Elodie Valade1, Mingyu Heo2, Saerom Kim2, Marta Valle1, Jungwon Woo2
1Parexel Clinical Pharmacology, Modeling and Simulation, 2MedPacto, Inc.
Objectives: Vactosertib is a highly selective, potent inhibitor of the protein serine/threonine kinase activity of transforming growth factor (TGF)-ß receptor type 1. Vactosertib is being developed as an anticancer agent for various solid and hematologic malignancies. Although vactosertib elicited concentration-dependent inhibition of human ether-à-go-go-related gene (hERG)-mediated potassium currents from 10 to 100 µM, no pathophysiological effects on respiratory rate, CNS function, or cardiovascular system were observed. To assess vactosertib QT interval prolongation potential, a concentration-QTc analysis was performed using retrospective data from available clinical trials in oncology patients. Methods: A total of 2 clinical studies in patients with advanced solid tumor or in patients with low and intermediate myelodysplastic syndrome were selected. Vactosertib was administered once daily (QD) or twice daily (BID) for 5 days on and 2 days off-treatment, repeated for 28-day cycles. Dose levels included 30, 60, 100, 140, 200, 260 and 340 mg QD and 100, 200 and 300 mg BID. No placebo data were available. In total, 529 matched vactosertib plasma concentrations and baseline-adjusted Fridericia corrected QT interval (?QTcF) from 36 subjects were available for analysis. An adapted pre-specified model for conc-QTc analysis was applied [1]. Model assumptions on heart rate (HR) effects, hysteresis and linearity were explored and sensitivity modelling analyses were performed when deemed appropriate. The developed models were evaluated with goodness-of-fit and visual predictive checks. Predictions of mean ?QTc and two-sided 90% confidence interval (CI) at concentrations of interest were calculated using nonparametric bootstraps. Data management was performed using SAS. Graphical analyses were performed using R version 3.3.3. Model development and simulations used NONMEM version VII level 4.4 (ICON, Hanover, MD), compiled using the GNU Fortran 4.6.3 for Windows, and Perlspeaks NONMEM (PsN) version 5.0.0. Results: Primary analysis: The adapted pre-specified linear mixed effects model for conc-QTc analysis including between-subject variability on intercept was used. The predicted ?QTcF (90% CI) were 5.46 ms (3.08-7.69 ms), 5.58 ms (3.17-7.84 ms) and 6.95 ms (4.15-9.72 ms) for the corresponding vactosertib geometric mean Cmax at steady state for 200 mg QD, 200 mg BID and 300 mg BID dosing regimens, respectively. The upper bounds of the two-sided 90% CI for the model-derived ?QTcF at the Cmax of interest were not predicted to reach the 10 ms threshold. Sensitivity analyses: A sensitivity analysis using predicted concentration in an effect compartment, added to a previously developed population pharmacokinetic model for vactosertib, was performed. The predicted ?QTcF (90% CI) were 7.31 ms (5.16-9.78 ms), 7.69 ms (5.42-10.28 ms) and 11.52 ms (8.13-15.41 ms) for the corresponding vactosertib geometric mean Cmax at steady state for 200 mg QD, 200 mg BID and 300 mg BID dosing regimens, respectively. A subject-specific HR correction was estimated using on- and off- treatment data to derive individual-specific corrected QT interval (QTcI) [2]. The predicted ?QTcI (90% CI) were 3.24 ms (0.978-5.35 ms), 3.39 ms (1.02-5.61 ms) and 5.09 ms (1.54-8.41 ms) for the corresponding vactosertib geometric mean Cmax at steady state for 200 mg QD, 200 mg BID and 300 mg BID dosing regimens, respectively. Conclusions: Considering vactosertib geometric mean Cmax at steady state, the upper bound of the two-sided 90% CI for the model-derived ?QTcF (primary analysis) or ?QTcI (sensitivity analysis) was not predicted to reach 10 ms. When accounting for a delay, the upper bound of the two-sided 90% CI for the model-derived ?QTcF was predicted to reach the 10 ms threshold for the 200 mg BID and 300 mg BID dosing regimens. Regardless of the type of analysis, the upper bound of the two-sided 90% CI for the model derived ?QTcF or ?QTcI was not predicted to reach the 20 ms threshold for vactosertib geometric mean Cmax at Cycle 1 Day 1 or at steady state for the dosing regimens clinically relevant (200 mg QD, 200 mg BID and 300 mg BID).
[1] Garnett C, Bonate P, Dang Q, Ferber G, Huang D, Liu J, Mehrotra D, Riley S, Sager P, Tornoe C, Wang Y. Scientific white paper on concentration-qtc modeling. Journal of Pharmacokinetics and Pharmacodynamics, 45:383–397, 2018. [2] Malik M, Hnatkova K, Batchvarov V. Differences between study-specific and subject specific heart rate corrections of the QT interval in investigations of drug induced QTc prolongation. Pacing Clin Electrophysiol., 27(6 Pt 1):791–800, 2004.
Reference: PAGE 33 (2025) Abstr 11716 [www.page-meeting.org/?abstract=11716]
Poster: Drug/Disease Modelling - Other Topics