Inflammation-Based Scores Predict Survival after CD3xCD20 Bispecific T Cell Engagers in R/R LBCL

Management of relapsed/refractory Large B Cell Lymphoma (r/r LCBL) after CD19 Chimeric Antigen Receptor (CAR) T cells represents a major clinical challenge. Glofitamab and Epcoritamab, two CD3xCD20 bispecific T cell-recruiting antibodies, have recently been approved for third-line treatment. However, real-world experience with these bispecifics is limited, and predictive biomarkers for efficacy are lacking. Given that both bispecifics and CAR-T cells mediate their antineoplastic activity by redirecting the patient’s T cells, we hypothesized that biomarkers associated with outcomes from CD19 CAR-T therapy would similarly predict outcomes post-CD20-directed bispecifics. To test this, we applied two established models from CAR-T cell therapy—CAR-HEMATOTOX (Rejeski et al, Blood 2021) and InflaMix (Raj et al, ASH Annual Meeting 2023) to our cohort of 63 patients treated with bispecifics.

In this international, multicenter observational study, we analyzed outcomes of r/r LBCL patients treated with Glofitamab (n=55) or Epcoritamab (n=8) monotherapy. Baseline features were assessed on the first day of the first treatment cycle. Parameters were derived from standard laboratory analysis. For a subset of patients, cytokines were measured using the Legendplex flow-cytometry based multiplex immunoassay. Kaplan-Meier estimates of progression-free (PFS) and overall survival (OS) were compared using the log-rank test. Univariate and multivariate analyses of PFS were conducted using the Cox proportional hazards model.

The CAR-HEMATOTOX model integrates five variables (ANC, hemoglobin, platelet count, C-reactive protein, and ferritin) determined at the time of lymphodepletion. The InflaMix model, based on up to 14 laboratory values assessed prior to CAR-T cell infusion, stratifies patients into inflammatory and non-inflammatory clusters. Both models discriminate between cohorts with significantly different clinical outcome.

Median age was 67 years (range 35-86). Patients had received a median of 3 prior therapy lines (range 1-8), 46 patients (73%) underwent prior CAR-T cell therapy and 21 patients (33%) were refractory to their latest line of therapy. The objective response rate was 54%, with 16 complete remissions (25%) and 18 partial remissions (29%). With a median follow up of 4.1 months, median PFS was 3.5 months and median OS was 13.5 months. Cytokine release syndrome (CRS) occurred in 38.6% of patients (22/57), high grade CRS (grade ≥3) was reported in 3 cases. ICANS was noted in 5.3% of patients (3/57), with one patient experiencing high grade ICANS.

In a univariate Cox regression analysis, the examined clinical factors (including sex, age, ECOG performance status, histology, Ann Arbor stage, presence of extranodal disease, number of prior lines of therapy and prior CAR-T exposure) were not significantly associated with PFS. Among the analyzed laboratory values, however, higher LDH (p<0.001), CRP (p=0.002), and ferritin levels (p=0.006) as well as low hemoglobin (p=0.021) were significantly associated with inferior PFS, highlighting inflammation as a key factor associated with response.

Based on data availability, CAR-HEMATOTOX and InflaMix were applied to 45 and 55 patients, respectively. Patients presenting with a CAR-HEMATOTOX score ≥2 (24/45) showed inferior PFS (1.9 months vs not reached, p=0.005) and OS (3.9 vs 13.7 months, p=0.027). When applying InflaMix, 38.2% of the patients were assigned to the inflammatory cluster, showing a significantly shorter median PFS (1.7 vs 6.9 months, p<0.001) and OS (2.7 months vs not reached, p<0.001). In a multivariable Cox regression analysis accounting for LDH, age, prior CAR-T exposure and bispecific product, higher LDH and assignment to the inflammatory cluster by InflaMix remained independent risk factors for inferior PFS. Notably, InflaMix had the most significant effect on PFS (p=0.005).

In summary, our multi-center international real-world data analysis demonstrates that the CAR-HEMATOTOX and the InflaMix models effectively identify LBCL patients with high risk of treatment failure, highlighting the importance of immune dysregulation in failure to bispecifics. Further validation of both models is ongoing. These scores may guide patient stratification into bispecifics monotherapy, combinatorial approaches and inclusion in clinical trials to further advance treatment approaches in LBCL.