined. Failures to detect clonotypic sequences were predominantly seen in PMBL (56%) and unclassified (31%) cases. The ABC subtype display a significantly lower SHM rate than the 3 other subtypes, especially in contrast to the PMBL subgroup that was characterized by a higher SHM rate (p

Description: Background and aim of the study Primary mediastinal B-cell lymphoma (PMBL) is an entity of aggressive B-cell lymphoma that is clinically and biologically distinct from the other molecular subtypes of diffuse large B-cell lymphoma (DLBCL). We recently detected by Whole exome sequencing a recurrent point mutation in the XPO1 (exportin 1) gene (also referred to as chromosome region maintenance 1; CRM1), which resulted in the Glu571Lys (p.E571K) missense substitution in 2 refractory/relapsed PMBL (Dubois et al., ICML 2015; Mareschal et al. AACR 2015). XPO1 is a member of the Karyopherin-b superfamily of nuclear transport proteins. XPO1 mediates the nuclear export of numerous RNAs and cellular regulatory proteins, including tumor suppressor proteins. This mutation is in the hydrophobic groove of XPO1 that binds to the leucine-rich nuclear export signal (NES) of cargo proteins. In this study, we investigated the prevalence, specificity, and biological / clinical relevance of XPO1 mutations in PMBL. Patients and methods High-throughput targeted or Sanger sequencing of 117 PMBL patients and 3 PMBL cell lines were performed. PMBL cases were defined either molecularly by gene expression profile (mPMBL cohort) or by standard histological method (hPMBL cohort) and enrolled in various LYSA (LYmphoma Study Association) clinical trials. To assess the frequency and specificity of XPO1 mutations, cases of classical Hodgkin lymphoma (cHL) and primary mediastinal grey zone lymphoma (MGZL) were analysed. Cell experiments were performed to assess the impact of the E571 mutation on the activity of selective inhibitor of nuclear export (SINE) molecules. Results XPO1 mutations were present in 28/117 (24%) PMBL cases but were rare in cHL cases (1/19, 5%) and absent from MGZL cases (0/20). A higher prevalence (50%) of the recurrent codon 571 variant (p.E571K) was observed in PMBL cases defined by gene expression profiling (n = 32), as compared to hPMBL cases (n = 85, 13%). No difference in age, International Prognostic Index (IPI) or bulky mass was observed between the PMBL patients harboring mutant and wild-type XPO1 in the overall cohort whereas a female predominance was noticed in the mPMBL cohort. Based on a median follow-up duration of 42 months, XPO1 mutant patients exhibited significantly decreased PFS (3y PFS = 74% [CI95% 55-100]) compared to wild-type patients (3y PFS = 94% [CI95% 83-100], p=0.049) in the mPMBL cohort. In 4/4 tested cases, the E571K variant was also detected in cell-free circulating plasmatic DNA, suggesting that the mutation can be used as a biomarker at the time of diagnosis and during follow-up. Importantly, the E571K variant was detected as a heterozygous mutation in MedB-1, a PMBL-derived cell line, whereas the two other PMBL cell lines tested, Karpas1106 and U-2940, did not display any variants in XPO1 exon 15. KPT-185, the SINE compound that blocks XPO1-dependent nuclear export, induced a dose-dependent decrease in cell proliferation and increased cell death in the PMBL cell lines harbouring wild type or mutated alleles. To test directly if XPO1 mutation from E571 to E571K alters XPO1 inhibition by SINE compounds, the mutated protein was tested in vitro. The E571XPO1 mutated allele was transiently transfected into osteosarcoma U2OS cells which stably express the fluorescently labelled XPO1 cargo REV. Cells were treated with the clinical SINE compound selinexor, which is currently in phase I/II clinical trials and nuclear localization of REV-GFP was analysed in red transfected cells. The results showed that the nuclear export of the mutated XPO1 protein was inhibited by selinexor similarly to the wild-type XPO1 protein (Figure 1). Conclusion Although the oncogenic properties of XPO1 mutations remain to be determined, their recurrent selection in PMBL strongly supports their involvement in the pathogenesis of this curable aggressive B-cell lymphoma. XPO1 mutations were primarily observed in young female patients who displayed a typical PMBL molecular signature. The E571K XPO1 mutation represents a novel hallmark of PMBL but does not seem to interfere with SINE activity. Rev-GFP (green fluorescent) expressing U2OS cells were transfected with wild type XPO1-RFP (red fluorescent protein), XPO1-C528S-RFP, XPO1-E571K-mCherry, and XPO1-E571G-mCherry. The cells were then treated with 1µM KPT-330 for 8 hours. Figure 1. Rev-GFP expressing U2OS cells transfected with XPO1 variants. Figure 1. Rev-GFP expressing U2OS cells transfected with XPO1 variants. Disclosures Landesman: Karyopharm Therapeutics: Employment. Senapedis:Karyopharm Therapeutics, Inc.: Employment, Patents & Royalties. Argueta:Karyopharm Therapeutics: Employment. Milpied:Celgene: Honoraria, Research Funding.

