ALBERT

Description: Introduction: Tumor boards have become a crucial institution in oncology practice to provide paramount interdisciplinary cancer treatment, stream-line patient (pt) entries and to ensure treatment according to clinical pathways (CP). We initiated a weekly MM-TB at our institution in 6/2012. Participating experts are hematologist-oncologists, pathologists/cytogenetic specialists, orthopedists, radiotherapists, immunologists/rheumatologists and, if needed, nephrologists, cardiologists and others. Pt applications to be discussed are centrally organized through our CCCF, with the TB advice being centrally stored within our electronic pt information system. Recommended TB advice is made according to best current literature/knowledge and international CP. The development of mandatory CCCF-CP and transparency of decision making are key quality criteria. Methods: This first analysis focused on a) discussed TB questions, b) given recommendations, c) pt characteristics, d) pts’, referring- and participating-physicians' satisfaction with the TB, e) inclusion of these challenging-to-treat pts in clinical trials (CT) and f) PFS/OS of TB pts as compared to the literature (Kumar SK. Leukemia 2012). Grades of recommendations were assigned using the GRADE criteria (Engelhardt M. Haematologica 2014) and meticulously assessed, as well as whether TB recommendations were pursued. Pts’, referring- and participating-physicians' satisfaction with the TB was evaluated via standardized questionnaires, the aimed sample size being n=100 for consecutive pts and ~n=30 each for participating and referring physicians. Results: From 6/2012-5/2014, 483 pts have been discussed within 90 MM-TB sessions, substantially increasing these from 2011 to 2012, 2013 and 2014 by 12-fold. Of the entire MM cohort seen at our institution, 60% of these challenging-to-treat pts were discussed within the TB in 2012, increasing to 71% in 2013. We have currently assessed 200 TB-protocols for pt characteristics, clinical outcome and adherence to TB decisions. Of those, 2% were presented for explicit diagnosis-finding, 17% had newly diagnosed MM, 41% relapsed/refractory MM and 40% had attained stable disease or better with their last-line therapy and were discussed to resolve their ongoing treatment. Expectedly, most pts (89%) were discussed for their next-line treatment, 43% due to strains with comorbidities, symptom control, side effects, diagnosis finding and MM-staging, and 11% due to various other reasons (multiple entries possible). Mean treatment lines of pts discussed in the TB was 2 (range 0-10), deciding on their 3rd-line-treatment. Within the TB cohort, 70% were presented once, but 30% several times (mean 2, range 2-4). Of these multiple presentations, most pts had relapsed or refractory MM, this rate further increasing towards the 3rd and 4th TB-presentation. The adherence to TB-recommendations was excellent with 93% of decisions being pursued. Reasons for adapted approaches were practicable issues or disagreement of the pt, family or referring physician. Of currently 80/100 interviewed pts, 95% were entirely satisfied with their care, treating oncologists/MM-expert team and very supportively perceived the MM-TB. Of note, 94% considered their cancer care ideally achieved by the TB, 92% that their local physician profited greatly and 88% that their personal preferences were also accounted for. Of 30 interviewed participating physicians, 97% considered themselves well-educated and their time well-spent. Of currently 18 referring physicians, 73% were unconditionally satisfied with all TB-diagnostics and -therapies, with the university centers' cooperation and 65% acknowledged no information loss. Of 288 pts assessed for their CT suitability, 28% were suggested by the TB to be included, with 53% actually being able to enter therein. Thus, 15% of our MM-TB cohort could be included in a CT, which is considerable since these were challenging-to-treat pts who had received extensive prior therapies and showed several comorbidities. This also confirms current CT accrual rates for cancer pts of 5-15%, which can be increased with well-structured TB. Conclusions: Our preliminary results suggest that this MM-TB is a highly relevant exchange platform and allows physicians from different disciplines to intensely and rewardingly collaborate for state-of-the-art cancer care. Disclosures No relevant conflicts of interest to declare.

