Evaluation of the potential for QTc prolongation with avelumab

Left Ventricular Ejection Fraction in Patients With Ovarian Cancer Treated With Avelumab, Pegylated Liposomal Doxorubicin, or Both

In the phase III JAVELIN Ovarian 200 trial, 566 patients with platinum-resistant/refractory ovarian cancer were randomized 1:1:1 to receive avelumab alone, avelumab plus pegylated liposomal doxorubicin (PLD), or PLD alone. Cardiac monitoring was included for all patients. We report left ventricular ejection fraction (LVEF) data from the trial. Grade ≥3 cardiac adverse events (AEs) occurred in 4 (2.1%), 1 (0.5%), and 0 patients in the avelumab, combination, and PLD arms, respectively. LVEF decreases of ≥10% to below institutional lower limit of normal at any time during treatment were observed in 1 (0.8%), 3 (1.9%), and 2 (1.5%) patients, respectively; 4 had subsequent assessments, and these showed transient decreases. No patient had a cardiovascular AE related to LVEF decrease. This analysis is, to our knowledge, the first analysis of LVEF in patients receiving immune checkpoint inhibitors.

CLINICALTRIALS.GOV IDENTIFIER

NCT02580058.

Concentration-QTc analysis for single arm studies

Concentration-QTc (C-QTc) modeling is being increasingly used in phase 1 studies. For studies without a placebo arm (single arm studies), the scientific whitepaper by Garnett et al. ( https://doi.org/10.1007/s10928-017-9558-5 ) states that time-matched baseline adjustments may minimize the effect of diurnal variation in QTc intervals, and categorical time effects are not needed in the model. However, how diurnal variations can be accounted for when only pre-dose baselines are available is unclear. This research investigates whether including categorical time effects in the model can adjust diurnal variation in single arm studies with pre-dose baselines, where QTc prolongation is evaluated at a concentration of interest based on ΔQTc at 24 h and ΔΔQTc (a model-derived difference in ΔQTc from concentration zero). To understand the operating characteristics for the models with and without categorical time effects, simulations were conducted under various scenarios considering oncology early phase studies. When the C-QTc relationship is linear, models without categorical time effects provided biased estimates for model parameters and inflated or decreased false negative rates (FNRs) depending on the pattern of diurnal variations in QTc intervals, whereas models with categorical time effects caused no biases and controlled the FNRs. For non-linear C-QTc relationships, ΔΔQTc estimations made using the model with categorical time effects were not robust. Thus, for single arm studies where only pre-dose baselines are available, we recommend collecting QTc measurements at 24 h and estimating ΔQTc at a concentration of interest at 24 h using the C-QTc model with categorical time effects.