ALBERT

Description: Haemophagocytic lymphohistiocytosis (HLH) is a life-threatening systemic inflammatory clinical syndrome that can be primary/familial or secondary to a variety of underlying conditions. The 2004 diagnostic criteria for HLH require 5 out of 8 of the following clinical and pathological variables to be present: fever, splenomegaly, cytopenia in at least 2 lineages, hypertriglyceridemia/hypofibrinogenemia, haemophagocytosis on pathology examination, low/absent NK cell activity, ferritin greater than 500µg/L or soluble IL-2 receptor (sCD25) greater than 2400 U/mL. At least 5 of these criteria must be met, as each one lacks specificity. With regard to the pathology detection of haemophagocytosis, the sensitivity and specificity reported in the literature were 83% and 60% respectively. Indeed, a degree of haemophagocytosis can be seen outside the context of HLH. Furthermore, some of the diagnostic features of HLH may not be useful when applied to single cases because they may be intrinsic to the underlying condition. Therefore, we wondered whether the histopathological diagnosis might be improved. We investigated whether testing for S100B expression in macrophages might render the detection of haemophagocytosis more specific and sensitive. S100B is a marker of macrophage activation, and can routinely be detected by immunohistochemistry. S100B has pro-inflammatory properties and has been shown to affect macrophage function. It is typically expressed in macrophages of Rosai-Dorfman disease apart of having also a tissue-specific expression pattern, marking melanocytes as well as Langerhans cells and other subsets of dendritic cells. A natural language search of the pathology database at our institution identified 32 patients with bone marrow samples reporting haemophagocytosis as evaluated on bone marrow smear preparations, between January 2002 and July 2015. Cases that did not have available paraffin-embedded bone marrow trephine biopsy and clinical data were excluded. The bone marrow samples of three patients without haemophagocytosis and without any clinical features of HLH were used as controls: a new diagnosis of acute leukemia, a follow-up of acute leukemia post-transplant and a known lymphoma patient with unexplained pancytopenia. Clinical parameters relevant to the diagnosis of HLH were evaluated. Cases were clinically categorized as diagnostic for HLH (≥5 criteria), clinically likely (

Description: Background: Standard treatment in PTCL pts is unsatisfactory due to a high rate of early progression. Alemtuzumab (A), a monoclonal anti CD52 antibody, has demonstrated efficacy in relapsed PTCL pts. The results of the final analyses of the ACT-1 phase III and the ACT-2 phase III trial, comparing standard CHOP to A-CHOP showed higher response rates, but did not yield significant differences in EFS, PFS and OS in both previously untreated young and elderly patients, but the planned sample sizes could not be reached due to low recruitment. We undertook a planned pooled analysis of both ACT trials comparing CHOP with A-CHOP to increase statistical power. Patients and Methods: Between 2007 and 2013, 252 pts from Austria, Belgium, Czech Republic, Denmark, Finland, France, Germany, Norway, Poland, Portugal, Sweden and The Netherlands were randomized to receive either 6 cycles CHOP or A-CHOP at 14 day intervals. The planned pooled analysis was performed to compare patient outcome adding A to CHOP with patients treated with CHOP alone. Primary endpoint within ACT-1 and ACT-2 trial was the 3 years event free survival (EFS), secondary endpoints were progression free survival (PFS) and overall survival (OS). EFS is defined as the time from randomization to disease progression, start of salvage treatment, additional (unplanned) treatments, relapse, or death of any cause. Results: 252 pts were randomized (ACT-1: 136; ACT-2: 116). Five patients received no study treatment; therefore, 247 pts were analyzed (CHOP: 124; A-CHOP: 123). Median age was 61 years. 62% were male. Histologies were 33% AITL, 34% PTCL NOS, 33% other subtypes. In the pooled data set, the two treatment groups were comparable. Infections grade ≥3 occurred more often in patients treated with Alemtuzumab (A-CHOP: 55% [95% CI: 45% - 64%] vs. CHOP 23% [16% - 32%]; p 1 (HR 2.1) and bulky disease (HR 2.1) were the most prominent unfavorable risk factors for EFS. With the sample size of 247 pts the power for detecting the planned EFS difference of 15% was 74%. Conclusion: Adding Alemtuzumab to CHOP increased the rate of CR in PTCL patients, albeit at the costs of higher treatment related toxicity. Addition of Alemtuzumab did not improve EFS, PFS, or OS. Female gender is associated with a significantly better prognosis. This is the largest prospective dataset collected for PTCL. Supported by Federal Ministry of Research BMBF FKZ 01KG0705 and unrestricted research grants by Genzyme-Sanofi and AMGEN Disclosures No relevant conflicts of interest to declare.