Description: The FOXO transcription factors control proliferation and apoptosis in different cell types. Their activity is regulated by posttranslational modifications, mainly by the PI3K-PKB pathway, which controls nuclear export and degradation. We show that FOXO1 is highly expressed in normal germinal center B cells as well as in non-Hodgkin lymphomas, including follicular lymphoma, diffuse large B-cell lymphoma, mucosa-associated lymphoid tissue non-Hodgkin lymphoma, B-cell chronic lymphocytic leukemia, and mantle cell lymphoma. In contrast, in 31 of 32 classical Hodgkin lymphoma (cHL) cases, Hodgkin and Reed-Sternberg cells were FOXO1 negative. Neoplastic cells of nodular lymphocyte-predominant Hodgkin lymphoma were negative in 14 of 20 cases. FOXO1 was down-regulated in cHL cell lines, whereas it was expressed in non-Hodgkin lymphoma cell lines at levels comparable with normal B cells. Ectopic expression of a constitutively active FOXO1 induced apoptosis in cHL cell lines and blocked proliferation, accompanied with cell-cycle arrest in the G0/G1 phase. We found that, in cHL cell lines, FOXO1 is inactivated by multiple mechanisms, including constitutive activation of AKT/PKB and MAPK/ERK kinases and up-regulation of microRNAs miR-96, miR-182, and miR-183. These results suggest that FOXO1 repression contributes to cHL lymphomagenesis.

Description: The survival of diffuse large B-cell lymphoma patients varies considerably, reflecting the molecular diversity of tumors. In view of the controversy whether cytologic features, immunohistochemical markers or gene expression signatures may capture this molecular diversity, we investigated which features provide prognostic information in a prospective trial in the R-CHOP treatment era. Within the cohort of DLBCLs patients treated in the RICOVER-60 trial of the German High-Grade Lymphoma Study Group (DSHNHL), we tested the prognostic impact of IB morphology in 949 patients. The expression of immunohistochemical markers CD5, CD10, BCL2, BCL6, human leukocyte antigen (HLA)–DR, interferon regulatory factor-4/multiple myeloma-1 (IRF4/MUM1), and Ki-67 was assessed in 506 patients. Expression of the immunohistochemical markers tested was of modest, if any, prognostic relevance. Moreover, the Hans algorithm using the expression patterns of CD10, BCL6, and interferon regulatory factor-4/multiple myeloma-1 failed to show prognostic significance in the entire cohort as well as in patient subgroups. IB morphology, however, emerged as a robust, significantly adverse prognostic factor in multivariate analysis, and its diagnosis showed a good reproducibility among expert hematopathologists. We conclude, therefore, that IB morphology in DLBCL is likely to capture some of the adverse molecular alterations that are currently not detectable in a routine diagnostic setting, and that its recognition has significant prognostic power.

Description: Introduction The combination of bortezomib, doxorubicin and dexamethasone (BDD) is well tolerated and induces a high overall response rate (ORR). Preclinical studies have demonstrated that vorinostat, a histone deacetylase inhibitor, is synergistic with bortezomib and doxorubicin. The aim of this phase I/II study was to determine the tolerability and activity of the combination of BDD with vorinostat (VBDD) in relapsed/refractory multiple myeloma (MM). Methods Patients received escalated vorinostat-doses (provided by MSD) at 100mg (dose level 0), 200mg (dose level +1) and 300mg (dose level+2) on days 1-4, 8-11, 15-18, combined with bortezomib 1.3mg/m2 day 1,8,15 (provided by Janssen), dexamethasone 40mg day 1,8,15,22 and doxorubicin on 9mg/m2 day 1+8. The primary objective was the maximum tolerated dose (MTD; 3+3 dose escalation design). Secondary objectives were safety, response assessed by EBMT and IMWG criteria, progression-free survival and overall survival. Correlative endpoints include prognostic MM-parameters, organ function, QoL-, comorbidity-assessments and translational studies (e.g. HDAC-activity in PB MNCs, Figure 1). Dose limiting toxicities (DLTs) were defined