Ansell, Stephen M.

Abstract Rituximab is a chimeric murine/human monoclonal antibody that binds to CD20 on B-lymphocytes. Binding of the Fab domain to B-cells directly induces apoptosis, while the Fc domain recruits immune effector functions to mediate cell lysis. Because rituximab therapy alone does not result in durable responses for all patients, combination therapies have been explored. We have previously shown in a phase I study that interleukin-12 (IL-12), which facilitates cytolytic T-cell responses and enhances the lytic activity of NK cells, can be safely combined with rituximab and that IL-12 significantly upregulated gamma interferon, CXCL10 (inducible protein-10) and NK cell activity in the peripheral blood. To confirm whether IL-12 could augment the immune mediated cell lysis induced by rituximab, a subsequent randomized phase II study of the combination was performed in patients with B-cell lymphoma. While the combination of IL-12 and rituximab was well tolerated with acceptable toxicity, only moderate disease activity was seen and the response rate to the combination was similar to that seen with rituximab alone. Additionally, the sequential administration of IL-12 at the time of disease progression after treatment with rituximab did not result in any clinical responses. This study was therefore performed to determine potential biologic reasons for the lack of increased clinical efficacy when IL-12 was added to rituximab therapy in patients with B-cell non-Hodgkin lymphoma. Of the 52 patients treated on the phase II study, 8 patients had matched tumor biopsies and peripheral blood specimens obtained prior to therapy and again 2 weeks after treatment was started. Six of the patients were receiving IL-12 plus rituximab at the time the specimens were obtained while 2 were receiving rituximab alone. Gene expression array analysis using the Affymetrix U133 plus chip was performed on RNA isolated from cells from involved lymph nodes and from peripheral blood mononuclear cells. Specimens from the peripheral blood of patients who received IL-12 in combination with rituximab showed a greater than 5-fold increase in the expression of multiple genes known to be upregulated by IL-12 signaling including interferon gamma, CXCL10, IFIT2 and 4 (interferon-induced protein with tetratricopeptide repeats 2 and 4), IL-8 and CXCL2 (macrophage inflammatory protein-2). These increases in gene expression were not seen in the peripheral blood of patients who received rituximab alone. Furthermore, the significant changes seen in cells obtained from the peripheral blood were not seen in cells obtained from lymph nodes involved by lymphoma, despite the samples being obtained from the same patient on the same day. In the tumor specimens, the changes in gene expression involved none of the genes downstream of IL-12 signaling. Instead, many of the upregulated genes were associated with cell cycle and spindle checkpoint proteins suggesting ongoing tumor cell proliferation. In conclusion, while IL-12 significantly upregulated gene expression in the peripheral blood, the same changes were not seen in the tumor. This would suggest that systemically administered IL-12 may not be effectively delivered to the site of tumor involvement. This finding may explain the lack of additional clinical benefit when IL-12 was added to rituximab as therapy for patients with non-Hodgkin lymphoma.

