P1210: Immune Correlates of Response to Glofitamab: Biomarker Findings From a Pivotal Phase Ii Expansion Study in Patients With Relapsed or Refractory (R/R) Diffuse Large B-Cell Lymphoma (Dlbcl)

Background: Glofitamab is a T-cell-engaging bispecific antibody with a novel 2:1 configuration that confers bivalency for CD20 (B cells) and monovalency for CD3 (T cells). In an ongoing Phase I/II study (NCT03075696), glofitamab demonstrated durable complete responses (CR) and a favorable safety profile in patients (pts) with R/R B-cell lymphomas (Phase I dose-escalation part; Hutchings, et al. 2021) and in pts with R/R DLBCL (pivotal Phase II expansion). Aims: To evaluate tumor and peripheral blood (PB) biomarkers associated with clinical response to glofitamab in the pivotal Phase II expansion. Methods: Pts with DLBCL (DLBCL, NOS, HGBCL, PMBCL, or trFL; n=107) and ≥2 prior therapies received intravenous (IV) obinutuzumab pretreatment (Gpt; 1000mg) 7 days before the first glofitamab dose. IV glofitamab was then given as step-up doses on Day (D) 1 (2.5mg) and D8 (10mg) of Cycle (C) 1 and at the target dose (30mg) on D1 of C2–12 (21-day cycles). Response was assessed by Independent Review Committee (Lugano 2014 criteria). All pts provided informed consent. Baseline (BL) tumor biopsies were assessed by CD20/PAX5 immunohistochemistry (n=64) and CD8/Ki67 immunofluorescence (IF) (n=61). PB biomarkers were evaluated by flow cytometry (n=87) and plasma cytokines by ELISA (n=70). Immune and stromal cell type were inferred from bulk tumor RNAseq samples (n=55) using xCell cell type enrichment analysis. All BL samples were collected prior to Gpt. Results: Analysis of BL tumor biopsies showed all pts were CD20 positive (range: ~35–98% CD20+ PAX5+ B cells) and CD20 levels were not associated with response. A trend towards higher CD8 counts was observed in pts with CR (by IF in total CD8 and CD8+Ki67- T cells). Gene expression analysis showed that responders had higher tumor microenvironment abundance enrichment scores, including higher CD4+ and CD8+ T-cell subsets scores (Figure). PB immune profiling revealed a positive association of BL counts of B cells (p<0.01), CD4 T cells (p=0.02), and CD4 effector memory (EM) cells (p=0.03) with response. There were no significant differences in BL counts of CD8 subsets, regulatory T cells, natural killer cells or monocytes. In pts with progressive disease (PD), there was a trend towards higher expression of the checkpoint marker PD1 on CD8 cells (p=0.05). Pts with CR had lower levels of IL-6 and C-reactive protein (CRP) at BL (test for association with CR: p=0.1 and p=0.01, respectively). Early on-treatment changes in PB were associated with response to glofitamab: induction of IL-6, IFNγ, IL-10, and TNFα in the 6 hours after the first glofitamab dose was strongest in pts with CR (IL-6, IFNγ, IL-10: 6- to 16-fold induction; TNFα: 1.5-fold induction). Transient margination of CD4 and CD8 T cells was strongest in pts with CR. Additional analyses, including ctDNA, are ongoing and updated data will be presented. Image:Summary/Conclusion: Biomarker results from the expansion cohort in pts with R/R DLBCL were consistent with those from the dose-escalation (Broeske et al. 2021), including comparable BL tumor CD20 expression across response categories, a trend towards higher BL tumor CD8 levels in pts with CR, and lower BL IL-6 and CRP in pts with CR. T-cell margination was strongest in pts with CR. Several novel response-associated biomarkers were identified in PB including higher BL B cell, CD4, and CD4 EM counts in responders and higher PD1 expression on CD8 in pts with PD. These data suggest a model where the clinical activity of glofitamab may be dependent upon immune contexture in the tumor and PB at the start of treatment.