Citations

Abstract Purpose: Regulatory T-cells (Treg) are essential to Tregs homeostasis and modulate the antitumor immune response in patients with lymphoma. However, the biology and prognostic impact of Tregs in splenic marginal zone lymphoma (SMZL) have not been studied. Experimental Design: Biopsy specimens from 24 patients with SMZL and 12 reactive spleens (rSP) from individuals without lymphoma were analyzed by using CITE-seq (cellular indexing of transcriptomes and epitopes by sequencing), CyTOF (mass cytometry) analysis, and flow cytometry to explore the phenotype, transcriptomic profile, and clinical significance of intratumoral Tregs and their subsets. The biological characteristics and cell signaling pathways of intratumoral Treg subsets were confirmed by in vitro functional assays. Results: We found that Tregs are more abundant in SMZL patients' spleens than rSP, and Tregs from patients with SMZL and rSP can be separated into CD161+Treg and CD26+Treg subsets. CD161+Tregs are increased in SMZL but have dysregulated immune function. We found that CD161+Treg and CD26+Tregs have unique gene expression and phenotypic profiles and are differentially correlated with patient outcomes. Specifically, increased CD161+Tregs are significantly associated with a favorable prognosis in patients with SMZL, whereas CD26+Tregs are associated with a poor prognosis. Furthermore, activation of the IL2/STAT5 pathway contributes to the induction of CD26+Tregs and can be reversed by STAT5 inhibition. Conclusions: IL2/STAT5-mediated expansion of CD26+Tregs contributes to a poor clinical outcome in SMZL and may represent a therapeutic opportunity in this disease.

Abstract Background: DLBCL is the most common non-Hodgkin lymphoma subtype in western countries and is clinically heterogeneous. Gene expression profiling has identified two major biologically distinctive DLBCL subtypes defined by their cell-of-origin (COO): germinal center B-cell (GCB) - characterized by BCL2 rearrangement and C-REL amplification, and activated B-cell (ABC) - characterized by constitutive activation of the NF-kB pathway. We evaluated putative DLBCL risk factors for etiologic heterogeneity as defined by COO. Methods: We used a clinic-based study of newly diagnosed NHL cases and frequency matched controls, enrolled from 2002-2014, with a total of 687 DLBCL cases and 2253 controls for this analysis. Using formalin-fixed, paraffin-embedded tumor tissue, we determined COO by either digital expression profiling (NanoString; classified as GCB, ABC, undetermined) or clinically using the Hans algorithm (immunohistochemical markers; classified as GCB, non-GCB, undetermined). Integrating these two sources of COO data, there were 271 GCB, 170 non-GCB, and 246 undetermined/missing (mainly due to lack of tissue) cases. Risk factor data were collected from self-administered questionnaires and included family history of hematologic malignancy; history of atopy, allergy, asthma, eczema, or autoimmune disease; blood transfusion; vaccination history; regular low-dose aspirin use; body mass index (BMI) 2 years prior diagnosis/enrollment; smoking history; alcohol consumption (never/current/former); leisure-time physical activity; and recreational sun exposure. Polytomous logistic regression was used to estimate odds ratios (ORs) and 95% confidence intervals for COO subtypes, adjusted for age, sex and residence. Results: The median age of cases was 64 years with 55% male; the median age of controls was 64 years and 53% male. The association with BMI (30+ vs. 18.5-24.9 kg/m2) was stronger for non-GCB (OR=2.21; 1.40-3.49) than for GCB (OR=1.70; 1.21-2.40) DLBCL (P-heterogeneity 0.02), while a family history of hematologic malignancy was associated with non-GCB (OR=2.27; 1.47-3.50) but not GCB (OR=1.30; 0.86-1.96) DLBCL, although the P-heterogeneity was not significant (P=0.13). For GCB, there was an inverse association with former (vs. never) alcohol use for GCB (OR=0.57; 0.39-0.83) but not for non-GCB (OR=0.98; 0.61-1.57) DLBCL (P-heterogeneity 0.007), and an inverse association with regular use of low-dose aspirin for GCB (OR=0.64; 0.47-0.87) but not for non-GCB (OR=0.88; 0.61-1.25) DLBCL (P-heterogeneity 0.05). No other risk factors evaluated showed important heterogeneity by COO. Conclusions: These initial results suggest most DLBCL risk factors do not show etiologic heterogeneity by DLBCL COO, although associations with BMI and family history with non-GCB DLBCL and alcohol and low-dose aspirin use with GCB DLBCL warrant follow-up. Citation Format: Geffen Kleinstern, Dennis P. Robinson, Lisa M. Rimsza, Melissa C. Larson, Rebecca L. King, Grzegorz S. Nowakowski, Carrie A. Thompson, Stephen M. Ansell, Matthew J. Maurer, Andrew L. Feldman, Susan L. Slager, Anne J. Novak, Thomas M. Habermann, James R. Cerhan. Evaluation of etiologic heterogeneity for risk of diffuse large B-cell lymphoma (DLBCL) subtype defined by cell-of-origin [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 741.