Description: [§ share last authorship] Background: In 2000-2010, the first large prospective trials in peripheral T-cell lymphoma (PTCL) showed outcomes burdened by high failure rates during induction. Concurrently, trials with the anti-CD52 monoclonal antibody alemtuzumab (ALZ) yielded promising responses in PTCL while demonstrating the feasibility of combining ALZ with CHOP. Hence, the Nordic Lymphoma Group initiated the randomized ACT-1 trial to test, in younger patients (pts) (18-65yrs), the addition of ALZ to CHOP + autologous stem cell transplant (ASCT). Primary endpoint was the 3 years event-free survival (EFS). Here, we present the final analysis of the ACT-1 trial (ClinicalTrials.gov: NCT00646854). Patients and Methods: Overall, 136 pts were randomized (43% of planned sample size due to slow accrual), five did not receive study treatment, and 131 were analyzed (ALZ-CHOP: 65; CHOP: 66). Due to lack of tumoral CD52 expression, anaplastic large cell lymphomas (ALCL) were not included in the ACT-1 trial. An amendment tapering ALZ dose from 360 mg (30 mg on days 1+2 of each CHOP course) to 120 mg (30 mg on day 1 of CHOP courses 1-4) was introduced early on due to systemic fungal infections in 2 pts. Of the 65 pts treated with ALZ-CHOP, 4 received the pre- and 61 (94%) the post-amendment dose. Monitoring for CMV- and EBV-DNA and antimicrobial prophylaxis were mandatory. Results: The median observation time for the Full Analysis Set was 66 months and the median age 51 yrs. The ALZ-CHOP and CHOP cohorts were well balanced with regard to classical prognostic factors and histological subtypes (PTCL-NOS 58% vs 54%, AILT 21% vs 25%, other 21% vs 21%). Feasibility: Neither CHOP nor ALZ-CHOP pts experienced substantial treatment delay. ALZ exposure did not affect stem cell harvest nor hematopoietic recovery. Grade 4 leucopenia was more frequent in ALZ-CHOP pts (73% vs 35%; p=0.001), whereas the occurrence of grade 3-4 anemia and thrombocytopenia did not differ significantly. After ALZ dose amendment, the frequency of bacterial and fungal infections of grade ≥3 was similar in both treatment arms. ALZ treated pts had more viral events (22/57=42% vs 4/23=17%), mainly due to asymptomatic CMV reactivations. The ratio of serious adverse events per ALZ-CHOP treated patient dropped markedly (from 3.25 to 0.86, comparable with 0.46 for CHOP) after dose amendment. Additional toxicity was mild and similar in both arms. Treatment related mortality was 4% (5% vs 3%). Efficacy: Complete remission (CR) was 52% in ALZ-CHOP vs 42% in CHOP. Primary refractory disease occurred for ALZ-CHOP and CHOP in 23% and 38% of pts, respectively. Overall, females had a significantly better outcome than males (p=0.004), also after adjustment for classical prognostic factors. Analyzing time-related endpoints without knowledge of CD52 expression, 3-years EFS, progression-free, and overall survival (PFS, OS) did not differ significantly between ALZ-CHOP and CHOP (EFS 35% vs 26%, PFS 37% vs 26%, OS 52% vs 50%). Fig.1A shows EFS by treatment arm, by gender, and by gender and treatment arm. Although not significantly different, EFS, PFS and OS values of ALZ-CHOP treated females in the ACT-1 trial were consistently higher than those of non-ALZ treated females or of males regardless of treatment group. RNA sequencing from evaluable pre-therapeutic tumor biopsies defined a signature of differentially expressed genes to be predictive of clinical outcome in ALZ-CHOP but not CHOP treated pts (n=33). Tumor microenvironment genes were prominent in determining response to ALZ. Tumors rich in B-cell milieu showed good responses, while the opposite was observed in tumors with signatures enriched with high endothelial cell genes (p

