Safety, pharmacokinetics (PK), pharmacodynamics (PD) and preliminary efficacy of AZD2936, a bispecific antibody targeting PD-1 and TIGIT, in checkpoint inhibitor (CPI)experienced advanced/metastatic non-small-cell lung cancer (NSCLC): First report of ARTEMIDE-01

9050 Background: PD-1 and TIGIT are implicated in cancer-related T cell immunosuppression. AZD2936 is a bispecific, humanized IgG1 (triple-mutated to minimize Fc effector function) targeting PD-1 and TIGIT. It has demonstrated encouraging activity compared to both anti-PD-1 and anti-PD-1/-TIGIT combinations in murine models. We report data from dose escalation (Part A) and an expansion cohort (Part B) of the first-in-human study ARTEMIDE-01 (NCT04995523). Methods: This open-label, multicenter study enrolled pts with advanced NSCLC who had prior CPI treatment and a PD-L1 tumor proportion score ≥1%. Part A evaluated doses of 70–1500 mg IV Q3W; Part B evaluated the recommended phase 2 dose (RP2D). Primary endpoints included safety, tolerability, dose-limiting toxicities (DLTs) and preliminary efficacy. Secondary endpoints included PK and PD, defined as percentage PD-1 and TIGIT receptor occupancy (RO). Results: As of Dec 5, 2022, 80 pts were enrolled (Part A, n=48; Part B, n=32). Pts were 62.5% male, median age 63.5 years; 72.5% had adenocarcinoma, 23.8% had squamous cell carcinoma; 96.3% had metastatic disease; and 22.5% had brain metastases. They had a median of 2 prior regimens. Median duration of therapy was 11 wks. AZD2936 was well tolerated with no DLTs. In total, 46.3% of pts had treatment-related adverse events (TRAEs), all grade 1–3; the most common were pruritus, rash and lipase increased (6.3% each). Serious TRAEs occurred in 3 pts (3.8%): immune system disorder, acute hepatitis and fatigue (n=1 each). AZD2936 systemic exposure increased in a near dose-proportional manner. Doses of ≥210 mg achieved ~90% PD-1 and TIGIT RO in peripheral T cells. The RP2D was determined to be 750 mg Q3W based on safety, preliminary efficacy, PK/PD and modeling analysis predicting intratumoral RO. Among 76 evaluable pts, 3 had a partial response and 30 had stable disease (Table). Time to response was 1.9–4.0 mos and duration of response was 2.1–6.4 mos. Conclusions: In this interim analysis, AZD2936 showed an acceptable safety profile and preliminary antitumor activity in pts with advanced/metastatic NSCLC previously treated with standard therapy including CPIs. Further exploration of AZD2936 in CPI-naïve NSCLC pts, including a randomized dose optimization cohort is ongoing. Clinical trial information: NCT04995523 . [Table: see text]