Abstract IgM MGUS and Waldenstrom Macroglobulinemia (WM) represent a disease spectrum with highly dissimilar therapeutic management ranging from observation to chemoimmunotherapy. The current classification relies solely on clinical features and does not explain the heterogeneity that exists within each of these conditions. To shed light on the biology that may account for the clinical differences, we used bone marrow (BM) CD19+/CD138+ sorted cells and matched BM plasma from 32 patients (pts) (7 IgM MGUS, 25 WM) and 5 healthy controls to perform the first comprehensive multiomics approach including whole exome sequencing, RNA seq, proteomics and metabolomics. Applying principal component analysis to gene expression profiling, most of WM pts clustered together, while a small subset of them grouped separately with MGUS pts, suggesting a biologic dichotomy within WM. The healthy controls formed a distinct group from most WM and MGUS. We then applied a non-negative matrix factorization consensus clustering to the gene expression data and identified three robust clusters. Cluster 1 (C1) included only WM pts, cluster 2 (C2) included both WM and MGUS pts, and cluster 3 (C3) included all controls as well as a small number of WM and MGUS pts. When mutations commonly identified in WM were analyzed, there was no difference among the three groups (excluding controls) in mutation burden of MYD88 L265P and CXCR4. Interestingly, aberrant expression of TNFAIP3 seemed a distinct feature of C1 as deletion of 6q (which encodes for TNFAIP3) and TNFAIP3 mutations were each significantly enriched in C1 (47%) compared to C2 (0%) and C3 (20%; p=0.04). Individual clusters associated with specific transcriptional signature and clinical features. While C1 displayed enrichment of cell growth, downregulation of inflammatory pathways (eg IL6 and IL8 signaling) and aggressive behavior, C2 showed increased inflammatory signaling and cell survival with indolent behavior. C3 had an intermediate feature with combined proliferative and inflammatory signatures. In accordance with the transcriptomic data, the hallmark of C1 was upregulation of proliferation-associated proteins (eg AKT, MAPK) and downregulation of inflammatory proteins (eg IL4, IL10) while the opposite was observed in C2. Once more, C3 confirmed intermediate features with combined upregulation of proliferation and inflammatory proteins. The metabolism was rewired towards fatty acid catabolism in C1, glycolysis in C2 and anabolism in C3. Accordingly, C1 showed undetectable concentration of 3-hydroxybutyric acid as opposed to C2 which had increased levels of malic and lactic acids, as end products of fatty acid oxidation and glycolysis respectively. Those metabolites had intermediate levels in C3. In conclusion, we identified three molecular clusters with distinct clinical, proteomic and metabolomic features, suggesting a potential biologic classification that may have therapeutic implications. Citation Format: Patrizia Mondello, Jonas Paludo, Joseph Novak, Kerstin Wenzl, Shahrzad Jalali, Jordan Krull, Esteban Braggio, Surendra Dasari, Michelle Manske, Jithma Abeykoon, Asher Chanan-Khan, Robert Kyle, Morie Gertz, Zhi Zhang Yang, Anne Novak, Stephen Ansell. Molecular clusters and functional drivers of IgM monoclonal gammopathies [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3487.

