Exposure–response analysis of safety and efficacy of capivasertib administered as monotherapy or in combination with fulvestrant
Carlos Fernandez (1), Marie Cullberg (2), Ignacio González-Garcia (1*), Diansong Zhou (3)
(1) Clinical Pharmacology and Quantitative Pharmacology, BioPharmaceuticals R&D, AstraZeneca R&D, Cambridge, UK, (2) Clinical Pharmacology and Quantitative Pharmacology, BioPharmaceuticals R&D, AstraZeneca R&D, Gothenburg, Sweden, (3) Clinical Pharmacology and Quantitative Pharmacology, BioPharmaceuticals R&D, AstraZeneca R&D, Boston, MA, USA. *Affiliation at time of study.
Objectives:
The PI3K/AKT/PTEN signalling pathway is dysregulated in various cancers. Capivasertib is a potent, selective inhibitor of all three AKT isoforms [1]. Several doses and schedules of capivasertib have been studied clinically (all twice daily; BD) [2–5]. Monotherapy has been given as continuous and intermittent schedules (4 days on/3 days off or 2 days on/5 days off). Across the range of doses and schedules studied, the most common clinically relevant adverse events (AE) were diarrhoea, rash and hyperglycaemia. In the Phase III CAPItello-291 trial in patients with aromatase inhibitor-resistant, hormone receptor-positive/human epidermal growth factor receptor 2-negative advanced breast cancer, capivasertib+fulvestrant significantly improved the dual primary endpoints of progression-free survival (PFS) vs placebo+fulvestrant in the overall and AKT pathway-altered population [6]; intermittent dosing was used in CAPItello-291. The objective of this analysis was to evaluate the capivasertib exposure–response relationship (ER) for safety and efficacy in patients with advanced/metastatic solid tumours.
.
Methods:
The ER for capivasertib monotherapy safety was explored by pooling data from 277 patients from three Phase I trials [2–4]. Patients received capivasertib at doses from 80 to 800 mg BD, with either a continuous BD daily schedule or one of two intermittent BD schedules: 4 days on/3 days off or 2 days on/5 days off.
The ER for capivasertib+fulvestrant safety was explored in pooled data from 414 patients in a Phase I trial [2] and CAPItello-291 [6]. The ER for capivasertib+fulvestrant efficacy was explored in 340 patients from CAPItello-291. In the combination trials, patients in the capivasertib arm received capivasertib 400 mg BD (4 days on/3 days off)+fulvestrant.
Patients with available capivasertib exposure metrics (area under the concentration–time curve [AUC], maximum concentration [Cmax], dose, weekly AUC and weekly dose) derived from population pharmacokinetic (PK) models were included in the analyses. Safety endpoints were AE leading to dose discontinuation, AE leading to dose modification, serious AE, grade ≥3 AE, grade ≥1 AE, grade ≥2 diarrhoea, grade ≥2 rash, grade ≥3 hyperglycaemia and increased blood glucose >13.9 mmol/L. Safety data were analysed using logistic regression models. Efficacy endpoints were PFS and objective response rate (ORR); analyses were based on Kaplan–Meier curves and Cox proportional hazard models (PFS) and logistic regression models (ORR). The effects of several covariates on capivasertib PK were also explored.
Results:
Safety analyses for capivasertib monotherapy trials (all dose levels) showed statistically significant relationships between capivasertib weekly dose and/or weekly AUC and all safety endpoints evaluated (p<0.005) except grade ≥1 AE, which were observed in 99.3% of patients. Weekly exposures were lower with intermittent vs continuous schedules. The intermittent 4 days on/3 days off schedule showed improved safety vs continuous dosing (diarrhoea, rash) and 2 days on/5 days off (hyperglycaemia). Increasing total weekly exposure increased the prospect of AEs, except hyperglycaemia which was better related to exposure during a dosing interval.
In the safety analysis for capivasertib+fulvestrant, the incidence of most events was similar to that in the monotherapy trials. Only the relationship between AUC and the prospect of AE leading to dose modifications was statistically significant, suggesting patients with high initial exposure required dose modification to adjust their exposure to a more normalised level. ER relationships for all safety endpoints were flat.
In the efficacy analyses, no significant relationships were identified between capivasertib exposure and PFS (univariate Cox regression p=0.129 for AUC, 0.147 for Cmax) or ORR (p=0.219 for AUC, 0.359 for Cmax). The effects of covariates, including race, age, weight, sex, renal function, hepatic function and dosing schedule, on capivasertib PK were all <20% and not predicted to be clinically relevant.
Conclusions:
Safety analysis of capivasertib monotherapy trials using a wide range of dose levels showed that the 4 days on/3 days off intermittent schedule was generally better tolerated vs a continuous schedule. Over the range of capivasertib exposure in the Phase III trial, no clinically relevant relationships were identified for efficacy or safety. This suggests there is no need for a priori capivasertib dose adjustment.
References:
[1] Davies BR, et al. Mol Cancer Ther 2012;11:873–887.
[2] Banerji U, et al. Clin Cancer Res 2018;24:2050–2059.
[3] Tamura K, et al. Cancer Chemother Pharmacol 2016;77:787–795.
[4] Dean E, et al. Cancer Chemother Pharmacol 2018; 81:873–883.
[5] Jones RH, et al. Lancet Oncol 2020;21:345–357.
[6] Turner NC, et al. Presented at the 2022 San Antonio Breast Cancer Symposium; December 6–10, 2022; San Antonio, TX. Abstract GS3-04.
Acknowledgements:
This study was funded by AstraZeneca. AstraZeneca-funded medical writing support was provided by Suzanne Patel, PhD, from BOLDSCIENCE Inc. Ignacio González-Garcia is currently employed by Cognigen, Buffalo, NY, USA. Capivasertib was discovered by AstraZeneca subsequent to a collaboration with Astex Therapeutics (and its collaboration with the Institute of Cancer Research and Cancer Research Technology Limited).