Description: Primary cold agglutinin disease (CAD) is a type of hemolytic anemia mediated by anti-I autoantibodies. Patients suffer from anemia as well as circulatory problems. However, the severity of disease differs greatly between patients. We recently demonstrated that primary CAD is caused by an underlying low grade B cell lymphoproliferative disease of the bone marrow with a typical histology that is different from lymphoplasmacytic lymphoma and, accordingly, does not display the MYD88 L265P mutation (Randen et al., Haematologica, 2013). The majority of patients display circulating monoclonal antibodies encoded by the immunoglobulin heavy chain gene IGHV4-34. The disease severity does not correlate with antibody titers, but seems to be determined by the thermal amplitude, i.e., the highest temperature at which the cold agglutinin binds to the antigen. The framework region 1 of IGHV4-34 encodes for a sequence that binds to I antigen. However, this does not explain the molecular basis of disease heterogeneity. We studied 27 patients with well-characterized primary CAD and sequenced immunoglobulin heavy as well as immunoglobulin light chains to find additional consensus regions that may determine anti-I reactivity. Bone marrow aspirates, or frozen bone marrow trephine biopsies and blood from 27 patients with well-documented primary CAD were collected. Monoclonal B cells were isolated by flow sorting (FACS Aria Ilu High speed sorter, Becton Dickinson). Viable cells were detected using the forward scatter versus side scatter dot plot. Subsequently, CD45 bright events with low side scatter features representing lymphocytes, were selected. Then, CD5 positive and CD19 negative events, i.e. T cells, were gated out using a CD5 versus CD19 dot plot leaving only B cells. Finally, monoclonal B cells were selected using the immunoglobulin light chain gate, either k or l. Clonally rearranged IGH genes were detected using the Somatic Hypermutation Assay v2.0 (Invivoscribe) and were then sequenced. Immunoglobulin light chain genes (IGL) were amplified by an in-house diagnostic protocol based on Biomed-2 primers (van Dongen et al., Leukemia, 2003). All sequences were analyzed using the IMGT database (www.imgt.org). Productive IGHV4-34 gene rearrangements were identified in 22/27 patients. In 4 patients, no productive rearrangement was identified, while in one patient a productive IGHV3-23 was seen. No significant homology of complementarity determining region 3 (CDR3) regions was found between IGHV sequences. The N-glycosylation sequence within the CDR2 region, affecting antigen-binding, was mutated in 8 patients whereas no mutations were present in 7 patients and mutations in flanking residues were seen in 6 patients. The latter mutations may modulate glycosylation efficacy. Clonal rearrangement of the IGKV3-20 was detected in 16/27 patients, clonal IGKV3-15 gene rearrangements were identified in 4/27 patients whereas other IGL genes were rearranged in 4/27 patients. No clonal IGL gene rearrangement was found in 3/27 patients. Of interest, 7 of the patients with IGKV3-20 rearrangement displayed highly homologous CDR3 regions. The latter was highly associated with an un-mutated N-glycosylation sequence of the respective IGHV4-34 sequence. In conclusion, our data show that in addition to IGHV, also IGLV usage is highly restricted in CAD. Furthermore, stereotyped IGLV sequences are seen that are mutually exclusive with mutated N-glycosylation sequences in the IGHV CDR2 sequence. These data indicate that multiple regions within the immunoglobulin heavy chain as well as immunoglobulin light chain contribute to I-antigen binding. The data suggest that subtle differences in these multiple binding sequences may contribute to the differences in thermal amplitude of I antigen binding of the antibody. The highly restricted usage of IGKV3-20 provides a rationale for vaccination with IGKV3-20 proteins, known to be immunogenic and being considered for treatment in other lymphoproliferative diseases (Martorelli et al., Clin Cancer Res, 2012). Disclosures No relevant conflicts of interest to declare.