Abstract Previous studies have found that the prevalence of subtypes of T cells may be associated with patient outcomes, but a comprehensive analysis of the immune profile in the TM of FL has not been done. In the present study, we identified groups of FL patients with discretely unique TMEs and determined whether the T-cell phenotypes in the TME differed among groups. Using a cohort of 82 FL patients with biopsy specimens collected before treatment, we defined the type of TME based on the content of major lineages (T, B, monocytes/macrophages and NK cells) determined by CyTOF analysis. Hierarchical clustering of this cohort stratified patients into 4 groups with different TME: group 1 (G1) included patients with a high percentage of monocyte/macrophages/NK cells; patients from G2 and G3 were enriched for intratumoral T and B cells, respectively. Patients with intermediate numbers of T and B cells were included in G4. CITRUS analysis revealed that T-cell clusters with phenotypes expressing KLRG1, CD57, PD-1dim and rich in TEMRA cells were significantly more abundant in G1 when compared to G2, G3 or G4. In contrast, T-cell clusters with phenotypes containing CD127, CD45RA, CCR7 or PD-1high cells were significantly less abundant in G1 when compared to other groups. When compared to G3 or G4, 2 classes of T cell clusters, all from CD4+ T cells, were significantly more (CD127+KLRG1-) or less (CD57+PD-1high) abundant in G2, respectively. Clusters with a phenotype rich in short-lived effector cells (SLEC) (from CD8+) were upregulated in G3 when compared to G4. These results suggest that patient groups with distinct TME exhibited variable T-cell phenotypes. To determine the role of T-cell differentiation in predicting patient outcome in FL, we identified T-cell subsets using tSNE plots based on the expression of T-cell maturation markers. We identified 18 subsets (S1-S18) of T cells (11 from CD4, 6 from CD8 and 1 from CD4-CD8-) in each sample. Four subsets - S4 (CD4+TN), S5 (CD4+TCM), S7 (CD8+TN) and S10 (CD8+MEPCs) - were considered to be naïve cells or cells in the early stages of differentiation. Four additional subsets - S8 (CD57+TFH), S9 (CD4+TEXH), S13 (CD8+SLECs) and S17 (CD4+PD-1+Treg) - were memory cells with expression of surface markers indicting late-stage differentiation. We found that the 4 subsets (S4: p=0.08, S5: p=0.01, S7: p=0.04 and S10: p<0.01) with an early-stage phenotype were associated with a favorable clinical outcome. In contrast, subsets (S8: p=0.02, S9: p=0.04, S13: p=0.07 and S17: p=0.06) with a late-stage differentiation phenotype had an unfavorable survival. Supporting this finding, we observed that increased numbers of CD45RA+ T cells correlated with a favorable survival. These results indicate that the differentiation stage may determine the role of T cells in predicting patient outcome in FL. Citation Format: Zhi-Zhang Yang, Hyo Jin Kim, Hongyan Wu, Xinyi Tang, Jordan Krull, Patrizia Mondello, Jose Villasboas, Anne Novak, Stephen Ansell. T-cell phenotype and differentiation vary in the tumor microenvironment of follicular lymphoma and are associated with patient outcome [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 2529.

7562 Background: Lymphoma is the 6th most common cancer and a top 10 cause of cancer deaths. To date, there are no blood test approaches to population screening, and accurate surveillance markers are limited. Methylated DNA markers (MDMs) are broadly informative for early detection of cancer but have not been extensively studied for lymphomas. We sought to discover and validate MDMs in DNA extracted from lymphoma tissue, then test these MDMs in archival blood plasma specimens. Methods: Reduced representation bisulfite sequencing (RRBS) was performed on DNA extracted from a discovery set of frozen tissues of classic Hodgkin lymphoma (CHL, n=24), non-Hodgkin lymphoma (NHL, n=78: T-cell-TCL (8); diffuse large B-cell-DLBCL (18); follicular-FL (12); mantle cell-MCL (20); marginal zone-MZL (15)), and lymphoma cell line samples (27); controls included benign lymph node (n=11) and healthy donor buffy coat (n=30). 