053: Concentration-QTc Analysis of Quemliclustat in Combination With/Without Zimberelimab & nab-Paclitaxel/Gemcitabine in Patients With Advanced Pancreatic Cancer

Statement of Purpose, Innovation or Hypothesis: Quemliclustat is a potent, selective, small-molecule inhibitor of soluble and membrane-bound CD73 developed with the aim of eliminating adenosine-mediated immunosuppression in the tumor microenvironment.  We assessed the effects of quemliclustat exposure on the corrected QT (QTc) interval by categorical analysis and modelling of the concentration-QTc relationship using data collected from a Phase 1 study.  

Description of Methods & Materials: Study AB680CSP0002 (ARC-8) was a Phase 1, open-label, dose-escalation, and dose-expansion study, designed to evaluate the safety, tolerability, PK, pharmacodynamics, and clinical activity of quemliclustat  in combination with zimberelimab (programmed cell death-1 inhibitor) and standard chemotherapy (nab-paclitaxel [NP] and gemcitabine [Gem]) in participants with advanced pancreatic cancer as first-line (1L) and second-line (2L) treatment for metastatic disease. Dose escalation/expansion phases evaluated quemliclustat doses ranging from 25-200 mg, Q2W administered as 1-hour IV infusion. Time matched pharmacokinetic samples and triplicate ECG measurements were obtained in all patients after the first dose and at steady-state.  A linear mixed-effects model was used to conduct the C-QTc evaluation using NONMEM software (version 7.5.0).¹ Pre-specified fixed effects evaluated in the linear mixed model were plasma concentration of quemliclustat, sex, time of day, and baseline QTcF. The final model-predicted point estimate and its 2-sided 90% CI for ΔQTcF at the geometric mean peak steady state values at clinically relevant doses 100 mg and 200 mg Q2W were calculated. No QTc prolongation was concluded if the upper bound of the 2-sided 90% CI (equivalent to the upper bound of the 1-sided 95% CI) of the model-predicted QTc effect (ΔQTcF) was below 10 msecs.

Data & Results: Time-matched PK and ECG measurements were available from 165 patients, with one patient excluded due to no recorded baseline QTc interval. The mean change-from-baseline HR (ΔHR) on quemliclustat showed no correlation with quemliclustat concentration suggesting there is no dose-related effect on HR. The time courses of mean quemliclustat plasma concentrations and arithmetic mean ΔQTcF across post-dose time-points for Cycle 1, did not show hysteresis. Observed QTcF and ΔQTcF values in other dose groups above various thresholds did not appear to be quemliclustat dose-related. The estimated population slope of the concentration-QTc relationship was 0.118 msec per μg/mL (27% RSE). The slope of quemliclustat concentration was statistically significant at significance level of 0.01 (two-sided), producing a change in NONMEM objective function value of -10.43.  The predicted ΔQTcF at the geometric mean peak steady state quemliclustat concentrations for the 100 mg and 200 mg Q2W dosing regimens were 2.27 msecs (90%CI: 1.26, 3.27) and 4.54 msecs (90%CI: 2.52, 6.57), respectively.

Interpretation, Conclusion or Significance: Based on linear mixed effects modeling, the upper bound of the 90% CI of the predicted effect is < 10 msecs and effect on ΔQTcF exceeding 10 msecs can be excluded within the observed ranges of plasma concentrations, up to 57 μg/mL. Quemliclustat (doses up to 200 mg Q2W) was not associated with QTcF prolongation.

Disclosures: Each author is employed by and owns shares in Arcus Bioscience Inc.

Citations/References: (1) Garnett C, Bonate PL, Dang Q, et al. Scientific white paper on concentration-QTc modeling. J Pharmacokinet Pharmacodyn. 2018;45(3):383-397.

Additional Information/Authors: NA