Abstract Background: Local antitumor immunity—as assessed by quantifying tumor-infiltrating immune cells—is increasingly recognized as a critical factor influencing prognosis and therapy response in TNBC. However, an understanding of systemic antitumor immune responses occurring in peripheral blood, and their influence on prognosis and chemotherapy response has not been rigorously studied. Methods: Cytometry by time-of-flight (CyTOFTM, Fluidigm) was used to examine viably cryopreserved peripheral blood mononuclear cell (PBMC) suspensions prospectively collected from patients with early-stage TNBC prior to initiation of standard neoadjuvant paclitaxel followed by doxorubicin and cyclophosphamide (NACT) as part of the BEAUTY study [1]. Samples were stained using a panel of metal-tagged antibodies, recognizing 30 surface proteins optimized for immune monitoring of human peripheral blood. Differential abundance analysis of immune cell subsets was carried out to evaluate differences between patients who achieved pCR versus those with residual disease after NACT, and between patients with known luminal androgen receptor (LAR) versus basal TNBC subtypes defined by bulk tumor RNA sequencing. Results: Viably cryopreserved PBMC samples from 40 treatment-naive TNBC patients were available for analysis. The median age was 52 years (range 32 - 73), with 6 (15%) patients having tumors classified as LAR TNBC, and the remaining 34 (85%) as basal TNBC. Overall, 21 (53%) patients achieved pCR after NACT. After acquisition on the mass cytometer, the median yield per sample was 626,815 single-cell events (range 42,786 - 1,035,575), with a median percent debris of 13.7% (range 14 - 58). Across the 40 PBMC samples, the total yield was 23,507,094 single-cell events. The median frequencies of major circulating immune cell subsets across the 40 TNBC patients were: T cells 53.9% (range 25.4 - 71.3), with 33.4% CD4+ T cells (range 11.4 - 46.7) and 10.3% CD8+ T cells (range 5.8 - 19.9); B cells 10.8% (3.3 - 32.6), NK cells 8.6% (1.7 - 17.0) and monocytes 10.6% (2.7 - 29.8). Examining pre-treatment blood samples, patients with residual disease after NACT exhibited a higher median frequency of baseline CD14+CD16- classical monocytes (7.5% vs. 4.1%, p=0.025) and a lower frequency of terminally-differentiated effector memory cytotoxic (CD8+) T cells (0.6% vs. 1.7%, p=0.038) compared to patients who achieved pCR. Patients with LAR TNBC also exhibited a higher frequency of CD14+CD16- classical monocytes (11.5% vs 4.3%, p=0.058), and in addition exhibited a lower frequency of central memory CD4+ T cells (10.4% vs 15.2%, p=0.048). No difference in CD8+ T cells was seen by LAR status. Additional associations of peripheral blood immune cell subsets and classic tumor pathological features will be presented at the meeting. Conclusion: To our knowledge, this is the first study focused on TNBC to demonstrate variation in peripheral blood immune cell populations by molecular TNBC subtype (LAR vs. basal), and by chemotherapy response. A higher frequency of circulating classical monocytes—which can infiltrate into tissues and give rise to macrophages—appears to be detrimental; whereas a higher frequency of circulating antigen-experienced memory CD8+ T cells seems to be protective, suggesting a putative role of this cell subset in TNBC anti-tumoral immunity. Reference: [1] Goetz MP et al. JNCI 2017, PMID:28376176 Citation Format: Roberto A Leon-Ferre, Kaitlyn McGrath, Jodi M Carter, Krishna R Kalari, Vera J Suman, Richard Weinshilboum, Liewei Wang, Keith L Knutson, Stephen M Ansell, Judy C Boughey, J C Villasboas, Matthew P Goetz. Deep phenotyping using CyTOF identifies peripheral blood immune signatures associated with clinical outcomes and molecular subtypes in patients with early-stage triple negative breast cancer (TNBC) [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P5-04-09.

TPS7570 Background: The DIAL study is testing the efficacy of dual immunomodulation in patients with advanced B cell non-Hodgkin lymphoma (B-NHL). Developed by the Cancer Therapy Evaluation Program (CTEP), the trial combines the use of a programmed cell death protein 1 (PD-1) inhibitor (nivolumab) with an agonist of the CD27 receptor (varlilumab) in a randomized phase 2 design. CD27, a co-stimulatory receptor, regulates T cell activation through interaction with CD70. T cell exhaustion plays a major role in immune evasion in B-NHL. Varlilumab is an agonistic IgG1 monoclonal antibody that recognizes CD27 leading to prevention or reversal of exhaustion. Varlilumab also demonstrates direct anti-tumoral activity in xenograft models of human lymphoma cell lines via antibody-dependent cell-mediated cytotoxicity. Phase 1 data supports the safety and tolerability of single-agent varlilumab in advanced hematologic malignancies. We hypothesize that CD27 activation synergizes with PD-1 inhibition resulting in a superior anti-lymphoma effect compared to PD-1 blockade alone. The study will also evaluate the effect of these agents on tumor and immune cells using IHC, mass cytometry (CyTOF), multiplex ELISA, imaging mass cytometry, and whole exome sequencing. Methods: The trial is enrolling patients with advanced aggressive B-NHL. Standard inclusion criteria and prior treatment with at least 2 lines of standard therapy are required. Prior autologous stem cell transplant and/or chimeric antigen receptor (CAR) T cell therapy is allowed. Patients with active CNS disease are excluded. Eligible patients will be randomized to treatment with single-agent nivolumab (group 1) or dual immunotherapy with nivolumab and varlilumab. Group 1 is allowed to cross-over at the time of progression. Nivolumab will be administered intravenously (IV) every 2 weeks (240 mg) for 4 months followed by monthly dosing thereafter (480 mg). Varlilumab will be given IV every 4 weeks (3 mg/kg). Response assessment will be done by PET-CT scan every 12 weeks. Primary outcome is overall response rate (ORR) according to the LYRIC criteria. The trial will enroll 48 patients per arm, allowing 80% power to detect at least 20% increase in ORR in the experimental arm (group 2) assuming a 25% ORR in the control arm (group 1). The trial is registered and open to participation to members of ETCTN and EDDOP. Clinical trial information: NCT03038672.