30 MDMs were ranked by fold-change and AUC and used to design methylation-specific PCR assays for biological validation on independent DNA samples extracted from FFPE tissue from 13 normal lymph node and 63 lymphoma samples and 36 buffy coats from healthy patients. Target enrichment long-probe quantitative-amplified signal assays were developed for 16 MDMs and then assayed in plasma-extracted, bisulfite-converted DNA samples from 390 independent treatment-naïve lymphoma patients (100 CHL and 290 NHL:100 DLBCL, 73 FL, 41 MCL, 41 MZL, 35 TCL), and 210 controls without cancer. Lymphoma plasma samples and 159 controls were provided by the Lymphoma SPORE (CA97274); 51 controls came from a 7-county population archive. Classification of lymphoma cases vs controls was modeled with random forest method and cross-validated across 500 bootstrap samples of each dataset. Results: For MDMs tested in DNA from independent biological validation samples, cross-validated random forest models identified 60/63 cases (95% sensitivity) and 12/13 normal controls (92% specificity). In plasmas, a panel of 16 MDMs ( ZNF503, VWA5B1, HOXA9_5195, GABRG3, ITGA5, MAX_chr17_793, BNC1_2407, CDK20, MAX_chr4_184, TPBG, DNAH14, SYT2, CACNG8, FAM110B, and NRN1) detected 78% (95% CI, 74-82%) of lymphoma cases at 90% specificity. Excluding MZL and TCL, sensitivity increased to 84% (80-88%) including 26/49 (53% (38-67%)) of stage I and 59/71 (83% (71-90%)) of stage II cases. Conclusions: MDMs show promise to detect lymphoma. These markers could be evaluated as part of multicancer early detection testing and could also be evaluated as response markers to treatment and subsequent surveillance.[Table: see text]

Abstract Introduction: Low-grade B-cell lymphomas (LGBCL), aside from follicular lymphoma and chronic lymphocytic leukemia/small lymphocytic lymphoma, account for approximately 10% of B-cell non-Hodgkin lymphomas and consist of several subtypes. While a majority of LGBCL cases have an overall favorable prognosis, we have previously shown that cases who have an event (relapse or progression, transformation, or re-treatment) within 24 months of diagnosis (EFS24) have an inferior overall survival (OS) compared to those achieving EFS24 (Tracy et al., AJH 2019;94:658-66). However, the underlying biological characteristics associated with early failure and aggressive disease across LGBCL subtypes are unknown. In this study, we used matched transcriptomic, genomic, and immune profiling data from LGBCL cases, the largest cohort to date, and asked whether there were unique biological phenotypes across different LGBCL subtypes and whether we could identify signatures associated with aggressive LGBCL. Validation of the prognostic utility of this signature was performed on a previously published, independent cohort of 63 pre-treatment LGBCL cases. Methods: Tumors from 64 newly diagnosed LGBCL patients from the Molecular Epidemiology Resource of the University of Iowa/Mayo Clinic Lymphoma Specialized Program of Research Excellence were included in this study (SMZL (n = 48), NMZL (n = 6), LPL (n = 5), B-NOS (n = 3), EMZL (n = 2)). RNA sequencing (RNAseq) data from 61 LGBCL tumors and 5 benign CD19+CD27+ memory B samples was subjected to NMF clustering to define groups. Differential expression and pathway analysis were used to identify biological characteristics of each cluster. CIBERSORT was used to identify immune cells in the tumor microenvironment. Whole exome sequencing (WES) was performed on 61 tumor-normal pairs. Singscore was used to assign a single score per patient representing gene expression of the survival-associated transcriptomic signature identified in this study. Results: NMF analysis of RNAseq data identified 5 clusters of patients, denoted LGBCL1-5 (Fig 1A). Patients from the same diagnostic subtype did not exclusively cluster together, with all LGBCL clusters comprised of patients from multiple subtypes (Fig 1B). Exploring the association between patient cluster and outcome, we observed significantly inferior event-free survival (EFS) (HR 2.24; 95% CI 1.01-4.98) and overall survival (OS) (HR 5.59; 95% CI 2.00-15.63) in LGBCL5 patients compared to LGBCL1-4 (Fig 1C). In addition, 80% of the transformation cases in our cohort were classified as LGBCL5 (Fig 1D). Differential expression and pathway analysis showed distinct processes significantly upregulated in each cluster (FDR < 0.05), with LGBCL5 demonstrating enrichment of cell cycle and mitosis pathways. CIBERSORT identified increased immune cell content in LGBCL3 and LGBCL5 compared to other clusters, with high frequencies of mast cells in both (p = 0.0002), increased CD8 T cells in LGBCL3 (p < 0.0001), and increased T follicular helper cells in LGBCL5 (p = 0.004). WES identified previously reported alterations in NOTCH, NFkB, and chromatin remodeling pathways and novel variants in LGBCL, including mutations in HNRNPK, CLTC, HLA-A, HLA-B and HLA-C. Assessment of alterations by cluster showed significant enrichment of TNFAIP3 (OR 5.54; 95% CI 1.20-28.14) and BCL2 alterations (OR 5.49; 95% CI 1.07-32.02) in LGBCL5 cluster. Finally, we identified a cell cycle-related transcriptomic signature of 108 genes upregulated in LGBCL5 and EFS24 failure cases. Cases with high expression of this signature showed significantly inferior EFS (HR 14.25; 95% CI 4.90-41.38) and OS (HR 7.82; 95% CI 2.40-25.44) compared to cases with