Description: Introduction: T-cell lymphomas (T-NHL) represent rarer entities compared to B-NHL, accounting for 5% -10% of NHL in Western countries and 15%-20% in Asia. They are divided into clinico-pathologic subtypes based on etiology, morphology, and clinical behavior. Because of the rarity and the lack of specific histologic features for the different subtypes, the diagnosis is difficult and clinical picture is usually very helpful to establish the diagnosis. We conducted a retrospective analysis of patients (pts) with T-NHL treated at our Centre, with the purpose of studying overall outcome and possible prognostic factors, including histologic subtypes, from our database. Patients and methods: Consecutive T-NHL pts (excluding Adult T-cell leukemia/lymphoma, NK/T NHL, and primary cutaneous T-NHL), receiving primary treatment at the Princess Margaret Cancer Centre (PMCC) between 2001-2014 were included. Data were extracted from a prospective patient database and the medical record regarding baseline characteristics, treatment, response and outcome. Response assessment was with CT imaging as per 1999 Working Group criteria. Results: Of a total of 2155 pts with aggressive histology NHL treated at PMCC between 2001-2014, 2031 pts had B-NHL and 124 (5.7%) T-NHL. Median age was 56 years (18-90), male/female ratio: 2.4; 63% presented with advanced stage (III-IV) disease, 22% had bone marrow involvement; 63% had elevated LDH and 44% had B symptoms. Observed subtypes were: Peripheral T-cell lymphoma, NOS 58 pts (PTCL NOS, 47%), Anaplastic large cell lymphoma, ALK-negative 16 pts (ALCL-ALK-, 13%), ALCL, ALK-positive 22 pts (ALCL-ALK+, 18%), Angioimmunoblastic T-cell lymphoma 13 pts (AITL, 10.5%), Enteropathy-associated T-cell lymphoma 7 pts (EATL, 5.6%), Hepatosplenic T-cell lymphoma 8 pts (HSTCL, 6%). 105/124 pts (85%) received induction chemotherapy; CHOP-like regimens were used in 94 pts (90%), and involved field radiation therapy (RT) was included in primary treatment in 24 pts (19%). Nineteen pts were treated palliatively, 5 pts with RT alone, 14 pts received palliative chemotherapy or supportive care only. Complete response (CR) was obtained in 63/105 pts (60%; Table 2), PR in 2 pts (1.9%) and 40 pts had no response or progressive disease (SD and PD; 38%). Considering together the most common subtypes (ALCL-ALK+/-, AITL, PTCL NOS), CR rate was 84% in limited stage vs 52% in advanced stage disease. Among patients with CR, 24 relapsed (38%). Fourteen pts received autologous stem cell transplant (8 at relapse, 6 for PD); 7/14 (50%) were alive at last follow-up. At a median follow-up of 5.3 years, 57/124 (46%) pts are alive. Cause of death was T-NHL in 50/124 (40%) pts. Two pts died of second malignancy (1.6%). Median overall survival (OS) and progression-free survival (PFS) were 4.57 years (95%CI: 2.23-9.53; 5yOS: 48%) and 1.5 years (0.87-3.17; 5yPFS: 37%), respectively (Table 2). For pts with limited stage disease median PFS was 4.57 years (5yPFS: 49%) and median OS 10.0 years (5yOS: 62%), while for pts with stage III/IV, median PFS was 0.82 years (5yPFS: 31%) and OS 1.81 years (5yOS: 38%). Median OS for pts who did not experience relapse (39/63; 62%) was 13.38 years (95%CI: 9.53-NA; 5yOS: 93%) vs 3.12 years (95%CI: 1.69-4.07 years; 5yOS: 25%) in pts who had relapse after CR1 (p

Description: Introduction. Recent genome profiling studies have increased our understanding of the mutation landscapes of myeloid malignancies. Molecular testing of AMLs (NPM1, FLT3-ITD, KIT) and MPNs (JAK2, CALR) constitute current diagnostic standard-of-care. Evidence for the diagnostic, prognostic and/or therapeutic impact of a growing set of genes and variants in myeloid malignancies allows for more accurate patient stratification and enhanced patient management. This has led to consideration of next-generation sequencing (NGS) approaches to simultaneously detect multiple variants in myeloid malignancies for use in the clinical diagnostic setting, to supplant single-gene molecular assays. We designed the Princess Margaret Advanced Genomics in Leukemia (AGILE) trial to prospectively assess the utility of NGS molecular profiling in the management of patients with myeloid malignancies. Methods. Patients for the AGILE trial are consented at the time of diagnosis using an REB approved written consent. Bone marrow or peripheral blood samples are collected at consent, accessioned within CoPath, and DNA extracted for NGS testing. NGS molecular profiling was performed using the TruSight Myeloid Sequencing Panel (TMSP; Illumina) on the MiSeq benchtop genome sequencer (Illumina) by the University Health Network Advanced Molecular Diagnostics Laboratory. The TMSP enables profiling of 54 genes (39 hotspot region; 15 complete coding region coverage) using amplicon-based library preparation and sequencing by synthesis. The TMSP detects the CALR 52 base pair deletion relevant to myelofibrosis, but not FLT3 internal tandem duplications greater than 30 base pair in size. Data were analyzed by NextGENe (v.2.3.1, SoftGenetics) and MiSeq Reporter v2.4.60. A specific script enabling alignment and calling of CALR deletions was added to the analysis to ensure there were no false negative calls. Additional testing and verification of CEBPA variants was performed by Sanger sequencing. Variants were interpreted according to Sukhai et al (Genetics in Medicine, 2015), reviewed by lab directors and reported in the Electronic Patient Record. Impact on patient care was defined as: potential for post-consolidation clinical trials; changes to frequency of monitoring; and, changes to transplant management. Cases were discussed in an interdisciplinary Genomic Tumor Board setting, at which NGS profiling data were reviewed in the context of all other diagnostic information for the patient, to determine impact on patient care. Results. Between February 11 and July 24, 2015, 162 patients were consented for AGILE; 148/162 were profiled by NGS, and to date 124/148 have been reviewed and interpreted. 