7559 Background: Waldenström macroglobulinemia (WM) is a rare indolent lymphoma commonly treated with rituximab (R)-based therapy. The use of rituximab maintenance (mR) in WM is controversial. We present a case-control study of patients (pts) with WM treated with mR. Methods: Pts evaluated at Mayo Clinic, Rochester with active WM that received mR between 1/2000 & 6/2018 were included. Cases comprised pts who received mR following R-based induction as primary therapy. Cases were matched based on the time of diagnosis in 1:2 ratio with a control group treated with R-based primary induction therapy without mR. Time to event analyses were performed from initiation of R-based induction. Results: Of 776 pts with active WM, 42 (5%) cases received mR and 84 pts were selected as controls. The median follow-up and the proportion of high risk pts were comparable between the two cohorts (Table). Pts in the mR cohort show a trend toward longer time to next therapy (TTNT) and a significantly longer overall survival (OS) compared to the control group (Table). The R-based induction therapies were comparable in the two cohorts (p = 0.6). Median duration of mR was 1.9 yrs (95% CI 1.6-2) and mR was used most frequently every (q) 2 (range 1-6) months. Of the 42 mR pts, 25 (60%) received an R-based combination for induction and 17 (40%) received R monotherapy as induction. Five (12%) pts discontinued mR due to toxicity, infections were reported in 13 pts (31%) during mR therapy and 3 pts (7%) received IVIg infusions for recurrent infections. Conclusions: R-based induction followed by mR demonstrates a longer OS in WM compared to R-treated control population not receiving mR, albeit at a high rate of infections. Despite limitations of a retrospective study, with a heterogeneously treated cohort, these data add to the body of literature supporting Rituximab maintenance. Results from an ongoing randomized controlled trial are awaited. [Table: see text]

7554 Background: Predicting early clinical failure in patients with untreated FL is important but difficult. Lymphoma involvement of EN sites is better detected by FDG-PET/CT than CT alone, but PET parameters are not part of the usual predictive indices. We aimed to determine the incidence and patterns of spleen and EN disease using PET/CT, and learn if they are useful in predicting early clinical failure. Methods: PET/CT images from 613 cases of newly diagnosed FL between 2003 – 2016 were retrospectively reviewed. The location, pattern, and number of EN sites, as well as splenic involvement, were recorded. Associations with outcomes were assessed using event-free survival (EFS) and overall survival (OS). Results: 49% (301/613) of patients had PET/CT-detected EN involvement, and 28% (171/613) had spleen involvement. Presence of ≥2 EN sites, spleen, bone, or soft tissue involvement, as well as a multifocal on diffuse pattern of bone involvement by imaging, were univariate predictors of EFS; presence of ≥2 EN sites and bone involvement pattern were also predictive of OS. In a multivariate analysis with FLIPI-2 factors, spleen involvement, pattern of bone involvement, and soft tissue involvement independently predicted a lower EFS (Table). When the multivariate analysis was performed using PRIMA-PI factors, the presence of ≥2 EN sites was an adverse independent prognostic factor for OS (HR 2.28; 95% CI 1.01-5.18; p=0.05). Conclusions: Baseline PET/CT identifies EN and spleen sites of disease that can predict early clinical failure of FL. These results may be useful to better identify high-risk patients and guide appropriate therapy. [Table: see text]

7509 Background: Waldenstrom macroglobulinemia (WM) is a rare lymphoma for which scant comparative data exist to guide frontline therapy. Herein, we compare 3 commonly used regimens in WM: R-Benda, DRC, and BDR in frontline setting. Methods: Patients (Pts) with active WM seen at Mayo Clinic between 2000 & 2018 who received R-Benda, DRC or BDR as primary therapy were included in this retrospective study. Response rates were assessed by Consensus Criteria. All time to event analyses were performed from the frontline therapy, using Kaplan-Meier method. Results: The study included 172 pts with active WM (R-Benda, n=67, DRC, n=75, BDR, n=30).The median follow-up for the entire cohort was 3.7 years (y) (95% CI 3.7-3.0). Baseline characteristics, including IPSS, and time to frontline therapy from WM diagnosis were similar across the 3 cohorts. Clinically relevant endpoints are shown in the Table. Hematologic and non-hematologic toxicities were similar across the 3 groups. Grade 3 neuropathy requiring treatment discontinuation was encountered in 13% pts treated with BDR. 56 pts received subsequent salvage therapy [(10% in R-Benda arm, 44% in DRC arm, & 53% in BDR arm]; 29% pts in the R-Benda arm and 30% pts in DRC arm received a PI-based regimen while 69% pts in the BDR arm received alkylator-rituximab based therapy. Conclusions: Outcomes (MRR, TTNT and EFS) with frontline R-Benda are superior in comparison to frontline DRC or BDR in patients with WM. Clinically relevant endpoints are not significantly different with DRC vs. BDR. The toxicity profile across the 3 groups was comparable. [Table: see text]

