Planned Interim Analysis of a Phase 2 Investigator-Initiated Trial of Anakinra to Prevent CRS and Neurotoxicity after Treatment with Lisocabtagene Maraleucel

Background CD19 CAR T-cell therapies remain limited by significant toxicities: cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). Data from our group and others suggested the recombinant IL-1 receptor antagonist anakinra may successfully prevent and treat CRS and ICANS. We report here our planned interim analysis of 15 patients (pts) enrolled in a phase 2 study of anakinra to prevent CRS and ICANS after lisocabtagene maraleucel (liso-cel; NCT04359784). Methods Adults with B-cell LBCL receiving liso-cel per the FDA label were eligible. Anakinra was administered at 200 mg/day SC from day 0 through +13 after liso-cel infusion. CRS and ICANS severity were assessed per ASTCT criteria. The primary endpoint was the absence of any grade CRS assessed using the Bayesian optimal phase 2 design. Results Nineteen pts were screened and 15 were enrolled (Table 1). All 15 pts received liso-cel and all doses of anakinra as planned without dose adjustment, interruption, injection-site reactions, or AEs attributed to anakinra. All pts were evaluable for toxicity assessment; 12 pts were evaluable for disease response assessment at day +28.Eight pts (53%) developed any grade CRS, including 1 case of grade 3 CRS. Four pts (27%) developed any grade ICANS, including 3 with grade 3 (Figure 1). We observed robust in vivo CAR expansion (median peak: 5.1 log10 transgene copies/μg DNA). We could confirm CAR transgene persistence in blood in all pts with available data past day +60 (n = 5; range 60-159 days; Figure 2).Study pts were retrospectively compared to pts who received liso-cel per standard of care off trial without prophylactic anakinra at our center (n = 49). The incidence of CRS was numerically lower in anakinra-treated compared to anakinra-untreated pts (53% vs. 65%, respectively) with comparable median duration (2.5 vs. 3 days, respectively). The incidence and duration of ICANS were similar (27% vs. 29%; median 9 vs. 7 days, respectively). Based on Locally Estimated Scatterplot Smoothing estimates, time from liso-cel infusion to CRP normalization was faster in pts receiving anakinra (9 days) compared to those receiving liso-cel alone (14 days). The best ORR by Lugano criteria was 67% (CR, n = 5 [42%]) in pts receiving prophylactic anakinra vs. 81% (CR, n = 21 [57%]) in pts receiving liso-cel alone. Conclusion Low-dose SC anakinra prophylaxis during liso-cel treatment was very well tolerated with preserved anti-tumor efficacy and in vivo CAR T-cell kinetics. While our primary endpoint was not reached at our planned interim analysis, we observed lower CRS rates and faster resolution of CRS in comparison to a real-world cohort of pts receiving liso-cel without prophylactic anakinra. To further enhance the preventative effects of anakinra, we plan to change the administration route from SC to IV, and to allow anakinra dose uptitration for the last 10 participants in the study.