62/124 (50%) of interpreted cases had a diagnosis of acute myeloid leukemia (AML); 21/124 (20%) with myeloproliferative neoplasms (MPNs); 13/124 (10%) with myelodysplastic syndromes (MDS); 6/124 (5%) with MDS/MPN; and, 15% with other hematologic malignancies. 90% of all cases profiled were informative for at least one variant (range 1-9 variants, average 3.1 variants/case). AML, MDS and MDS/MPN cases exhibited slightly more variants (3.4-4.4 variants/case) than did MPN cases (2.6 variants/case). Overall, 69% of variants were potentially actionable (Sukhai et al, 2015: 23% class 1; 8% class 2; 38% class 3), with a large fraction of cases (90/124, 72.6%) demonstrating at least one class 1 or class 3 variant. Additionally, 73/124 (58.9%) of patients exhibited actionable, class 1, variants not currently being identified by routine molecular diagnostics. In AMLs and MPNs, 88-90% of cases exhibited at least one potentially actionable variant; NGS profiling was more informative in AMLs (62% of cases exhibiting potentially actionable variants not profiled in standard of care testing, compared to 12% of MPN cases). Conclusions. We report the results of a prospective analysis of integrated NGS profiling in the context of diagnosis and management of patients with myeloid malignancies. Using a targeted NGS panel, molecular profiling of patients yielded significant information benefit over current standard approaches in 58.9% of cases analyzed, enabling potential impact on patient management. These data highlight the utility of NGS profiling to complement the initial diagnostic evaluation of myeloid malignancies. Disclosures Gupta: Incyte: Honoraria, Research Funding; Novartis: Consultancy, Membership on an entity's Board of Directors or advisory committees.

Description: Background. Survival of patients with high-risk diffuse large B-cell lymphoma (DLBCL) is suboptimal, and the risk of central nervous system (CNS) progression is relatively high. We investigated the efficacy of dose-dense chemoimmunotherapy and systemic CNS prophylaxis in two Nordic trials including patients less than 65 years with high-risk DLBCL. We combined individual patient data from these trials to compare clinical outcome and biological prognostic factors in patients treated with CNS prophylaxis given in the beginning (CHIC) versus at the end (CRY-04) of therapy. Patients and methods. In CRY-04 study, patients were treated with six courses of R-CHOEP14 followed by HD-Mtx and HD-Ara-C. In CHIC trial, treatment started with two courses of HD-Mtx in combination with R-CHOP14, followed by four courses of R-CHOEP-14 and one course of R-HD-AraC. In addition, liposomal AraC was administered intrathecally at courses 1, 3 and 5. For the correlative studies, formalin fixed paraffin embedded pretreatment tumor samples were analyzed by fluorescent in situ hybridization for BCL2 and c-MYC breakpoints and by immunochemistry for CD10, BCL6, MUM1, MYC and BCL2 expression. Germinal center B-cell-like (GCB)/non-GCB) subclassification was performed according to Hans algorithm. Results. Among 303 patients enrolled in the trials (CRY-04, n=160 and CHIC, n=143), 295 (CRY-04, n=154 and CHIC, n=139) were evaluable for baseline characteristics and outcome. Median age (54 and 56 years, p=0.222), male/female ratio, stage, and aaIPI scores were comparable in the two cohorts. CHIC regimen improved outcome over CRY-04; the findings included 4-year estimates of PFS (81% vs 66%, p=0.003), OS (83% and 79%, p=ns) and cumulative incidence rates of CNS progression (2.4% and 5.0%, p=ns). Treatment with the CHIC regimen reduced the risk of systemic progression (aaIPI adjusted RR=0.489, 95%CI 0.308-0.777, p=0.002). PFS benefit with CHIC over CRY-04 was observed across pre-specified subgroups, and particularly in patients less than 60 years old (p=0.008). In the entire study population, dual protein expression (DPE) of BCL2 and MYC was the only parameter to be significantly correlated with a worse PFS (4-y PFS 77% vs 50%, p=0.024; RR=2.300, 95% CI 1.088-4.860, p=0.029). Neither any single immunohistochemical marker nor the GCB/non-GCB subtype or MYC/BCL2-translocations significantly affected outcome. However, when treatment interaction was tested, MYC/BCL2 double hit status (DHL; 13%) predicted poor outcome among patients treated with CRY-04 regimen compared with patients who received CHIC regimen (4-y PFS; 38% vs 78%, p=0.086). GCB subtype and BCL2 positivity were also associated with better outcome in the CHIC cohort (4 y PFS; 63% vs 84%, p=0.011 and 61% vs 80%, p=0.007, respectively), whereas there were no significant survival differences between these regimens among the patients with non-GCB subtype, BCL2 negative DLBCL or DPE lymphomas. Conclusions. Our results derived from trial data with homogenous treatment support the use of HD-Mtx in the beginning rather than at the end of therapy. The survival benefit related to CHIC regimen over CRY-04 is due to better systemic control of the disease, and at least partly linked to improved survival among patients with GCB subtype, BCL2 positivity and DHL. Disclosures Leppa: Takeda: Consultancy, Research Funding; Janssen: Consultancy, Research Funding; Bayer: Research Funding; Roche: Consultancy, Honoraria, Research Funding; Celgene: Consultancy. Holte:Novartis Pharmaceuticals Corporation: Membership on an entity's Board of Directors or advisory committees; Roche, Norway: Research Funding.