as any possibly drug related adverse events (AEs) ≥grade 3 (CTCAE) within the 1st cycle. After completion of 6 cycles, patients could continue with bortezomib maintenance therapy or proceed to (most often 2nd) ASCT. Results To date, 18/30 patients have been enrolled (median age 63 years [range 54-75], 55% men). The median Karnofsky Performance Status was 90% (range 70-100%). Median prior therapy lines were substantial with 3 (range 1-8): bortezomib, thalidomide or lenalidomide were given in 88% and 24% each, respectively; 94% of patients had undergone prior SCTs. Cytogenetic abnormalities included del(17p) (n=2), t(4;14) (n=2), gain(1q) (n=2), t(11;14) (n=4) and hyperdiploidy (n=7). No DLTs have been observed to date; with 3 patients each being included in dose level 0 and dose level +1 and the following patients safely proceeding to dose level +2. Six SAEs occurred in 4/18 patients (22%): bacteraemia (n=1) and herpes zoster reactivation (n=1) were suspected to be related to all VBDD-drugs. No causal relationship to study drugs was suspected for pneumonia (n=2), 1 syncope and 1 death due to PD with persisting plasma cell leukemia. The ORR (〉PR) and clinical benefit rate (SD, PR, CR) was 65% and 89%, respectively. At a median follow-up of 8 months (range 3-23), there have been only 2 patients with PD (refractory MM + plasma cell leukemia). The analysis of HDAC activity after VBDD initiation demonstrated downregulation in 6/8 (75%) patients (Figure 1). Further analyses will determine, whether HDAC activity and treatment response may correlate and whether this HDAC downregulation may precede and/or indicate depth of response Conclusions VBDD is a well tolerated and effective regimen in heavily pretreated relapsed/refractory MM patients. There have been no observed DLT and the MTD of vorinostat was set at 300mg, with all reported SAEs being in line with the known safety profile of the investigated drugs. Our alternative vorinostat-schedule (dosing of 4 days on and 4 days off) induced excellent tolerability and seems to enhance the antimyeloma response, warranting completion of this study. Disclosures: Kleber: MSD, Janssen-Cilag: Research Funding. Off Label Use: Preclinical studies have demonstrated that vorinostat, a histone deacetylase inhibitor, is synergistic with bortezomib and doxorubicin. The aim of this phase I/II study was to determine the tolerability and activity of the combination of BDD with vorinostat (VBDD) in relapsed/refractory multiple myeloma. Vorinostat is off-label use for MM patients, all other drugs are in label use. Waesch:MSD, Janssen-Cilag: Research Funding. Engelhardt:MSD, Janssen-Cilag: Research Funding.

Description: IL-33 is a recently discovered cytokine involved in induction of Th2 responses and functions as an alarmin. Despite numerous recent studies targeting IL-33, its role in vivo is incompletely understood. Here we investigated inflammatory responses to intraperitoneal IL-33 injections in wild-type and mast cell–deficient mice. We found that wild-type mice, but not mast cell–deficient Wsh/Wsh mice, respond to IL-33 treatment with neutrophil infiltration to the peritoneum, whereas other investigated cell types remained unchanged. In Wsh/Wsh mice, the IL-33–induced innate neutrophil response could be rescued by local reconstitution with wild-type but not with T1/ST2−/− mast cells, demonstrating a mast cell–dependent mechanism. Furthermore, we found this mechanism to be partially dependent on mast cell–derived TNF, as we observed reduced neutrophil infiltration in Wsh/Wsh mice reconstituted with TNF−/− bone marrow–derived mast cells compared with those reconstituted with wild-type bone marrow–derived mast cells. In agreement with our in vivo findings, we demonstrate that humanneutrophils migrate toward the supernatant of IL-33–treated human mast cells. Taken together, our findings reveal that IL-33 activates mast cells in vivo to recruit neutrophils, a mechanism dependent on IL-33R expression on peritoneal mast cells. Mast cells activated in vivo by IL-33 probably play an important role in inflammatory reactions.

