Abstract 2155: Joint modeling of safety and peripheral mode-of-action (MoA) biomarkers to support RP2D identification in Phase 1 study of SAR444245 (SAR’245) as monotherapy (mono) or combined with pembrolizumab (pembro) in patients with advanced solid tumors

Abstract Background: SAR’245 is a clinical-stage site-specific pegylated human IL-2 that blocks IL-2 alpha receptor binding but retains near-native binding affinity for beta/gamma IL-2 receptor subunits. When administered in preclinical models, a unique ‘T-cell remodeling’ MoA characterized by robust increase of CD8+Teff/CD4+Treg coupled with potent NK-cell activation/expansion was observed. Circulating tumor DNA (ctDNA) can be used as a non-invasive biomarker of early clinical response whilst overcoming challenges of obtaining repeat tumor biopsies from patients. Previously, we reported results of the Phase 1 HAMMER study (NCT04009681); herein, we describe an innovative integrative approach that considers peripheral key MoA biomarkers with objective response and dose-limiting toxicity (DLT) rate to help identify the recommended Phase 2 dose (RP2D) for SAR’245. Methods: SAR’245 was given IV as mono Q3W [Cohort B] or Q3W + IV pembro 200 mg Q3W/400 mg Q6W [Cohort C]. Joint modeling was carried out to account for the relationship between dose and 1) MoA biomarkers, and NK cells in blood measured by flow cytometry; 2) response surrogate biomarker: ctDNA measured by Guardant Omni 500 panel; and 3) DLTs. In the model, a latent variable was used to model correlation between DLT and the MoA or response surrogate biomarkers, Bayesian approach was used to derive posterior probabilities (PP) at each dose level of the target region (defined by >20% probability of fold change of biomarker values post-treatment above a predefined threshold and DLT rate <33%) and the PP of the overdose region (defined by DLT rate ≥33%). RP2D dose level was determined by maximizing the PP of target region among doses with PP of the overdose region probability <40%. Results: Samples from 35 subjects (Cohort B) and 34 subjects (Cohort C) were available. The results from SAR’245 mono suggests that the CD8/CD4 ratio and the concentrations of NK, CD8, and CD4 achieve maximum probability of reaching meaningful modulation (based on pre-defined threshold) around 32 ug/kg. When SAR’245 was combined with pembro, the results with PoM biomarkers showed the best performance at 24-32 µg/kg, while results with ctDNA showed the best results at 16-24 µg/kg. When all parameters were considered, either 24 or 32 µg/kg could serve as an adequate dose at Q3W scheduling. Conclusions: In early oncology studies, joint modeling using non-invasive biomarkers, including MoA and response biomarkers, and a safety profile can inform dose-response relationships and support RP2D selection. This innovative integrative modeling will guide clinical study design. Studies of SAR’245 that further explore the dosing and scheduling are on-going. Disclosures: This study was sponsored by Synthorx, a Sanofi company. Citation Format: Siqing Fu, Gerald S. Falchook, Minal Barve, Meredith McKean, Tira J. Tan, Charlotte Lemech, Cheng E. Chee, Neyssa Marina, Giovanni Abbadessa, Robin Meng, Federico Rotolo, Hong Wang, Jason Deng, Wenting Wang, Rui Wang, Tarek Meniawy. Joint modeling of safety and peripheral mode-of-action (MoA) biomarkers to support RP2D identification in Phase 1 study of SAR444245 (SAR’245) as monotherapy (mono) or combined with pembrolizumab (pembro) in patients with advanced solid tumors [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2023; Part 1 (Regular and Invited Abstracts); 2023 Apr 14-19; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2023;83(7_Suppl):Abstract nr 2155.