Description: Primary cold agglutinin disease (CAD) is a hemolytic anemia mediated by monoclonal anti-I autoantibodies. CAD is caused by an underlying low grade B-cell lymphoproliferative disease of the bone marrow with a typical histology that is different from lymphoplasmacytic lymphoma and, accordingly, does not display the MYD88 L265P mutation (Randen et al., Haematologica, 2014). Since CAD is a clonal lymphoproliferative disorder, we studied the mutational landscape to further characterize the disease and identify potential novel treatment approaches. We prospectively collected bone marrow samples of CAD patients, enrolled in a clinical trial (CAD5; www.clinicaltrials.gov, NCT02689986). Exome sequencing of six cases was performed and findings were confirmed in ten additional cases using targeted sequencing. For these analyses, clonal B cells and normal T cells, used as control, were purified from bone marrow samples using fluorescent activated cell sorting. All mutations were verified by Sanger sequencing. Whole-exome sequencing was performed at BGI Tech Solutions (Hongkong) using the Agilent SureSelect Human All Exon V4 Reagent Kit and Illumina HiSeq technology. The bioinformatics pipeline consistent of BWA alignment tool (aligned to hg19); Picard tools FixMateInformation and MarkDuplicates; the Genome Analysis Toolkit (GATK) IndelRealigner and BaseRecalibrator; somatic variant detection tools Strelka, MuTect and Pindel; the annotation tool SnpEff. Recurrent somatic mutations were found in KMT2D (11/16 cases, 69%) and CARD11 (5/16, 31%) (Table 1, Figure 1-2). 7/16 (44%) of KMT2D mutations were deemed high impact mutations by SnpEff and to result in inactive protein. 2/16 (12,5%) of KMT2D mutations are missense mutations and are predicted to impair SET domain function. 2/16 (12,5%) of KMT2D mutations are classified by SnpEff as low impact mutations of which functional tests are necessary to demonstrate potential consequences for protein function. Of interest, two additional patients showed rare germline KMT2D variants that have also been seen in Kabuki syndrome patients, although these patients do not have Kabuki syndrome. CARD11 was somatically mutated in 5/16 (31%) cases. Four of those patients had a concurrent KMT2D mutation. All CARD 11 mutations were classified as moderate impact mutations by SnpEff. Mutations were tightly clustered in a 20bp sequence of the coiled-coil domain sequence (Table 1, Figure 2). KMT2D is a histone lysine methyl transferase that represses B cell lymphoma development. Mono-allelic mutations of KMT2D seem to act in a dominant fashion and cause partial loss of protein expression with cell growth advantage. KMT2D is frequently mutated in follicular lymphoma, diffuse large B cell lymphoma and nodal marginal zone lymphoma. Mono-allelic constitutional mutations cause Kabuki syndrome, characterized by distinct facial characteristics and multiple organ malformations. Patients frequently develop immune-mediated thrombocytopenia as well as auto-immune hemolytic anemia. CARD11 coiled-coil domain mutations result in constitutive NF-kB activation and enhanced NF-kB activity upon antigen receptor stimulation. Mutations were previously detected in diffuse large B cell lymphomas of activated B cell origin. Mono-allelic CARD11 coiled coil mutations are not oncogenic per se in mice and humans, but result in B-cell proliferation and auto-antibody production. Since four CAD patients showed concurrent KMT2D and CARD11 mutations, it seems likely that the mutations act in concert with anti-I B-cell receptor stimulation to contribute to CAD-associated lymphoproliferative disease. In conclusion, we demonstrated a high frequency of KMT2D and CARD11 mutations in the bone marrow B cell lymphoproliferative disease of patients with CAD. These results confirm that CAD-associated B cell lymphoproliferative disease is a distinct disease different from other known B cell lymphoproliferative diseases of the bone marrow, most notably lymphoplasmacytic lymphoma. The identification of these recurrent mutations in CAD may allow the design of novel treatment modalities. Table 1 Mutations in KMT2D and CARD11 gene in CAD. Table 1. Mutations in KMT2D and CARD11 gene in CAD. Figure 1 Mutations within KMT2D gene and protein detected in CAD patients. Figure 1. Mutations within KMT2D gene and protein detected in CAD patients. Figure 2 Mutations within CARD11 protein detected in CAD patients. Figure 2. Mutations within CARD11 protein detected in CAD patients. Disclosures No relevant conflicts of interest to declare.