Description: Abstract 1559 Follicular lymphoma (FL) is the most common indolent subtype of non-Hodgkin's lymphoma in the western world. The genetic hallmark of FL is the t(14;18)(q32;q21) translocation leading to the deregulation of BCL2 expression which occurs in up to 90% of the grade 1–2 FL. However, a minority of FL without BCL2 gene rearrangement harbour genetic abnormalities involving the BCL6 gene (5–15% of the cases) and show distinct pathological features. The activation of Signal Transducer and Activator of Transcription 6 (STAT-6) has been observed in Primary Mediastinal B-cell Lymphoma (PMBL) and Hodgkin Lymphoma but also in FL. Because missense mutations of STAT6 DNA binding domain have been described in PMBL, we searched for such mutations in FL. Using a PCR HRM (High Resolution Melting) assay as a screening tool and conventional Sanger sequencing in all cases with abnormal denaturation curves, we analyzed the frequency of STAT6 mutations in FL samples. We focused our screening on exon 12, which encodes part of the DNA binding domain and which had been shown to be a hotspot mutation in PMBL. A series of 40 FL lymphomas samples diagnosed at the University Hospital of Créteil (Henri Mondor) and Centre Henri Becquerel in Rouen were retrieved. These tumors showed characteristic histopathological features of FL according to the World Health Organization (WHO) classification. DNA analysis was performed on DNA extracted from fresh/frozen samples (Rouen) or FFPE (formalin-fixed and paraffin-embedded) tissues (Créteil) with standard procedures. We detected 5 (12%) mutated tumors in this series of FL. These 5 mutations were single missense mutations targeting amino acids 419–421. All mutations were observed in histological grade 1or 2 lymphomas, and no mutations were found in the 9 cases classified as 3A or 3B. Only one classical FL out of 15 cases (6%) with BCL2 rearrangement showed STAT6 mutation. Strikingly, the 4 other mutated cases showed specific features. There were 3 female and 1 male, with a mean age of 52 years, and all presented with inguinal involvement and stage III or IV Ann Arbor disease. Morphologically, these cases displayed a follicular growth pattern. The immunophenotype was CD20+, CD5-, CD10+ (3/4), BCL6+, CD23+ (3/4) but BCL2 was negative (4/4). Cytogenetically, these 4 cases were characterized by BCL6 gene rearrangement without BCL2 gene rearrangement by interphase FISH on FFPE tissue sections. Thus, STAT6 mutations were observed in 4/11 (36%) grade 1–2, and 0/6 grade 3A/3B BCL2 negative FL with BCL6 rearrangement. No case was found to be mutated in the FL group without BCL2 and BCL6 rearrangement (8 cases). In conclusion, this is the first time that mutations of STAT6 are found in FL and interestingly, they target a rare group of FL with distinct pathological and cytogenetical features. Further investigations are required to identify the mutational mechanisms involved and the oncogenic function associated with these mutations. Disclosures: No relevant conflicts of interest to declare.

Description: Beside the rare true leukemic forms of follicular lymphoma (FL), low levels of circulating tumor cells (CTC) are detected by PCR in the vast majority of FL patients at diagnosis. By a regular recirculating process, those CTC could reflect the total solid tumor mass. Alternatively, they could represent a subpopulation of tumor cells with distinct molecular profile including adhesion molecules and, as such, could add to the prognostic value of solid tumoral mass previously reported in FL (Dupuis, JCO 2013). In order to address these issues, we retrospectively selected FL patients treated in our institution between 2007 and 2014 who were simultaneously evaluated for both t(14;18) cells in peripheral blood (PB) and FDG-PET/CT tumor mass at diagnosis or at relapse. Absolute quantification of t(14;18) positive cells was performed using quantitative droplet digital PCR (ddPCR). Total metabolic tumor volume (TMTV) and total lesion glycolysis (TLG) were calculated from FDG-PET/CT, and baseline clinical characteristics and outcomes were collected. One hundred fourteen patients fulfilled the inclusion criteria. Using a routine 10-4 sensitive biomed 2 t(14;18) PCR assay, 75 had a positive PCR, either in peripheral blood alone (n=37), in peripheral blood and tumor biopsy (n=27) or in tumor biopsy but not PB (n= 11). Absolute quantitative ddPCR was performed for the 56 t(14;18) MBR+ patients. Clinical characteristics are given in Table 1. Median CTC value (number of t(14;18) positive cells out of total peripheral blood nucleated cells) was 1.6 10-3 (range 0-0.96), with only 5 CTC (-) patients and 13 patients with 〉10% CTC . Median TMTV was 267 cm3 (range 4.61-1900) and median TLG was 1473 (range 10.24-5912). A positive correlation was found between number of CTC and TMTV (R2 = 0.49; P 432 cm3 and TLG 〉 2717 tend to be associated with poorer OS (Fig. 2). The combined presence of 〉 6% CTC and TLG 〉 2717 allowed to identify a group of patients with 3-year OS of 71%, compared with 100% when both criteria were negative or dissociated (P= 0.01) (Fig. 2). In the subset of 42 patients with an untreated and untransformed FL, incremental prognostic value of circulating mass and metabolic tumor burden remained significant (P=0.03). Disclosures Dupuis: ABBVIE: Membership on an entity's Board of Directors or advisory committees; ROCHE: Speakers Bureau.