low expression in our discovery cohort. This observation was reproduced in an independent validation cohort, where patients with high expression of this signature demonstrated significantly inferior EFS (HR 5.70; 95% CI 1.49-21.79) and OS (HR 10.07; 95% CI 2.00-50.61). Conclusions: In this study, we are the first to define mechanisms of pathogenesis in LGBCL with shared transcriptomic, genomic, and immune profiles present across LGBCL subtypes. We then further defined a gene expression signature associated with inferior patient outcome, with application of this signature to an independent validation cohort demonstrating proof of concept and utility of this signature as a prognostic marker in LGBCL patients. Figure 1 Figure 1. Disclosures Maurer: Morphosys: Membership on an entity's Board of Directors or advisory committees, Research Funding; Genentech: Research Funding; Kite Pharma: Membership on an entity's Board of Directors or advisory committees; Pfizer: Membership on an entity's Board of Directors or advisory committees; BMS: Research Funding; Nanostring: Research Funding. Paludo: Karyopharm: Research Funding. Habermann: Tess Therapeutics: Other: Data Monitoring Committee; Morphosys: Other: Scientific Advisory Board; Incyte: Other: Scientific Advisory Board; Seagen: Other: Data Monitoring Committee; Loxo Oncology: Other: Scientific Advisory Board; Eli Lilly & Co.,: Other: Scientific Advisor. Link: MEI: Consultancy; Genentech/Roche: Consultancy, Research Funding; Novartis, Jannsen: Research Funding. Rimsza: NanoString Technologies: Other: Fee-for-service contract. Ansell: Bristol Myers Squibb, ADC Therapeutics, Seattle Genetics, Regeneron, Affimed, AI Therapeutics, Pfizer, Trillium and Takeda: Research Funding. Cerhan: Genentech: Research Funding; Regeneron Genetics Center: Other: Research Collaboration; Celgene/BMS: Other: Connect Lymphoma Scientific Steering Committee, Research Funding; NanoString: Research Funding. Novak: Celgene/BMS: Research Funding.

Abstract The tumor microenvironment plays a crucial role in mediating the tumor immune response, thereby affecting patient outcomes in follicular lymphoma (FL). Using CyTOF, we analyzed single cell suspensions created from pre-treatment biopsies in a cohort of 82 FL patients and retrospectively extracted patient clinical parameters. Hierarchical analysis of the tumor microenvironment composition stratified these patients into 4 groups: group 1 represents patients with high percentage of monocyte/macrophages/NK cells; patients from group 2 and 3 are rich of intratumoral T and B cells, respectively. Patients with intermediate number of T and B cells are classed into group 4. Intratumoral T cells from these 4 groups exhibited distinct phenotypical characteristics. Group 1 was enriched with KLRG1 +CD8 + T cells, while group 4 showed increased number of CXCR3 - T EM and CD4 + T N cells. Patients from group 2 and 3 featured with increased number of ICOS + T reg cells and CD57 + T FH cells, respectively. We observed that higher numbers of T cells significantly correlated with early disease stage, lack of B-symptoms, event-free survival at 24 months (EFS24) and a favorable overall and EFS survival in patients with histological grade 1 and 2. However, this association was not seen in patients with histological grade 3a and b, suggesting that T cell-mediated anti-tumor immunity is preserved in low grade patients, but overridden in higher grade patients. CITRUS analysis revealed that a cluster that is CD57 + and PD-1 high was significantly more abundant in patients who had disease progression than patients who had a complete response to therapy, suggesting a role of CD57 + T FH cells in promoting malignant cell growth in FL. Consistent with this finding, we observed that higher number of CD57 + T FH cells correlated with an inferior survival in FL. Using CITE-seq technology, we found that CD57 + T FH cells exhibited a substantially different transcriptome when compared to CD57 - T FH cells. Genes that were differentially upregulated in CD57 + T FH cells when compared to CD57 - T FH cells included genes involved in cell survival, compromised inflammatory response, and metabolism activation including GZMK, CCL4, CST7, DUSP2, LGALS3, CYTOR, CHI3L2, SYNE2, CXCL13, CD27 and FABP5. Taken together, our results indicate different tumor microenvironments among FL patient groups that is associated with variable T-cell phenotype. We found that CD57 + T FH cells play an important role in predicting disease progression and patient prognosis in FL. Figure 1 Figure 1. Disclosures Novak: Celgene/BMS: Research Funding. Ansell: Bristol Myers Squibb, ADC Therapeutics, Seattle Genetics, Regeneron, Affimed, AI Therapeutics, Pfizer, Trillium and Takeda: Research Funding.