Description: The transcription factor KLF4 may act both as an oncogene and a tumor suppressor in a tissue-depending manner. In T- and pre-B-cell lymphoma, KLF4 was found to act as tumor suppressor. We found the KLF4 promoter methylated in B-cell lymphoma cell lines and in primary cases of B-cell lymphomas, namely, follicular lymphoma, diffuse large B-cell lymphoma, Burkitt lymphoma, and in classic Hodgkin lymphoma (cHL) cases. Promoter hypermethylation was associated with silencing of KLF4 expression. Conditional overexpression of KLF4 in Burkitt lymphoma cell lines moderately retarded proliferation, via cell-cycle arrest in G0/G1. In the cHL cell lines, KLF4 induced massive cell death that could partially be inhibited with Z-VAD.fmk. A quantitative reverse-transcribed polymerase chain reaction array revealed KLF4 target genes, including the proapoptotic gene BAK1. Using an shRNA-mediated knock-down approach, we found that BAK1 is largely responsible for KLF4-induced apoptosis. In addition, we found that KLF4 negatively regulates CXCL10, CD86, and MSC/ABF-1 genes. These genes are specifically up-regulated in HRS cells of cHL and known to be involved in establishing the cHL phenotype. We conclude that epigenetic silencing of KLF4 in B-cell lymphomas and particularly in cHL may favor lymphoma survival by loosening cell-cycle control and protecting from apoptosis.

Description: Abstract 2417 Primary mediastinal B-cell lymphoma (PMBL) is a subtype of diffuse large B-cell lymphoma (DLBCL) that affects predominantly young women (Swerdlow et al. 2008). Despite improvements due to addition of rituximab, which has become state of the art treatment, 20% of PMBL patients succumb to disease progression or relapse. Notably, here are currently no registered trials that are actively recruiting PMBL-patients and a better understanding of the underlying pathobiology may identify novel therapeutic targets and provide an alternative to dose escalation (Steidl and Gascoyne 2011). BCL6 is a key germinal center B-cell transcription factor that suppresses genes involved in lymphocyte activation, differentiation, cell cycle arrest and DNA damage response gene. BCL6 is aberrantly expressed in certain DLBCL subgroups and BCL6 overexpression is sufficient for lymphomagenesis in mice (Cattoretti et al. 2005). In cellular- and murine DLBCL models, targeting of BCL6 via retroinverted BCL6 peptid inhibitor (RI-BPI) appears effective (Polo et al. 2004; Cerchietti et al. 2010). In conjunction with the relatively restricted expression pattern of BCL6, these data collectively suggest BCL6 as a candidate for targeted therapy in BCL6-positive lymphomas. Despite substantial work on BCL6 in lymphomas, the function of BCL6 in PMBL is unknown. To address the BCL6 function in PMBL, we performed BCL6 depletion by siRNA in all three available PMBL cell lines: K1106, U-2940 and MedB-1. We found that BCL6 acts pro-proliferative and anti-apoptotic; however, PMBL models were only partially dependent on and not addicted to BCL6. Given that BCL6 expression in all PMBL cell lines is variable with a notable fraction of BCL6-negative cells, we argued that increasing the fraction of BCL6-positive cells might increase the level of BCL6-dependence. Since IL-4/STAT6 signaling upregulates BCL6 in mouse lymphocytes (Schroder et al. 2002), we treated PMBL cell lines with IL-4 (or IL-13) and, as expected, observed increased phosphorylated (p)STAT6 levels. Surprisingly, the pSTAT6 increase was not associated with higher – but with drastically lower BCL6 protein levels. Moreover, in untreated cells, co-localization studies for pSTAT6- and BCL6 demonstrated staining in mutually exclusive subsets of cells (Figure 1A), suggesting negative interaction between BCL6 and pSTAT6. Other STAT family members were already shown to participate in the transcriptional regulation of BCL6. Thus, we examined binding of STAT6 to the proximal promoter of BCL6 in all PMBL cell lines using shift assay and chromatin immunoprecipitation. We found that STAT6 can bind all five GAS binding sites within the BCL6 promoter in vitro and in all PMBL cell lines STAT6 was bound to proximal BCL6 promoter in vivo. Furthermore, transient STAT6 depletion by siRNA and/or ectopic expression of constitutively active STAT6 confirms that pSTAT6 is sufficient for transcriptional repression of BCL6. Co-localization studies in primary patient samples demonstrated mutually exclusive BCL6/pSTAT6 distribution as a visual hallmark of the repression mechanism (Figure 1B, C). Thus, our data demonstrate for the first time that constitutively active STAT6 transcriptionally represses BCL6 in PMBL. In conjunction with functional data, the delineated repression mechanism may prevent addiction to one single oncogenic pathway (i.e. BCL6) in PMBL. Figure 1. Mutually exclusive distribution of BCL6 and pSTAT6 in PMBL Figure 1. Mutually exclusive distribution of BCL6 and pSTAT6 in PMBL Disclosures: No relevant conflicts of interest to declare.

