Pharmacokinetic-tumor growth inhibition modeling of raludotatug deruxtecan (R-DXd) to support phase 3 dose selection in platinum-resistant ovarian cancer (PROC)

Background: R-DXd is a novel cadherin 6 (CDH6)-directed antibody-drug conjugate (ADC) under clinical investigation. In ongoing Phase 1 DS6000-A-U101 (NCT04707248) and Phase 2/3 REJOICE-Ovarian01 (NCT06161025) studies, preliminary data showed promising results in patients with platinum-resistant ovarian cancer (PROC). To support dose selection in PROC, a pharmacokinetic-tumor growth inhibition (PK-TGI) model was developed to investigate the effect of R-DXd exposure on tumor size and predict percent change from baseline in tumor size across R-DXd doses of interest (4.8, 5.6, and 6.4 mg/kg), as well as explore the effect of CDH6 expression levels on tumor response. Methods: The analysis dataset comprised 229 patients enrolled from either study who completed at least 18 weeks of follow-up (ie, 3 postbaseline tumor scans). Time-course data of PK (ADC and released payload [DXd] plasma concentration) and tumor size (sum of longest diameters [SLD], assessed by blinded independent central review [BICR]) were collected from PROC patients treated with R-DXd monotherapy at 1.6-9.6 mg/kg IV Q3W in DS6000-A-U101 and the dose optimization part of REJOICE-Ovarian01. The PK-TGI model was developed using nonlinear mixed-effects modeling and adapted from previously published TGI modeling frameworks by Claret et al. Simulations accounting for parameter uncertainties were conducted to predict tumor response of R-DXd treatment across dose levels for a typical reference patient. Results: Model parameters were estimated with good precision, and goodness-of-fit plots showed that model predictions describe data appropriately. The final model was comprised of three components: 1) a drug-induced tumor killing rate ( ) driven by R-DXd concentration, which is 2) attenuated over time by a resistance term ( ), and 3) tumor growth described as a saturated logistic function, , in which represents the intrinsic growth rate, tumor size, and the maximum tumor size. The simulations demonstrated a clear exposure-response relationship, with higher doses of R-DXd associated with stronger tumor shrinkage: predicted median tumor shrinkage at 24 weeks was −52%, −58%, and −64% for 4.8, 5.6, and 6.4 mg/kg, respectively. Additionally, baseline CDH6 protein expression was identified as a significant covariate on parameter, positively associated with tumor shrinkage, with its effect tending to plateau with increasing expression levels. Conclusions: The PK-TGI model could adequately characterize changes in tumor size as a function of R-DXd concentration as well as quantify the effect of target expression on tumor shrinkage. Combined with the totality of clinical data and an overall assessment of benefit-risk, the analyses supported the selection of 5.6 mg/kg Q3W as the optimal monotherapy dose of R-DXd for the Phase 3 part of the REJOICE-Ovarian01 study.