Description: Background: The LY.12 randomized phase 3 clinical trial defined gemcitabine, dexamethasone and cisplatin (GDP) as an effective outpatient salvage chemotherapy regimen in relapsed/refractory (R/R) patients with aggressive lymphomas who are candidates for autologous stem cell transplant (ASCT) (Crump et al. JCO 2014). When the anti-CD20 antibody rituximab (R) was added to GDP, the ORR was 45.6% by CT imaging and 51.9% of patients were able to receive ASCT. Obinutuzumab (O) is a type 2 CD20 antibody that has demonstrated superiority to R in some studies in indolent lymphomas and is active in R-refractory follicular lymphoma. Improvements in the outcome of salvage therapy have tested alternative CD20 antibodies (Van Imhoff, JCO 2017), to date without success. We report a single centre, single arm clinical trial of O-GDP to assess safety and efficacy in R/R aggressive B cell lymphoma. Methods: Transplant eligible patients with DLBCL and transformed indolent lymphoma were treated with O-GDP for two cycles, followed by response assessment by CT. Non-progressors received a third cycle of O-GDP for stem cell mobilization and a PET/CT scan was obtained after stem cell collection. Responders then proceeded to ASCT per investigator decision. O was given at 1000 mg weekly during the first cycle of GDP and then on day 1 of cycles 2 and 3. Responses were determined by Lugano criteria using investigator assessment. The primary outcome was ORR by CT imaging after 2 cycles. The pre-specified statistical analysis stated that the trial will be declared positive if the ORR was 〉60%, negative if the ORR was 60%. The overall rate of proceeding to ASCT with O-GDP salvage in this trial was 65.5% compared to that previously reported for R-GDP at 51.9%. Disclosures Scott: NanoString: Patents & Royalties: Named Inventor on a patent licensed to NanoString Technologies, Research Funding; Roche: Research Funding; Janssen: Research Funding; Celgene: Consultancy, Honoraria. Kuruvilla:BMS: Consultancy, Honoraria; Abbvie: Consultancy; Leukemia and Lymphoma Society Canada: Research Funding; Seattle Genetics: Consultancy, Honoraria; Amgen: Honoraria; Celgene: Honoraria; Karyopharm: Honoraria; Janssen: Consultancy, Honoraria, Research Funding; Princess Margaret Cancer Foundation: Research Funding; Gilead: Consultancy, Honoraria; Lundbeck: Honoraria; Roche: Consultancy, Honoraria, Research Funding; Merck: Consultancy, Honoraria.