Description: In contrast to other B cell lymphomas, Hodgkin and Reed-Sternberg (HRS) cells, the malignant cells of classical Hodgkin lymphoma (cHL), have mainly lost their B cell identity. Recently, we reported that the transcription factor FOXO1, indispensable for B cell development and differentiation, is downregulated in HRS cells and expression of FOXO1 results in growth arrest and apoptosis in cHL cell lines. To find molecular targets of FOXO1 we performed gene expression profiling of 5 cHL cell lines expressing constitutively active FOXO1 fused to ligand binding domain of estrogen receptor (FOXO1(A3)ER). We found that FOXO1 activation led to downregulation of genes, which are normally upregulated in cHL, such as CD30/TNFRSF8 and the proto-oncogene MYC. On the other hand, FOXO1 activated expression of germinal center-specific genes including BCL6, AICDA, BACH2, and GCET2. The most surprising finding was induction of BLIMP-1/PRDM1, the master regulator of plasma cell differentiation and tumor suppressor in activated B cell-like diffuse large B cell lymphoma (ABC-DLBCL). A positive correlation of FOXO1 and BLIMP-1 expression in microdissected HRS cells further indicated a role for FOXO1 in BLIMP-1 regulation. Of note, HRS cells do not undergo plasma cell differentiation despite constitutive expression of NF-κB, IRF4, and STAT3, known inducers of BLIMP-1. However, BLIMP-1 levels remain low in HRS cells thereby preventing terminal differentiation. The mechanisms leading to BLIMP-1 repression in cHL are poorly understood. We found that in most cHL cell lines, FOXO1 specifically upregulated BLIMP-1α representing the full-length isoform of BLIMP-1. In addition, luciferase reporter assay showed that FOXO1 activates the BLIMP-1α promoter in the cHL cell line L428. Overexpression of BLIMP-1α using a lentiviral vector strongly inhibited growth of cHL cell lines. In line with the fact that BLIMP-1 and MYC form a negative feedback loop, we found that ectopic BLIMP-1α expression resulted in downregulation of MYC protein levels, which most likely contributes to the anti-tumor effect of BLIMP-1α. On the other hand, MYC inhibition in cHL cell lines led to BLIMP-1 upregulation. Thus, our data indicate that FOXO1, BLIMP-1 and MYC constitute a negative feed-forward regulatory loop, which is controlled by FOXO1. Searching for additional mechanisms of BLIMP-1α downregulation in cHL we found that BLIMP-1α promoter was hypermethylated in two cHL cell lines but not in microdissected HRS cells. This indicates that hypermethylation is not critical for BLIMP-1α repression. Interestingly, the functionally impaired BLIMP-1 variant BLIMP-1β, which is normally not detected in normal B cells, was upregulated in cHL cell lines. BLIMP-1β is also expressed in ABC-DLBCL and has been suggested to block BLIMP-1α in a dominant-negative manner. We speculate that BLIMP-1β attenuates function of residual BLIMP-1α in cHL. However, further investigation is required. Taken together, we show for the first time that FOXO1 induces expression of BLIMP-1. We identified BLIMP-1α as a tumor suppressor which downregulates expression of the proto-oncogene MYC in cHL. Therefore, FOXO1 might exert its tumor suppressor function at least in part by upregulation of BLIMP-1α. Further work on BLIMP-1α regulation by FOXO1 and investigation of a potential role of FOXO1 in plasma cell differentiation is warranted. Disclosures: No relevant conflicts of interest to declare.

