Exposure-Response Analysis of Datopotamab Deruxtecan (Dato-DXd) in Combination with Pembrolizumab with or without Platinum Chemotherapy in Patients with Advanced or Metastatic Non-small Cell Lung Cancer (NSCLC)

Objectives: Dato-DXd is a TROP2-directed ADC, recently approved by the US FDA for the treatment of HR+ HER2- unresectable or metastatic breast cancer. Dato-DXd has demonstrated encouraging antitumor activity and tolerable safety in an ongoing Phase 1b study TROPION-Lung02 [ 1, 2] in patients with advanced or metastatic NSCLC, receiving 4 or 6 mg/kg Q3W doses of Dato-DXd in combination with pembrolizumab (doublet) or pembrolizumab and platinum chemotherapy (triplet). The aim of this work was to characterize Dato-DXd’s ER relationship with tumor size (TS) and overall response rate (ORR) in patients with NSCLC receiving the doublet or triplet regimen and identify the relevant baseline characteristics for the antitumor response. Methods: A previously developed population PK model [ 3] was used to derive exposure metrics of Dato-DXd. TS response was modeled using the tumor growth inhibition (TGI) framework described by Han et al. [ 4] and ORR was modeled using logistic regression analysis. Full covariate modeling approach was used in both cases. Covariate effects in the TGI and ORR model were assessed by predicting the % change in TS at nadir from baseline (%TSnadir) and the probability of best overall response (BOR) of complete or partial response (CR/PR) respectively. Results: Among 142 patients, 70 (49.3%) and 72 (50.7%) patients received doublet and triplet regimen, respectively; 96 (67.6%) and 46 (32.4%) patients had zero and ≥1 prior line of therapy (PLT), respectively. Among 46 patients who received ≥1 PLT, 18 (39.1%) patients had their last prior line of therapy with an immuno-oncology (IO) drug (LPIO). PD-L1 expression levels (locally tested) were < 1% or missing in 59 patients (41.5%), 1–49% in 52 patients (36.6%), and ≥50% in 31 patients (21.8%). A TGI model with multiplicative drug effects of Dato-DXd, pembrolizumab, and chemotherapy best described the ER for TS. An initial sigmoidal Emax model linking Dato-DXd cycle-wise AUC with TGI revealed the saturation of drug effect at the first exposure quartile, leading to the implementation of an exposure-independent effect of Dato-DXd in the TGI model to avoid over-parameterization. A log-linear drug effect model with Dato-DXd AUC in Cycle 1 best described the ER for ORR. Univariate analyses of TGI and ORR models revealed that higher PD-L1 expression (1-49% and ≥50%), no PLT, or LPIO had significant effects on the TS reduction and increased probability of BOR being CR/PR. The estimated %TSnadir for each of these covariate effects were >15%, which was significantly higher than 4.6% %TSnadir for the reference subject who received Dato-DXd monotherapy, had PD-L1 expression < 1%, and received ≥1 PLT without last line being IO. Conclusions: This ER analysis is the first characterization of antitumor activities for Dato-DXd in combination regimens with pembrolizumab, with or without platinum chemotherapy, in patients with NSCLC. Key baseline patient characteristics associated with antitumor response were identified and warrant further evaluation in ongoing Phase 3 studies of Dato-DXd in these combination regimens.