Abstract Introduction: Central nervous system (CNS) relapse is an uncommon but devastating complication occurring in about 2-10% of patients with diffuse large B-cell lymphoma (DLBCL). The CNS-IPI is a validated prognostic tool used to identify patients at high risk of CNS relapse. MYC and BCL2/BCL6 rearrangements (double-hit lymphoma- DHL) and non-GCB (NGCB) cell of origin may be at higher risk of CNS relapse. However, the ability of these molecular data to refine CNS-IPI risk prediction has not been assessed. In this study, we aimed to describe the impact of DHL and NGCB cell of origin in the incidence of CNS relapse in patients with newly diagnosed DLBCL. Methods: A retrospective review was conducted to identify patients with newly diagnosed DLBCL from 01/01/2002 - 06/30/2015 using the University of Iowa/Mayo Clinic Lymphoma Molecular Epidemiology Resource database. Patients with primary CNS lymphoma or with CNS involvement at diagnosis were excluded. All patients received chemoimmunotherapy. Clinical, biological, laboratory and radiological data were abstracted including the occurrence of CNS relapse. CNS IPI risk score categories were defined as low- (≤1) intermediate- (2-3) and high-risk (≥4) as originally described (Schmitz et al, JCO 2016). Cell of origin (COO) classification into GCB and non-GCB subtypes was performed using Hans algorithm and/or NanoString data. Characteristics between groups were compared using Kruskal-Wallis or Chi-squared test as appropriate. Survival analysis was performed using Cox proportional hazards model and the Kaplan Meier method. Statistical analysis was performed using SAS version 9.4M5 and R version 4.0.3. Results: We identified a total of 1,360 patients with newly diagnosed DLBCL. Patients had a median age of 62 years (range 18-93 years), were predominantly male (56%), had a performance status <2 (84%), had stage III/IV disease (61%) and an elevated LDH (56%) at diagnosis. Fluorescent in situ hybridization data was present for 65% of patients. Among patients with available data, DHL and NGCG COO represented 2.8% and 36% of the cohort, respectively. CNS IPI high- and intermediate-risk categories were present in 12.4% and 52% of the patients, respectively. R-CHOP was the most common upfront chemotherapy (n=980, 72%). The median follow-up for the entire cohort was 84.9 months (IQR: 58.9-120.9). Overall, 50 patients (3.6%) had a CNS relapse during the study period. Lymphoma relapsed in the CNS in 3.0% (95% CI:2.1-4.4), 4.7% (95% CI: 3.4-6.3) and 4.9% (95% CI: 3.7-6.6)% of patients with CNS-IPI score of intermediate/high risk at 1, 3 and 5 years following the diagnosis of DLBCL. In patients with low risk CNS-IPI, 1.1% (95% CI: 0.4-2.5), 1.7% (95% CI: 0.9-3.4) and 1.7% (95% CI: 0.9-3.4) had a CNS relapse at 1, 3 and 5 years following DLBC diagnosis. Patients with CNS relapse were more likely to have elevated LDH (76% vs 56%, p=0.008), advanced stage (78% vs 61%, p=0.01), adrenal involvement (8% vs 2.4%, p=0.014) and have a CNS IPI intermediate risk (68% vs 51%), and high risk (16% vs 12.3%, p=0.014) at the time of initial diagnosis compared to patients without CNS relapse. Patients with CNS relapse had similar rates of DHL (4% vs 2.7%, p=0.17) and NGCB cell of origin (42% vs 36.2%, p=0.52) compared to patients without CNS relapse. In the multivariable analysis, the strongest predictors for CNS relapse were CNS IPI high risk (Hazard Ratio [HR]= 3.6, 95%CI= 1.4-9.7) and CNS IPI intermediate risk (HR= 3.15, 95%CI= 1.46-6.81). DHL and NGCB cell of origin were not associated with a significantly higher risk of CNS relapse (HR= 1.79, [95%CI=0.42-7.59] and HR= 1.26, [95%CI=0.62-2.57], respectively). Conclusion: CNS relapse was an uncommon occurrence in patients with DLBCL who were treated with chemoimmunotherapy. CNS IPI intermediate- and high-risk scores remain the strongest predictors of CNS relapse in patients with DLBCL after adjusting to biological high-risk variables including DHL and NGCB cell of origin. Larger studies are required to validate our findings due to the low incidence of CNS relapse in our cohort. Figure 1 Figure 1. Disclosures Farooq: Kite, a Gilead Company: Honoraria. Maurer: Kite Pharma: Membership on an entity's Board of Directors or