Description: Abstract 1632 Background: OFA is a fully human monoclonal antibody that binds to both the large and small extracellular loops of CD20. OFA is currently approved for patients (pts) with refractory chronic lymphocytic leukemia and has demonstrated activity in non-Hodgkin's lymphomas, including follicular lymphoma (FL). We previously reported results of a phase II study of OFA in combination with CHOP (cyclophosphamide 750 mg/m2, doxorubicin 50 mg/m2, vincristine 1.4 mg/m2, prednisone 100 mg daily for 5 days) chemotherapy (O-CHOP) in pts with previously untreated FL (Czuczman et al. Br J Haematol. 2012;157:438). We now report updated efficacy, safety and pharmacokinetic (PK) follow-up data for this study. This trial is registered at www.clinicaltrials.gov (NCT00494780). Methods: Fifty-nine pts with previously untreated FL were randomized to OFA 500 mg (n = 29) or 1000 mg (n = 30) on day 1, with CHOP on day 3, every 3 weeks for 6 cycles. The primary end point was overall response rate (ORR), as assessed by an independent end points review committee. Secondary end points included complete response (CR), progression-free survival (PFS), overall survival, adverse events (AEs) and PK. Follow-up assessments after therapy were done every 3 months (mo) until mo 12 and then every 6 mo until alternative FL therapy or mo 60. Positron emission tomography (PET) was done at baseline and 3 mo after last therapy. Blood samples for PK analyses were collected to determine OFA serum concentrations, and noncompartmental methods were used to estimate PK parameter values. Results: Fifty-eight pts received therapy; 1 pt in the 1000-mg group withdrew before initiation of therapy. The ORR was 90% for the 500-mg group (n=29) and 100% for the 1000-mg group (n=29); 55% of pts achieved CR or unconfirmed CR (CRu), including 67% of pts with a Follicular Lymphoma International Prognostic Index (FLIPI) score of 3–5. At baseline, 57 pts were PET positive, and 49 pts underwent repeat PET scans after therapy. Forty of 49 pts (82%) became PET negative, including 27 of 29 (93%) pts who achieved CR/CRu and 13 of 20 pts (65%) who achieved partial response (PR). With a median follow-up of 33.8 mo, the median PFS for the 500-mg group was 27.6 mo and the median PFS for the 1000-mg group was not reached (P=0.46). Median PFS for pts with FLIPI scores of 0–1 (n=17), 2 (n=20) and 3–5 (n=21) was not reached, 27.6 mo and 27.6 mo, respectively (P=0.68). Median PFS for pts (n=32) who achieved CR/CRu was also not reached and was 28.3 mo for pts (n=23) achieving PR. Median PFS for PR pts who were PET positive and PET negative after therapy was not reached and 28.3 mo, respectively. No deaths have been reported. No hematologic serious AEs (SAEs) were experienced during the follow-up period. During the follow-up period, non-hematologic SAEs were reported in 1 pt in the 500-mg group (pneumonia) and 5 pts in the 1000-mg group (abdominal hernia, erysipelas, intervertebral disc protrusion, meniscus lesion and vulval cancer); none were ofatumumab-related. After repeated dosing, OFA clearance values were 6.3 and 5.9 mL/h, and half-life values were 27.2 and 26.8 days in the 500-mg and 1000-mg groups, respectively. Conclusions: O-CHOP achieved durable remissions in previously untreated pts with FL. There were no observed PK or PFS differences between the 500-mg and 1000-mg arms, but the study was not powered to detect such differences. O-CHOP was effective in pts with high-risk FLIPI scores, and CR/CRu and PFS rates were not affected by FLIPI score. PET status after therapy did not predict PFS in responding pts, although the study was too small to make such a determination. These results indicate that O-CHOP should be studied as a therapy for FL pts with high-risk FLIPI scores. Disclosures: Czuczman: GlaxoSmithKline: Advisory board Other, Honoraria. Off Label Use: Ofatumumab in follicular lymphoma. Belada:GlaxoSmithKline: Research Funding. Mayer:Roche: Consultancy, Research Funding; GlaxoSmithKline: Consultancy, Research Funding. Gupta:GlaxoSmithKline: Employment. Lin:GlaxoSmithKline: Employment, Equity Ownership. Winter:GlaxoSmithKline: Employment, Equity Ownership. Goldstein:GlaxoSmithKline: Employment, Equity Ownership. Jewell:GlaxoSmithKline: Employment, Equity Ownership. Lisby:Genmab: Employment, Equity Ownership.