8538 Background: Follicular lymphoma (FL) is a heterogeneous disease with a variable prognosis. Transformation to an aggressive lymphoma is associated with a poor prognosis. Previous studies have identified tumor microenvironment cells such as CD68+ macrophages, PD-1+ and FOXP3+ cells as prognostic in FL. However, results have been conflicting and infrequently address the risk of transformation. In this study we analyzed various components of the microenvironment and their association with transformation. Methods: Patients with FL that later transformed were identified through Mayo Clinic lymphoma database. Tissue specimens at diagnosis were stained with CD68, CD11c, CD21, CXCL13, FOXP3, PD-1 and CD14 and characterized by pattern of location and semi-quantitative cell content. Time to transformation (TTT) and overall survival (OS) were assessed by Kaplan-Meir analysis. Significant variables were further analyzed by cox proportional hazards model. Results: 58 patients were included with a median TTT of 4.7 years (range, 0.4-20). Presence of CD14+ cells vs absence and their location (follicular vs non-follicular) were associated with shorter TTT (4.5 vs 6.2 yrs., p=0.037 and 3.8 vs 5.9 yrs. p= 0.027, respectively). The quantity of PD-1+ cell content was not associated with TTT or OS. However, localization of PD-1 to the follicle compared to diffuse was associated with a longer TTT (6.1 vs 3.6 yrs. p=0.033) and superior OS (9.7 vs 4.6 yrs. p=0.009). On multivariate analysis, the pattern of CD14 and PD-1 staining remained significantly associated with shorter TTT (Table). Conclusions: In follicular lymphoma, a diffuse pattern of PD-1+ cells is associated with inferior TTT and OS. CD14+ cells localized to the follicle is associated with a shorter TTT. After accounting for FLIPI score, both these factors remained significant. These results identify two independent predictors of the rate of transformation in follicular lymphoma and suggest it is location rather than quantity of CD14+ or PD-1+ cells that influence outcomes. [Table: see text]

7041 Background: Prognostic factors for diffuse large B-cell lymphoma (DLBCL) at day 100 post-autologous peripheral blood hematopoietic stem cell transplantation (APBHSCT) have not been evaluated. We previously reported that absolute monocyte (AMC)/lymphocyte count (ALC) prognostic score (AMLCPS) at diagnosis of DLBCL patients is a prognostic factor of survival, stratifying patients into three risk groups: low- (AMC<630/μL and ALC>1000/μL), intermediate- (AMC ≥ 630/μL or ALC ≤ 1000/μL) and high-risk (AMC ≥ 630/μL and ALC ≤ 1000/μL) (Leukemia 25: 1502-9, 2011). Therefore, we evaluated day 100 AMLCPS as a prognostic factor of survival by landmark analysis from day 100 post-APBHSCT in DLBCL patients. Methods: DLBCL patients that achieved complete response by day 100 post-APBHSCT qualified for the study. From 2000 to 2007, 134 consecutive DLBCL patients were evaluated. Overall survival (OS) and progression free survival (PFS) were calculated using Kaplan-Meier analysis. Results: The median follow up from day 100 for the cohort was 5.5 years (range: 0.17-12.17 years) and for living patients (N=93) was 6.9 years (range: 2.5-12.17 years). Day 100 AMLCPS was able to stratify patients into three risk groups low-, intermediate-, and high-risk for OS [low: median OS = not reached, 5-year survival rate = 94% (95%CI 83.3%-98.1%); intermediate: median OS = not reached, 5-year survival rate = 70% (95%CI 58.0%-79.8%); high: median OS = 2.2 years, 5-year survival rate = 13% (95%CI 3.4%-40.5%); p<0.0001] and PFS [low: median PFS = not reached, 5-year progression free rate = 87% (95%CI 74.5%-94.3%); intermediate: median PFS = 10.9 years, 5-year progression free rate = 67% (95%CI 55.4%-77.5%); high: median PFS = 1 year, 5-year progression free rate = 13% (95%CI 3.4%-40.5%); p<0.0001]. The day 100 AMLCPS was balanced in relation to the International Prognostic Index (IPI; p=0.22), infused CD34+ stem cells (p=0.92), and disease status pre-APBHSCT (complete remission versus partial response; p=0.11). Conclusions: The day 100 AMLCPS is a simple biomarker that can help identify DLBCL patients post-APBHSCT with poor clinical outcomes.