advisory committees; Morphosys: Membership on an entity's Board of Directors or advisory committees, Research Funding; Celgene: Research Funding; Genentech: Research Funding; Pfizer: Consultancy, Membership on an entity's Board of Directors or advisory committees; Nanostring: Research Funding. Witzig: Karyopharm Therapeutics, Celgene/BMS, Incyte, Epizyme: Consultancy, Membership on an entity's Board of Directors or advisory committees; Celgene/BMS, Acerta Pharma, Kura Oncology, Acrotech Biopharma, Karyopharm Therapeutics: Research Funding. Habermann: Tess Therapeutics: Other: Data Monitoring Committee; Seagen: Other: Data Monitoring Committee; Incyte: Other: Scientific Advisory Board; Morphosys: Other: Scientific Advisory Board; Loxo Oncology: Other: Scientific Advisory Board; Eli Lilly & Co.,: Other: Scientific Advisor. Novak: Celgene/BMS: Research Funding.

Abstract Background: Follicular lymphoma (FL) exhibits significant clinical, cellular, molecular, and genetic heterogeneity. Our understanding of FL biology and molecular classifications of FL, to date, has largely been driven by pathologic classification (Grade 1-3b), genetic classification (m7-FLIPI), or gene expression profiling (IR-1/2; Huet-23), along with limited studies on small cohorts or targeted panels. In order to further understand the biological underpinnings and complexity of FL, large-scale and integrated whole exome sequencing (WES) and RNA sequencing (RNAseq) studies are needed. Using a highly-annotated cohort of 93 FL tumors with matched RNAseq, WES, and CyTOF data, we have explored the transcriptomic signature of purified FL B cells and identified unique molecular subsets that are defined by distinct pathway activation, immune content, and genomic signatures. Methods: Frozen cell suspensions from 93 untreated FL (Grade 1-3b) patients' tumor biopsies, enrolled in the University of Iowa/Mayo Clinic Lymphoma SPORE, were used for the study. DNA was isolated from whole tumor cell suspensions, and RNA was isolated from both whole tumor and B cell-enriched cell suspensions. RNAseq and WES were performed in the Mayo Clinic Genome Analysis Core. RNAseq and WES data were processed using a standard pipeline and novel driver genes were identified using Chasm+ driver analysis. Copy number variants were identified from WES data using GISTIC 2.0. NMF clustering and single sample gene set testing for B cell lineage and tumor microenvironment (TME) signatures were performed in R using the NMF and singscore packages. Results: Unsupervised clustering of RNAseq data identified three distinct expression programs which separated patient B cell samples into 3 groups: Group 1 (G1, n=20), Group 2 (G2, n=24), Group 3 (G3, n=43). While no clinical attributes were defined by any single group, G1 consisted of cases that had less aggressive characteristics (63% Stage I-II, 79% FLIPI 0-1). To identify unique transcriptional pathways driving the three expression programs, we scored gene level contributions to NMF expression programs and employed gene set enrichment analysis. This revealed significant pathway association with type-I IFN signaling (G1), DNA repair and stress response (G2), and epigenetic modulation (G3) as differentiating programs between the 3 groups (FDR q<0.001). VIPER master regulator activity inferencing revealed that these pathways were likely being controlled by differential activity in NF-kB, IRFs, STAT1, BCL6, and FOXO1. Each program significantly enriched for, but were not defined by, portions of specific germinal center programs, such as pre-memory (G1), light-zone-to-dark-zone transition (G2), and a pre-light-zone intermediate (G3). We next assessed the connection between B cell programs and the tumor microenvironment (TME) using available paired CyTOF data on 67 cases, which revealed an active TME in G1, with an abundance of CD8 T cell and NK cell populations, a wide variety of immune content in G2 that consisted mostly of Tfh and myeloid cells, and a poorly populated immune compartment in G3 compared to G1 and G2. Finally, somatic driver mutations and copy number alterations from WES were identified and associated with the three clusters. The three groups distinguished themselves by significant enrichment of copy number