Exposure-response (E-R)-based product-profile (PP)-driven clinical utility index (CUI) to support phase III dose selection in oncology
Rui Zhu (1), Bill Poland (2), Russ Wada (2), Qi Liu (1), Luna Musib (1), Daniel Maslyar (1), Wei Yu (1), Han Ma (1), Angelica Quartino (1), Jin Yan Jin (1), Nageshwar Budha (1)
(1) Genentech, Inc., South San Francisco, CA, USA; (2) Certara, CA, USA.
Introduction/Objectives:
Benefit-risk assessment of a therapeutic agent can support decision-making in both drug development and regulatory evaluation. The traditional way of assessing benefit-risk involves characterizing dose/exposure-response relationships on primary efficacy and safety endpoints separately and then qualitatively comparing those relationships to support decision-making of dose/formulation selection, go/no go in clinical development, etc. However, with this implicit approach, it is difficult to balance benefit with risk when there are multiple important attributes. Therefore, the CUI, a more structured quantitative approach that brings all the attributes onto the same scale and reduces them to a single measure, allows more transparent and efficient benefit-risk assessment and decision-making [1-4].
The aims of this work were to characterize ipatasertib E-R relationships in a phase II study and to quantitatively assess benefit-risk using the CUI approach to support ipatasertib phase III dose selection in combination with abiraterone in patients with metastatic castration-resistant prostate cancer (mCRPC).
Methods:
Data used in the analyses were from a double-blinded, randomized phase II part of a phase Ib/II clinical trial, where patients (n=253) received placebo, or ipatasertib 200 mg or 400 mg daily in combination with abiraterone [5]. Logistic regression and Cox proportional-hazards models characterized E-R relationships for safety (various adverse events [AEs]) and efficacy (radiographic progression-free survival [rPFS]) endpoints, respectively. To capture the effect of dose modifications, an exposure metric based on the actual dosing (AUCactual) was used in the E-R modeling. Dose-intensity (DI) models were developed to characterize the DI relationship across treatment arms. E-R models were coupled with their corresponding DI models to project the dose-response relationships over the range of 0 to 500 mg daily ipatasertib dose. In addition, E-R analyses and dose-response projections with (AUCactual) and without (AUC based on planned dose [AUCplanned]) considering dose modifications were compared. A utility measure was developed for each important safety or efficacy attribute selected based on ipatasertib PPs, and they were weighted and combined for the CUI calculation. Combined results from overall utility profiles and probability of reaching PP utility levels were used to determine the dose with optimal benefit-risk balance.
Results:
The E-R analysis of the rPFS hazard ratio (HR) demonstrated a statistically non-significant (P>0.05) association between ipatasertib exposure and rPFS HR, with a slight trend of higher exposure leading to a lower rPFS HR. The E-R analyses of AEs indicated a statistically significant (P<0.05) association between exposure and AEs tested. The probability of having AEs of various grades increased with increasing exposure. Comparison between the E-R relationships from models with AUCactual and AUCplanned indicated that the E-R trends were generally flatter in the latter, but the dose-response projections from both models were very similar with slightly larger variability in the latter.
Based on the PPs, rPFS HR, diarrhea, and rash were selected as key attributes with cutoff/tradeoff values. Given the AEs are generally manageable and reversible, a slightly higher weight was chosen for efficacy (0.6) than for AEs (diarrhea [0.3] plus rash [0.1]). Sensitivity analyses were conducted to test the following 4 pre-defined scenarios: (1) weight assignment for rPFS:diarrhea:rash=0.6:0.3:0.1, grade ≥2 AEs, (2) 0.6:0.3:0.1, grade ≥3 AEs, (3) excluding rash with 0.6:0.4:0, grade ≥2 diarrhea, and (4) excluding rash with 0.6:0.4:0, grade ≥3 diarrhea. Results from all 4 scenarios supported 400 mg daily as the optimal dose.
Conclusions:
This E-R-based PP-driven CUI framework may be useful to support dose selection in clinical drug development with multiple attributes to balance. In practice, pre-defined PP can be a good anchor point to help team reach agreement on the key components for CUI analysis, eg, important attributes, weights, and clinically-meaningful cutoff/tradeoff values. Comparison of E-R modeling and simulation results from exposure metrics with and without considering dose modifications can be used to guide exposure metric selection in E-R analyses in trials with sizable dose modifications.
References:
[1] Khan, A.A., et al. AAPS J 11, 33-8 (2009).
[2] Poland, B. et al. Clin Pharmacol Ther 86, 105-8 (2009).
[3] Ouellet, D. Expert Opin Drug Saf 9, 289-300 (2010).
[4] de Greef-van der Sandt, I. et al. Clin Pharmacol Ther 99, 442-51 (2016).
[5] Bono, J.S.D. et al. Journal of Clinical Oncology 34, Abstract No. 5017 (2016).