TPS3113 Background: Programmed death-1 (PD-1) is an immune checkpoint receptor that inhibits T-cell activation upon interaction with its ligands PD-L1 or PD-L2. Increased PD-L1 expression has been reported with various hematologic malignancies and may prevent the host immune response from exerting an antitumor effect on the malignant cells. Nivolumab, a fully human IgG4 monoclonal PD-1 receptor blocking antibody, has demonstrated antitumor activity in pts with solid tumors including melanoma, renal cell carcinoma, and non-small cell lung carcinoma. We hypothesized that nivolumab could also mediate antitumor activity in pts with hematologic malignancies, a significant area of unmet medical need. We describe a phase I study to evaluate the effects of nivolumab in pts with select hematologic malignancies. Methods: This open-label, two-part study will enroll approximately 100 pts. During dose escalation, successive cohorts of pts with relapsed or refractory hematologic malignancies will be treated using a 6+3 escalation design. Pts will receive 1 or 3 mg/kg nivolumab IV every 2 weeks (wks) (the first dose will be followed by a 3-wk evaluation period), for 2 years, with the potential for an additional year of therapy for pts who progress during the follow-up period. Subsequently, 5 tumor-specific cohorts of 16 pts will be enrolled at the maximum tolerated dose (MTD) for multiple myeloma, B-cell lymphoma, T-cell lymphoma, Hodgkin lymphoma/primary mediastinal B-cell lymphoma, and chronic myelogenous leukemia. Response will be assessed at wks 4, 8, 16, 24, and every 16 wks thereafter. The primary study objective is to establish dose limiting toxicities, the MTD, and the recommended phase II nivolumab dose. Secondary objectives are to characterize nivolumab pharmacokinetics, immunogenicity, preliminary antitumor activity, and the potential association between PD-L1 expression on tumor cells and clinical efficacy. Exploratory objectives include investigation of the immunoregulatory effects of nivolumab in peripheral blood, bone marrow, and/or tumor. Pts are currently being enrolled at 3 mg/kg. Clinical trial information: NCT01592370.

8053 Background: Lenalidomide (LEN) is an immunomodulatory agent that has shown significant single-agent anti-lymphoma activity in patients with relapsed disease and in combination may enhance the efficacy of rituximab by various mechanisms. There is limited data on the role of LEN in combination regimens for treatment-naïve patients. This phase II single arm trial was designed to evaluate tumor response, toxicity (AE) and survival with a combination of LEN, rituximab, cyclophosphamide (CTX) and dexamethasone (DEX) (LR-CD) in symptomatic untreated patients with low grade B-cell non-Hodgkin lymphoma. Methods: Eligibility required age ≥18, ECOG PS ≤2, confirmed diagnosis of low-grade B-cell lymphoma (FL1-2, SLL, MZL or LPL/WM), measurable nodes ≥2cm or IgM ≥400mg/dL(LPL/WM), ANC ≥1400/mm3, platelets ≥100,000/mm3, creatinine ≤2mg/dL, signed informed consent and in need of therapy. Treatment consisted of IV rituximab 375mg/m2 D1, oral LEN 20mg D1-21, CTX 250mg/m2 D1, 8, 15, DEX 40mg D1, 8, 15, 22 and ASA 325 mg daily in a 28 day cycle. Treatment continued 2 cycles beyond best response (max 12). Toxicity was assessed by NCI CTCAE v3.0. Results: 28 patients have enrolled at Mayo Clinic with 25 evaluable for toxicity and 21 for response: Median age 65(43-83), 72% male, FL 24%, MZL 28%, LPL/WM 38%, and SLL 4%. Median number of cycles given was 5. Overall response in 21 patients with response data is 95% (95% CI 76-100%) with 19% CR (95% CI 5-42%), and 76% PR (confirmed and unconfirmed, 95% CI 53-92%). The ORR in 8 patients with LPL/WM with response data was 88%, all PR (95% CI 47-100%). At a median f/u of 14.7 months 96% are alive without progression. 1 death (unrelated) occurred 7.7 months after completing 12 cycles of treatment. The most common grade ≥3 AEs were neutropenia (37.5%), leukopenia (16.7%), anemia (12.5%) and fatigue (12.5%). Conclusions: The combination LR-CD with the novel agent LEN is feasible, well tolerated and produces high response rates in symptomatic patients with low grade B-cell NHL. Toxicities are manageable and similar to reported data of single agent LEN.