Tremelimumab (T) + durvalumab (D) + chemotherapy (CT) vs pembrolizumab (P) + CT in 1L non-squamous (NSQ) metastatic NSCLC (mNSCLC) with <i>STK11</i> , <i>KEAP1</i> , and/or <i>KRAS</i> mutations (mut): Interim analysis (IA) of the phase 2b TRITON study.

Abstract

8515 Background: Pts with STK11 , KEAP1 and/or KRAS -mutated mNSCLC may benefit from the addition of anti-CTLA-4 to anti-PD-(L)1-based chemo-immunotherapy. In the phase 3 POSEIDON study, 1L T+D+CT significantly improved OS vs CT in pts with mNSCLC. Exploratory analyses showed sustained OS improvement with T+D+CT vs CT in subgroups with STK11 , KEAP1 and/or KRAS mut; in each mut subgroup, magnitude of OS benefit with T+D+CT vs CT was numerically greater than with D+CT vs CT. The phase 2b, open-label, multicenter, US-based TRITON study is comparing 1L T+D+CT vs P+CT in pts with NSQ mNSCLC and STK11, KEAP1 and/or KRAS mut. Here we report results of a planned IA (data cutoff [DCO] 15 mo after 1st pt randomized) of objective response rate (ORR), duration of response (DoR) and safety. Methods: Pts with treatment [tx]-naïve, EGFR / ALK wild-type, NSQ mNSCLC and STK11 , KEAP1 and/or KRAS mut (ECOG PS 0/1) were randomized 1:1 to T+D+CT or P+CT. T+D+CT arm: T 75 mg + D 1500 mg + pemetrexed-platinum Q3W for 4 cycles, then maintenance D + pemetrexed Q4W until disease progression (PD); additional doses of T given at week 16 and, optionally, at mo 24. P+CT arm: P 200 mg + pemetrexed-platinum Q3W for 4 cycles, then maintenance P + pemetrexed Q3W until PD, for up to 24 mo. Randomization was stratified by mut type and tumor PD-L1 expression (≥1% vs &lt;1%). The primary endpoint is PFS (RECIST v1.1; investigator-assessed) with planned sample size ~100 pts. Key secondary endpoints include OS, ORR, DoR and safety. Results: At DCO (12 Nov 2025), 41 pts were randomized to T+D+CT and 43 to P+CT. Overall, 27.4%, 21.4% and 78.6% of pts had STK11 , KEAP1 and KRAS mut (not mutually exclusive); 39.3% had PD-L1 &lt;1%. In the T+D+CT vs P+CT arms, median (range) age was 69 (47–82) vs 69 (51–86) y, 48.8% vs 62.8% pts were male, 70.7% vs 72.1% were White and 14.6% vs 20.9% Black or African American, 78.0% vs 62.8% had ECOG PS 1. Median (range) safety follow-up for D/P was 5.6 (0.0–14.0)/5.1 (0.0–17.5) mo; median no. of D/P doses was 8/7. ORR (95% CI) was 39.0% (24.1–54.0) in the T+D+CT arm vs 34.9% (20.6–49.1) in the P+CT arm [unconfirmed ORR 48.8% (33.5–64.1) vs 41.9% (27.1–56.6)]. Median DoR (95% CI) was not reached (6.3–NR) vs 6.4 (4.2–NR) mo; 100% vs 58.3% pts remained in response at 6 mo. In KRAS mut-only pts (n=52; exploratory), ORR was 48.0% (12/25) with T+D+CT vs 33.3% (9/27) with P+CT. The sponsor remains blinded to PFS at this IA. With T+D+CT vs P+CT, 41.5% vs 41.9% of pts had Grade 3/4 AEs possibly related to tx (TRAEs); 2.4% vs 4.7% had TRAEs leading to tx discontinuation and 0% vs 2.3% had TRAEs leading to death. Conclusions: At IA, ORR and DoR results from the prospective TRITON study support a role for the addition of anti-CTLA-4 (T) to 1L anti-PD-L1 (D) + CT in pts with STK11 , KEAP1 and/or KRAS -mutated mNSCLC. Safety was similar in the two arms and consistent with known safety profiles. Clinical trial information: NCT06008093 .