Description: Background Despite recent advances in molecular profiling of diffuse large B-cell lymphoma (DLBCL), only a few biomarkers currently have an impact on treatment. In a previous study we could observe the heterogeneity of DLBCL in the plasma immunoprofile from DLBCL patients, by use of recombinant antibody microarrays. By unsupervised hierarchical clustering, an immunoprofile of 23 plasma proteins could divide patients into two subgroups with significantly different overall survival (OS). In this study we aimed to expand the immunoprofiling with longitudinal plasma samples from high risk DLBCL patients, to search for novel prognostic and potentially predictive markers. Material and Methods Plasma samples from 126 high risk DLBCL patients included in a phase II clinical trial of the Nordic Lymphoma Group (CRY04) were collected at diagnosis, during and after treatment, and in the event of progression or relapse. Inclusion criteria were age 18-65 years, primary diffuse large B-cell lymphoma without CNS involvement, age-adjusted International Prognostic Index (aaIPI) 2-3 and WHO performance score 0-3. Six courses of R-CHOEP-14 were given followed by systemic CNS prophylaxis with one course of high-dose cytarabine and one course of high-dose methotrexate. Additionally, age and sex matched controls (n = 40) were included. Two hundred and eighty-three recombinant scFv antibody fragments directed against 97 known serum antigens, mainly immunoregulatory proteins, were selected from phage display libraries to be used in the antibody microarrays. Results Immunoprofiles distinguishing DLBCL patients from healthy controls were identified. Furthermore, a different protein expression was found in patients with aaIPI 3 versus 2 as well as in patients with shorter versus longer progression free survival (PFS). The kinases CDK-2 and BTK were upregulated both in patients with higher aaIPI score and in patients with shorter PFS. Comparing the patients who developed disease progression with those who did not, seven differentially deregulated proteins were identified including the phosphatase PTP-1B, the chemokine MCP-1, and the cytokines IL-4, IL-6 and IL-12, all previously implicated in B-cell lymphoma pathogenesis. In addition, results showed that subdivision of patients according to immunoprofile as defined in our previous study, could be performed also in the present patient cohort. The immunoprofile comprises T-helper (TH)1 cytokines, TH2 cytokines, complement proteins, chemokines, enzymes, and membrane proteins. The 23 included proteins are β-galactosidase, C1 esterase inhibitor, C4, C5, Cystatin C, Eotaxin, GLP-1 R, GM-CSF, HLA-DR/DP, IgM, IL-1ra, IL- 2, IL- 3, IL- 6, IL- 10, IL- 12, Leptin, MCP-1, MCP-3, Mucin-1, PSA, TNF-α, and TNF-β. Although not significant, a potential association to PFS and OS could be observed between the two generated subgroups. Finally, expression of IL-10 was shown to improve the prognostic value of aaIPI regarding both PFS and OS. Conclusion Protein expression profiling of plasma from high risk DLBCL patients provided novel insights into the biology of DLBCL. New candidate prognostic markers were identified which could potentially guide us in the prediction of outcome and in the choice of treatment for DLBCL patients. However, as this study was aimed for discovery, the findings must be validated in future studies with independent patient cohorts. Disclosures No relevant conflicts of interest to declare.