Description: Introduction: Enteropathy-associated T-cell Lymphoma (EATL) is a rare lymphoma of the GI tract accounting for less than 1% of all non-Hodgkin's lymphomas (NHL). EATL type I is seen predominately in patients (pts) from parts of Europe associated with a high incidence of celiac disease; type II is less frequently associated with celiac disease and thus a more worldwide distribution. EATL is difficult to treat with very poor prognosis. The purpose of this study is to understand the epidemiology and outcomes of this rare lymphoma in a North American multicultural setting at a tertiary cancer care centre. Methods: All pts with EATL who were assessed and treated at Princess Margaret Cancer Centre in Toronto, Canada, between January 1, 1996 and January 30, 2015 were retrospectively reviewed with case ascertainmentfrom a pathology database as well as a prospectively populatedlymphoma database. Pathological diagnoses were made according to the WHO classification and central pathology review was undertaken when available. Staging procedures included CT scans and bone marrow biopsy. Simple descriptive statistics were calculated as well as progression-free survival (PFS) and overall survival (OS). Results: 4852 new NHLs were seen for primary treatment at our centre during this time period, of which 287 were T cell lymphomas; 11 had EATL (0.23% of all NHL cases and 3.8% of all T-cell lymphomas). Median age at presentation was 61 years (range 51-85) with the majority being male (64%) (Table 1). All 11 pts presented with abdominal pain from small bowel perforation or obstruction. The most common site of EATL was jejunum (55%). Four pts (36%) had EATL type 1 and three had positive celiac serology. None had a history of celiac disease prior to diagnosis. 46% and 36% of pts presented with a high Ann Arbour stage (stage III or IV) and high Lugano stage (stage III or IV) respectively. No pt had a high ECOG score at presentation. Central pathology review was available for 8 patients (table 1). Interestingly, only 36% of cases had an intermediate-high risk International Prognostic Index (IPI) score of 3 and no pts had a high risk score of 4 or 5. Only one pt had an elevated LDH. All type 1 pts were CD56-, while type 2 pts were all CD8+ and most (86%) were CD56+. Ten pts (91%) received CHOP chemotherapy (one with the addition of etoposide and one of an IL2 receptor antibody) and one pt received palliative therapy. Only 5 of 11 responded to therapy (45%), with 2 patients achieving complete remission. All patients have progressed through treatment or relapsed (6 pts primary refractory and 5 responders), with one late relapse 14 years post response. At relapse, patients often presented with another bowel perforation. Upon relapse, four (36%) pts received salvage chemotherapy and 1 patient received an autologous stem cell transplant but progressed 16 months later and died from his disease. Ten patients (91%) have died, all from lymphoma; one patient (the late relapser) is presently undergoing salvage chemotherapy. Both median PFS and OS for all pts were 7 months. Two year OS and PFS were 18% and 9% respectively. Table 1 provides PFS and OS for EATL type separately. Conclusions: EATL is an aggressive form of lymphoma with unique presentation and very poor prognosis despite frequently presenting with limited disease and low IPI score. In our centre, type II EATL was more common, potentially underlying the multicultural composition of our city. CHOP with or without etoposide remains inadequate therapy for this aggressive lymphoma, and alternative induction regimens are needed. Few patients are stem cell transplant candidates at relapse due to age and poor performance status. Table 1. Patient characteristics Feature EATL1 (N=4) EATL 2 (N=7) Age (yrs) median (range) 65 (55-71) 61 (51-85) Gender (%) male 1 (25) 6 (86) Celiac status (%) positive 3(75) 0 Pathology CD3+ 4(100) 7(100) CD7+ 3(75) 5(71) CD4+/CD8- 1(25) 0 CD4-/CD8+ 2(50) 7(100) CD4-/CD8- 1(25) 0 CD56+ 0 6(86) T-cell receptor clonal gene rearrangement 1/2(50) 4/4(100) Symptoms (%) abdominal pain 4(100) 7(100) constitutional symptoms 3(75) 1(17) lymphadenopathy 2(50) 1(14) bone marrow 1 (25) 0 Primary Location (%) duodenum 2 (50) 3 (43) jejunum 1 (25) 5 (71) ileum 3 (75) 0 Monofocal (%) yes 3 (75) 3 (43) Anne Arbor Staging (%) III or IV 2 (50) 3 (43) Lugano Staging (%) III or IV 2 (50) 2 (29) IPI (%) score 3-5 1(25) 3 (43) Response to tx CR/CRu/PR 1(25) 4 (57) PFS median (mo) 6 7 2 year PFS 25% 0% OS median (mo) 7 8 2 year OS 25% 14% Disclosures Prica: Janssen: Honoraria; Celgene: Honoraria. Crump:Celgene: Consultancy; Seattle Genetics: Consultancy; Roche Canada: Consultancy. Kukreti:Celgene: Honoraria; Roche: Honoraria; Lundbeck: Honoraria; Amgen: Honoraria; Janssen Ortho: Honoraria. Kuruvilla:Lundbeck: Honoraria; Karyopharm: Honoraria; Merck: Honoraria; Bristol-Myers Squibb: Honoraria; Janssen: Consultancy, Honoraria; Gilead: Consultancy; Seattle Genetics: Consultancy, Honoraria; Hoffmann LaRoche: Consultancy, Honoraria, Research Funding.