Evaluation of Baseline CAR-Hematotox Scores to Predict Increased Severe Infection Risk in Patients with Relapsed/Refractory Follicular Lymphoma Treated with Odronextamab

Introduction

Severe infections and cytokine release syndrome (CRS) are adverse events commonly observed following treatment with T-cell–engaging therapies such as CAR-T therapies and CD3-targeting bispecific antibodies (bsAbs). The CAR-HEMATOTOX model, which includes markers of hematopoietic reserve (absolute neutrophil count [ANC], platelet count, hemoglobin) and inflammation (C-reactive protein [CRP], ferritin), measured at baseline prior to lymphodepletion, has been shown to help identify patients at heightened risk of severe infections and poor outcomes following CAR-T therapy (Rejeski K et al. J Immunother Cancer 2022). The model was trained and validated using real-world data from patients with large B-cell lymphoma receiving commercial CAR-T therapies (Rejeski K et al. Blood 2021).

Odronextamab, a CD20×CD3 bsAb, has been investigated in Phase 1 and 2 studies in relapsed/refractory (R/R) follicular lymphoma (FL) and diffuse large B-cell lymphoma and exhibits a similar safety profile in both settings. Here, we evaluate whether baseline CAR‑HEMATOTOX scores can help identify patients with R/R FL at risk of severe infections or CRS events when receiving odronextamab.

Methods

This retrospective analysis of the ELM-2 study evaluated infections in 153 patients with R/R FL. All patients received steroid premedication, infection prophylaxis, and odronextamab monotherapy as previously described (Taszner M. EHA 2024). Baseline CAR-HEMATOTOX scores were calculated for each patient, with 1 point assigned for ANC ≤1200/µL, hemoglobin ≤9 g/L, platelet count 75–175 × 10⁹/L, CRP ≥3 mg/dL, and ferritin 650–2000 ng/mL, and 2 points assigned for platelet count <75 × 10⁹/L and ferritin >2000 ng/mL. Patients were stratified into 2 groups based on the sum of scores: low risk (HT^low, 0–1) and high risk (HT^high, ≥2). Infection rates were analyzed by risk score for different periods: events occurring during the first 30, 60, and 90 days of treatment, and until end of study. CRS rates were analyzed by risk score for events occurring within the first 30 days.

Results

Of the 153 patients evaluated, complete baseline scores were available for 91% (n=139). A majority of patients were HT^low (82.7%) at baseline. The median age was 61 years (range 22–84), median number of prior lines of therapy was 3 (range 2–13), and 26% of patients had received a prior autologous stem cell transplant. The median values for hematologic baseline measures were generally similar to the median values reported in the original CAR‑HEMATOTOX validation set. However, the proportion of patients with FL who were HT^high in ELM-2 at baseline was substantially lower than the proportion reported in the CAR-HEMATOTOX population (17.3% vs 48.9%). The median duration of odronextamab exposure was 38.14 weeks.

In the first 30 days, Grade ≥3 infections were observed slightly more frequently in the HT^high compared with the HT^low group (8.3% vs 6.1%, nominal P-value 0.654). This difference was more pronounced at 60 days (20.8% vs 10.4%, nominal P-value 0.174) and 90 days (25.0% vs 14.8%, nominal P-value 0.233). When excluding COVID-19 infections, there was approximately double the incidence of Grade ≥3 infections at 30 days (8.3% vs 4.3%, nominal P-value 0.347), 60 days (16.7% vs 8.7%, nominal P-value 0.264), and 90 days (20.8% vs 12.2%, nominal P-value 0.324) in the HT^high vs HT^low group.

The rate of Grade ≥2 CRS events was higher in the HT^high compared with the HT^low group (25.0% vs 16.5%, nominal P-value 0.381), and this was consistent across different step-up regimens.

Conclusions

Simple, accessible tools are needed to help identify patients who may have greater susceptibility to infections and other cytokine-mediated toxicities when receiving odronextamab. This is the first study evaluating CAR-HEMATOTOX scores for identifying patients at risk of severe infections following treatment with T-cell–engaging bsAbs in FL. Preliminary results of this exploratory, post hoc analysis indicate that HT^high may be useful for identifying patients at high risk of developing infections during the first 90 days of treatment with odronextamab. Additional analyses, including clinical phenotypes of neutrophil recovery, cumulative infection rates, IgG levels, impact of steroid utilization, infection prophylaxis, re-baselining CAR-HEMATOTOX at 90 days, and association with long-term infection risk, will be presented.