alterations (TNFAIP3-loss , 1q23-gain, 1q32-gain) in G2, while 10q-loss and mutations in BCL2 and chromatin modifiers (KMT2D and CREBBP) enriched in G3. G1, overall, had lower alteration burden and had weak associations with any specific alterations, suggesting an alternative mechanism for driving the G1 program. Conclusion: In this study, we have identified three unique FL tumor B cell groups, defined by specific transcriptional programs. Pathways such as inflammation, DNA damage response, and chromatin modification contribute most to the differences between B cell samples and group membership. Additionally, each program associated with specific genetic events as well as TME composition, highlighting potential drivers of these B cell states. This study improves the understanding of the mechanisms driving FL tumors and motivates further investigation into transcriptional consequences of genetic events as well as potential tumor intrinsic factors that may influence the TME. Figure 1 Figure 1. Disclosures Maurer: BMS: Research Funding; Genentech: Research Funding; Morphosys: Membership on an entity's Board of Directors or advisory committees, Research Funding; Kite Pharma: Membership on an entity's Board of Directors or advisory committees; Pfizer: Membership on an entity's Board of Directors or advisory committees; Nanostring: Research Funding. Rimsza: NanoString Technologies: Other: Fee-for-service contract. Link: MEI: Consultancy; Genentech/Roche: Consultancy, Research Funding; Novartis, Jannsen: Research Funding. Habermann: Tess Therapeutics: Other: Data Monitoring Committee; Seagen: Other: Data Monitoring Committee; Incyte: Other: Scientific Advisory Board; Morphosys: Other: Scientific Advisory Board; Loxo Oncology: Other: Scientific Advisory Board; Eli Lilly & Co.,: Other: Scientific Advisor. Ansell: Bristol Myers Squibb, ADC Therapeutics, Seattle Genetics, Regeneron, Affimed, AI Therapeutics, Pfizer, Trillium and Takeda: Research Funding. King: Celgene/BMS: Research Funding. Cerhan: Genentech: Research Funding; Regeneron Genetics Center: Other: Research Collaboration; Celgene/BMS: Other: Connect Lymphoma Scientific Steering Committee, Research Funding; NanoString: Research Funding. Novak: Celgene/BMS: Research Funding.

Abstract Background: Vaccinations have been hypothesized to play a role in lymphoma etiology, but there are few studies, mixed results, and limited data on lymphoma subtypes. Herein, we investigate the association of vaccinations with risk of major lymphoma subtypes. Methods: We studied 2,461 lymphoma cases and 2,253 controls enrolled from 2002 to 2014. Participants self-reported history of vaccinations against hepatitis A, hepatitis B, yellow fever, and influenza. Polytomous logistic regression was used to estimate OR and 95% confidence intervals (CI), adjusting for potential confounders. Results: After multivariable adjustment, vaccination against influenza was inversely associated with lymphoma (OR = 0.82; 95% CI, 0.66–1.02), which was stronger for last vaccination 1+ years before enrollment (OR = 0.71; 95% CI, 0.56–0.91) and for >5 influenza vaccinations (OR = 0.56; 95% CI, 0.46–0.68). Ever vaccination against hepatitis A (OR = 0.81; 95% CI, 0.66–1.00) but not hepatitis B (OR = 0.97; 95% CI, 0.81–1.18) was associated with lymphoma risk, although more recent vaccinations were inversely associated with lymphoma risk for both hepatitis A (<6 years before enrollment, OR = 0.56; 95% CI, 0.40–0.77) and hepatitis B (<9 years before enrollment, OR = 0.72; 95% CI, 0.55–0.93). Ever vaccination against yellow fever was inversely associated with risk (OR = 0.73; 95% CI, 0.55–0.96), and this did not vary by time since last vaccination. Although there was no overall statistical evidence for heterogeneity of vaccination history by lymphoma subtype, the only statistically significant inverse associations were observed for influenza and yellow fever vaccinations with diffuse large B-cell and follicular lymphoma. Conclusions: Selected vaccinations were inversely associated with lymphoma risk, with time since last vaccination relevant for some of these vaccines. Impact: Vaccinations against hepatitis A, hepatitis B, yellow fever, and influenza are unlikely to increase lymphoma risk.