Exposure-response analysis and quantitative systems pharmacology modeling for an optimal dose selection of surovatamig in patients with DLBCL

Introduction: Surovatamig (formerly AZD0486) is a novel, IgG4 fully human CD19xCD3 bispecific T-cell engager that is being evaluated in an ongoing phase 1 study (NCT04594642) in patients (pts) with relapsed or refractory (R/R) B-cell non-Hodgkin lymphoma (B-NHL). Here, we present our quantitative approaches for determining an optimal dose for surovatamig in pts with R/R diffuse large B-cell lymphoma (DLBCL) with population pharmacokinetics (popPK), exposure-response (ER) analysis, and quantitative systems pharmacology (QSP) modeling. Methods: Pts with R/R B-NHL received escalating doses of surovatamig IV with fixed dosing, single step-up dosing (SUD), or double SUD schedules in cycle 1, followed by target dose (TD) every 2 weeks in 28-day cycles for up to 24 months. PK data from serial sampling was used to develop a popPK model in NONMEM ® . ER relationships between surovatamig PK and overall response rate (ORR) or complete response rate (CRR) were characterized in pts with DLBCL. Further, QSP modeling was developed to link the surovatamig PK to the trimeric complex formation. A total of 185 pts with all histology (DLBCL, follicular lymphoma [FL], and mantle cell lymphoma/marginal zone lymphoma) were analyzed for PK and safety, and 100 pts with DLBCL were analyzed for efficacy (based on data cutoff of August 2025). Results: Following IV infusion, the observed mean half-life of surovatamig is ~11 days across the dose cohorts after the cycle 1 day 15 dose. A dose-proportional increase in exposure is observed across the dose ranges (≥0.27 mg). Surovatamig PK was best described by a 2-compartment PK model with linear clearance and was comparable in pts with FL and DLBCL. No covariates (eg, disease type, race, age) other than body weight were identified to have a significant impact on PK parameters (clearance and volume of distribution). Exposure-efficacy analysis in pts with R/R DLBCL showed higher probability of ORR and CRR with increasing exposures (eg, C avgcycle1&2 or C trough ) across the dose range, supporting a TD of 25 mg in pts with DLBCL. Based on the QSP model calibrated with preclinical and emerging clinical data (eg, sum of the long diameter of target lesions), the maximal median trimer concentration at the tumor site increases across the TD of 2.4-37.5 mg. Predictions suggested no significant increments in ORR by increasing the TD from 25 mg to 37.5 mg. Overall, QSP modeling and simulation results support the TD of 25 mg as an optimal recommended phase 2 dose (RP2D) for efficacy. Exposure-safety analysis showed no clear association of exposure after TD with Gr2+ CRS or ICANS, or Gr4+ neutropenia. Conclusion: ER analysis and QSP modelling confirmed the clinical benefit with increased TD from 2.4 to 25 mg in pts with R/R DLBCL, with no clear TD-dependent clinically relevant toxicity, supporting 25 mg as the RP2D.