Elacestrant Population Pharmacokinetic and Exposure-Response Modelling Analyses to Support Elacestrant Dose Recommendations in Metastatic Breast Cancer (mBC) patients
E. Borella (1), C. Piana (1), P. Mazzei (1), S. Bouchene (1)*, M.Pitsiu (2), Y.L. Ho (3), C. Farrell (2), K. Grzegorzewski (1), F. Perabo (3), N. Habboubi (1)
(1) Stemline Therapeutics/Menarini group, (2) ICON Early Development Services, (3) Radius Health **No longer working for this company
Objectives: Elacestrant is a selective estrogen receptor degrader (SERD) developed for the treatment of estrogen receptor-positive breast cancer. Preliminary safety and efficacy exposure-response (E-R) analyses conducted using Phase 1b data supported the selection of the 400 mg** once daily (QD) dose for the Phase 3 study in metastatic breast cancer (mBC) patients (EMERALD, RAD1901-308, NCT03778931). The aims of this work were (1) to develop a population PK (popPK) model for elacestrant using pooled Phase 1 - 3 data, (2) to perform efficacy and safety E-R analyses using EMERALD Phase 3 data, and (3) to confirm the adequacy of elacestrant 400 mg QD dose in the target patient population.
Methods: A popPK model was developed to describe elacestrant disposition in healthy subjects (HS) and mBC patients using non-linear mixed effects modelling with NONMEM. The effects of demographics, lab parameters, concomitant medications, and formulation on elacestrant PK were investigated. The popPK model predictive performance was evaluated using goodness of fit plots, parameter estimates precision and visual predictive checks. Elacestrant model-predicted exposure was used in logistic regression models, implemented in R, to characterize the E-R relationships for efficacy (clinical benefit [CB] defined as achievement of complete response, partial response, or stable disease) and safety (adverse event [AE] of interest, namely nausea). Baseline ECOG status, visceral metastases, Estrogen Receptor 1 mutational status, prior treatments and lines of therapy were also assessed as predictors of response.
Results: 4171 quantifiable elacestrant concentrations from a total of 447 HS and mBC patients were used in the popPK analysis. Elacestrant PK was adequately described by a 2-compartment model with linear elimination and first-order absorption with lag time. The estimated PK parameters (relative standard error %) were apparent central clearance (CL/F) of 186 L/h (2.17%), apparent central volume of distribution (Vc/F) of 422 L (17.9%), apparent peripheral clearance (Q/F) of 231 L/h (17.2%), apparent peripheral volume of distribution (Vp/F) of 5411 L (5.37%), absorption rate constant (ka) of 0.0997 h-1 (14.4%), and lag time (Tlag) of 0.812 h (3.36%). Inter-individual variability was estimated on all parameters. Dose effect on relative bioavailability was structurally included in the model as well as body weight on all clearances and volumes. Age had a significant effect on central clearance and sex on peripheral clearance. None of those covariate effects warranted any dose adjustment. A total of 224 patients with CB data and 232 patients with AE data were included in the E-R analyses of EMERALD data. The mean (standard deviation) model-predicted AUC0-24h,ss (steady state daily AUC), AUCav (average daily AUC), and Conc4h (concentration 4h post-dose) in EMERALD patients were 2.44 (1.08) μg.h/mL, 2.32 (1.05) μg.h/mL, and 119 (51.9) ng/mL. Graphical comparison of model-predicted AUC0-24h,ss of EMERALD patients with the logistic curve of the probability of CB vs AUC0-24h,ss from Phase 1b studies shows that no differences in probability of CB are expected throughout the EMERALD exposure range, suggesting that maximum effectiveness has been achieved in the majority of patients. The comparison of model-predicted AUC0-24h,ss of EMERALD patients with the logistic curve of the probability of AEs vs AUC0-24h,ss from Phase 1b studies shows that the probability of experiencing severe AEs remains below 50% in the entire EMERALD exposure range, while it markedly increases for doses higher than 400 mg QD. As expected, following the administration of elacestrant 400 mg QD, increasing elacestrant exposure was not associated with a higher probability of CB (for AUCav, Odds Ratio [95% CI]=0.993 [0.718-1.374], p=0.97) or nausea (for Conc4h, Odds Ratio [95% CI]=0.999 [0.994-1.004], p=0.74).
Conclusions: The developed PopPK model suitably described the elacestrant PK in HS and mBC patients. The E-R analyses for efficacy and safety confirmed that all patients in EMERALD study were adequately exposed to elacestrant, thus confirming the suitability of the 400 mg QD dose in the target patient population.
**400 mg elacestrant dihydrochloride equivalent to 345 mg elacestrant